New to Operating?

Check out these reader favorites:

• Operating Economics: The Optimization Company (Amazon)

• Operating by Warren & Charlie

• Winning the Loser’s Game: The Creator Economy and the Companies We All Need to Build

• 0 to 55,000 - The First 90 Days Playbook - What I’d Do Differently Knowing What I Know Now

Become a paid subscriber to join the Operating Skool community and attend our weekly live Office Hour Q&A’s w/ John.

Subscribe now

Operating Stories No. 07 · By John Brewton

In October 2025, Mark Walter closed his purchase of the Los Angeles Lakers at a $10 billion valuation. It is the highest price ever paid for an American sports franchise, $3.9 billion above the Boston Celtics sale to Bill Chisholm seven months earlier. The Buss family, who had controlled the franchise since 1979, kept roughly 15 percent and Jeanie Buss kept the governor seat.

Forty-six years before that closing, Jerry Buss had purchased the Lakers, the Forum, the LA Kings, and a 13,000-acre Sierra Nevada ranch from Jack Kent Cooke for $67.5 million. Roughly $16 million of that total was allocated to the basketball franchise itself.

That $16 million compounded into the $10 billion sale. The math is 625x over 46 years, or a 14.9 percent compound annual growth rate. For comparison, a $16 million S&P 500 buy-and-hold over the same period would have grown to roughly $1.9 billion. The Buss family bought one private illiquid asset and outperformed the S&P 500 by 5x. Plus seventeen championships.

Eight months before Walter closed, the Lakers acquired Luka Dončić from the Dallas Mavericks in a three-team trade that reporters called the most unexpected transaction in American sports history. Luka was 25, an All-NBA First Team selection in his prime, and had not requested a trade.

That sequence looks like luck. It is not. It is the seventh execution of the same operating decision the franchise has made for 78 years.

The Operating Thesis

The Los Angeles Lakers are not a basketball franchise. They are an 80-year case study in concentrated talent acquisition. The franchise has won 17 championships across seven generational acquisitions. Every era has been anchored by a single decisive operating decision to acquire one transcendent star. The infrastructure (coaches, role players, ownership succession, the city itself) exists to protect and amplify that asset.

The basketball is the byproduct.

Seven Decisions

1. 1947 · George Mikan: First pick in the NBL dispersal draft after the Chicago Gears folded. Five championships in six years. The franchise template.

2. 1968 · Wilt Chamberlain: Trade from Philadelphia for Imhoff, Clark, Chambers, and cash. First explicit star-stack with West and Baylor. Won the 1972 title with a 33-game winning streak that still stands.

3. 1979 · Magic Johnson: First overall pick. Cooke held the pick from a prior trade with New Orleans. Selected Magic on June 25. Buss completed his $67.5 million purchase four days later. Five titles in nine years. The franchise became Showtime, explicitly Hollywood-engineered.

4. 1996 · Shaq plus Kobe: A coordinated single decision in a single calendar week. Signed Shaq to $121 million, the largest contract in NBA history at the time. Traded Vlade Divac to Charlotte for the rights to Kobe Bryant. Three consecutive championships, 2000 to 2002.

5. 2008 · Pau Gasol: Trade from Memphis for Kwame Brown, Javaris Crittenton, the rights to Marc Gasol, and two future first-round picks. Popovich called it a heist. Two titles in three years.

6. 2018 · LeBron James: Free agency, four years and $154 million. The first acquisition where the franchise’s accumulated brand gravity was the recruiting mechanism. No assets traded, no punitive contract absorbed. The seventeen banners did the work. Won the 2020 bubble title.

7. 2025 · Luka Doncic: Sent Anthony Davis, Max Christie, and a 2029 first to Dallas. Acquired the era’s third-best player in his prime, age 25, while the previous-generation star (LeBron) was still on the roster. The first time the cycle did not require a generational gap.

Subscribe now

Read more

All visuals for this article were produced in my Claude studio. If you’d like to be able to do the same, make sure you register for the Compound Content Studio, launching this Friday.

The Compound Content Studio™ is a live, 2‑week AI immersion where we build your Claude-powered content system that multiplies your ideas into an endless stream of on-brand content.

We kick off May 1st, and there are only 50 spots (25 of which have already been sold). If you want your content to compound rather than reset to zero every day, this is where you build the system that finally makes it happen.

This is the system I built to create all of my content across social platforms. Jumpstart and scale your brand and content this May!

Grab Your Studio Seat Here

48 Hours

On February 13, 2026, Anthropic publicly launched Cowork, a Claude-powered collaborative workspace that can read, write, and operate across a user’s applications simultaneously. Within 24 hours, the combined market capitalization of the ten largest US enterprise software companies fell by $285 billion. Adobe, Salesforce, Intuit, Microsoft. They all moved the same direction.

That $285 billion is not the most interesting number. The most interesting number came eight weeks later. On April 16, 2026, Anthropic released Claude Opus 4.7. On April 17, it released Claude Design, an Anthropic Labs product that lets a founder describe a pitch deck in plain English and produce a working prototype. Figma’s stock dropped 7.28% on the day. Three days before the launch, Anthropic’s Chief Product Officer, Mike Krieger, had quietly resigned from Figma’s board. The market missed the signal. The operating lesson is that Krieger did not resign because he was getting uncomfortable. He resigned because he already knew what was about to ship.

Anthropic now generates $30 billion in annualized revenue. It carries a $380 billion valuation and has received unsolicited offers at $800 billion. It has raised $65 billion across eleven funding rounds in five years. Eight of the ten Fortune 10 companies are customers. More than 500 businesses spend over $1 million per year on Claude. Claude Code went from zero to $2.5 billion in annualized billings in nine months.

For an operator, the right question about Anthropic is not how they did this. The right question is what they chose correctly that their competitors chose incorrectly. The answer is three things. A method decision. A constraint decision. A vertical integration decision. Every other operating outcome flows from those three.

Operating Thesis

First, Anthropic treated safety as the product, not a constraint on the product. Constitutional AI, the Responsible Scaling Policy, and the Long-Term Benefit Trust were not brand elements. They were the technical and governance foundation on which enterprise trust compounded. Every pharmaceutical company, bank, and public-sector buyer that runs Claude in a regulated workflow runs it because Anthropic’s safety disclosures made procurement possible.

Second, Anthropic built from the model up, not from the chatbot down. Every product release, in sequence, is a move up the stack from the model to the workspace to the outcome. Claude 1 sold capability. Claude 3.5 sold a workspace. Claude Code sold a developer. Cowork sold a department. Claude Design sold a designer. Each move captured more of the value that used to flow to the application layer.

Third, Anthropic treated each capability release as a governance decision first and a revenue decision second. When a model was too dangerous to ship broadly, it did not ship broadly. Claude Mythos, announced on April 7, 2026, was restricted to forty cybersecurity partners because its capabilities threatened operating systems globally. A normal company would have sold it. Anthropic withheld it and gained trust instead.

Three decisions. Five years. $380 billion. The pattern maps to three historical operators who made the same three decisions in their own eras.

Velocity

The revenue arc is what it is. $100 million annualized at the end of 2022. $1 billion by December 2024. $5 billion by August 2025. $7 billion by October 2025. $30 billion by April 2026. Three consecutive years of roughly 10x annual revenue growth at this scale is unverified in the enterprise technology record. Salesforce took eleven years to hit $5 billion in revenue. Amazon Web Services took about seven years to hit $10 billion. Anthropic hit $10 billion in roughly 24 months after crossing $1 billion.

The driver is enterprise concentration, not consumer subscriptions. Accounts generating more than $100,000 in annual recurring revenue grew 7x in 2025 over 2024. The total enterprise customer base grew from fewer than 1,000 in late 2023 to more than 300,000 in September 2025. That is a 300x expansion in under two years.

Claude Code is the inside story. It launched as a research preview in February 2025 and hit general availability in May 2025. It went from zero to over $2.5 billion in annualized billings in roughly nine months. Business subscriptions on Claude Code quadrupled in the six weeks following January 1, 2026. That is the fastest revenue ramp of any product in enterprise software history.

Here is the paradox. Anthropic will spend more than $50 billion on compute infrastructure in 2026 alone against $30 billion in annualized revenue. Dario Amodei publicly said in February 2026 that a 12-month delay in AI progress would make him personally bankrupt. The company is simultaneously one of the fastest-growing technology companies in history and one of the largest consumers of capital per dollar of revenue ever. Both things are true. The flywheel requires it.

New to Operating?

Check out these reader favorites:

• Operating Economics: The Optimization Company (Amazon)

• Operating by Warren & Charlie

• Winning the Loser’s Game: The Creator Economy and the Companies We All Need to Build

• 0 to 55,000 - The First 90 Days Playbook - What I’d Do Differently Knowing What I Know Now

Become a paid subscriber to join the Operating Skool community and attend our weekly live Office Hour Q&A’s w/ John.

Subscribe now

Three Product Inflection Points

The full product history has fifteen releases. Three of them mattered most.

Constitutional AI, December 2022. Before Anthropic shipped a product, it published a paper. “Constitutional AI: Harmlessness from AI Feedback” described a method for training large language models to critique and revise their own outputs against a written set of principles, replacing most human annotation with a principled self-improvement loop. The technique cut human labeling costs by 10x to 100x while producing models that were simultaneously more helpful and less harmful than prior approaches. Every Claude model since has been trained on this foundation. The strategic implication was procurement, not research. Constitutional AI gave Anthropic’s enterprise sales team verifiable, documented claims about model behavior that pharmaceutical, financial, legal, and public-sector buyers required before deploying AI in consequential workflows. It turned safety into a product feature.

Model Context Protocol, November 2024. Anthropic released MCP as an open standard for connecting AI systems with external data and tools. MCP replaced fragmented, bespoke API integrations with a universal protocol. Any AI agent built on MCP could connect to any MCP-compatible data source without custom connector code. In December 2025, Anthropic donated MCP to the Agentic AI Foundation, a directed fund within the Linux Foundation. This was the platform move. By establishing MCP as the standard, Anthropic positioned itself as the architect of the infrastructure layer through which AI agents operate. The economics are analogous to controlling HTTP or the dominant JDBC driver.

Cowork and Claude Design, February and April 2026. Cowork is a collaborative workspace where Claude reads, writes, and acts across applications simultaneously. A Claude-powered agent inside Cowork can open a user’s email, pull data from a spreadsheet, run analysis, draft a response, and file a follow-up task, all without handoffs. Claude Design lets a founder or product manager describe a design in plain English and produce a functional interactive prototype. Inputs include text prompts, uploaded images, DOCX, PPTX, XLSX files, and a web capture tool for grabbing live UI elements from existing websites. Claude Design is included at no extra charge for Claude Pro, Max, Team, and Enterprise subscribers. Cowork is the vertical move into the business user’s deliverable. Claude Design is the vertical move into the designer’s deliverable. Both are Rockefeller moves.

Three Parallels

Every high-velocity operating company has antecedents. Anthropic’s three most useful antecedents are the Traitorous Eight who founded Fairchild Semiconductor in 1957, Andy Grove’s memory-to-microprocessor pivot at Intel in 1985, and John D. Rockefeller’s climb up the Standard Oil stack from 1870 to 1890. Each one made a method decision that looked wrong in the short run and built an industry in the long run. Each one matches a specific Anthropic move, year for year.

The Traitorous Eight: method over hierarchy

In September 1957, eight scientists walked out on William Shockley. In late 2020 and early 2021, eight researchers walked out on OpenAI. The parallel is not the number. It is the substance. Shockley believed authority and hierarchy produced scientific output. Robert Noyce and the other seven believed peer review and diffusion of authority produced it. The eight who left OpenAI, Dario Amodei, Daniela Amodei, Jared Kaplan, Tom Brown, Sam McCandlish, Jack Clark, Benjamin Mann, and Christopher Olah, believed scaling compute alone was insufficient without a parallel rigorous focus on alignment. Both walkouts were method disagreements. Both founder cohorts built the company that was right about the method.

Fairchild’s alumni went on to found more than 65 companies including Intel, AMD, and Kleiner Perkins. Anthropic’s diffusion effect is already visible in year five. Model Context Protocol, donated to the Linux Foundation in December 2025, is the integrated circuit of the agent era. The Long-Term Benefit Trust is the governance equivalent of Fairchild’s stock option plan.

Andy Grove at Intel: trust as the moat

By 1985, Intel faced commodity pressure from Japanese DRAM manufacturers. Memory margins collapsed to 1%. Grove and Gordon Moore had the conversation. If the board replaced them, a new CEO would exit memory. So they walked out the door, came back in, and did it themselves. Revenue fell from $1.63 billion in 1984 to $1.27 billion in 1986 before recovering to $3.13 billion in 1989. By 1992, Intel was the largest semiconductor company in the world. Grove documented the framework in Only the Paranoid Survive.

Anthropic ran the Grove move twice. First in mid-2022 when it withheld Claude as OpenAI was about to release ChatGPT. Anthropic walked away from the quarter’s revenue because it had correctly identified the constraint was trust, not speed. Second in April 2026 when it restricted Claude Mythos to forty cybersecurity partners through Project Glasswing. Anthropic had a model capable of hundreds of millions in annual revenue if released broadly. It released it narrowly. Both decisions made trust the product. The enterprise perception that Claude is the only frontier model you can safely deploy in regulated contexts is the moat. That moat is worth more than Mythos revenue would have been.

Rockefeller and Standard Oil: the stack climb

Rockefeller did not build Standard Oil by refining oil better than his competitors. He built it by controlling the stack above and below refining. By 1890, Standard Oil controlled 88% of refined oil flows in the United States, most of the pipelines, a material share of rail bulk freight, and the retail distribution network that ended at the customer’s kerosene lamp.

Anthropic’s playbook is a compressed Standard Oil. The model is the refinery. Claude.ai is the retail distribution. Memory and Projects are the storage tanks. Skills are the refined products. MCP is the pipeline network. Claude Code is the vertical move into the developer’s output. Cowork is the vertical move into the business user’s output. Claude Design is the vertical move into the designer’s output.

The Standard Oil pattern is why the Cowork release wiped $285 billion in market cap in 24 hours. The market understood, correctly, that Anthropic had just completed the Rockefeller move on the SaaS layer. Every enterprise software category charging per-seat for a specific UI on top of specific data had just been informed that the UI was optional and the data was accessible through a new control layer.

For the full business history underneath these three parallels, including the Cleveland Massacre’s six-week timeline, Grove’s six-force framework for strategic inflection points, and the structural reasons the Fairchild diaspora produced 65+ companies, see the Historical Parallels Compendium in the asset hub below.

Four operating principles

Answer the correct question completely, not the approximate question fast. The operating discipline is to identify the correct question before anyone else does. For Anthropic, the correct question was not how to build the best chatbot. The correct question was how to build AI that enterprises can actually deploy in regulated workflows. Constitutional AI, the Responsible Scaling Policy, the Public Benefit Corporation structure, the Long-Term Benefit Trust, and the MCP protocol are all different angles on the same correct answer. Each took longer than the equivalent ChatGPT feature would have taken. Each produces more durable value.

Make the constraint the product. In every business there is a constraint competitors treat as a cost that you can treat as the product. For Anthropic, safety was that constraint. The received wisdom in AI in 2021 was that safety was an overhead, a drag on capability that you had to accept because regulators would eventually demand it. Anthropic inverted the assumption. Safety became the thing enterprises were buying. Constitutional AI was a feature, not a tax. The Responsible Scaling Policy was a selling point, not an apology. Find the overhead line item in your own business that your competitors all treat the same way. That line item is probably your product.

Climb the stack toward the deliverable. Every business that sells a tool gets commoditized. Every business that sells an outcome gets to charge for the value of the outcome. You start by selling the tool. You add a workspace that captures the user’s state. You add integrations that make the workspace the default. You climb down the stack to the deliverable the user was going to produce with your tool anyway. Each step up the stack increases switching costs. Each step down the stack captures value that was flowing to adjacent vendors. For Anthropic, the climb from Claude 1 (tool) to Cowork (outcome) is the most compressed version of that pattern on record.

Let governance precede revenue, every time. Every time Anthropic has gained a new capability, it has made a governance decision before making a revenue decision. Claude 2022 was held back. Mythos 2026 was restricted to forty partners. Opus 4.7 shipped with automatic detection and blocking of high-risk cybersecurity requests. Each decision was made against a public, specific, and auditable written standard. Each decision forewent short-term revenue to reinforce a long-term trust position. Your strongest long-term asset is the thing your competitors are willing to compromise that you are not.

The Operating Stories analysis continues below for paid subscribers, including the broader implications for professional services, the operating lesson for founders, and the full asset hub: three research compendiums (Product and Capability, Financial and Strategic, and the Historical Parallels business history compendium covering the Fairchild founding, Grove’s DRAM exit, and Rockefeller’s stack climb in full), the five single-page operating guides for Claude Projects, Skills, Cowork, Claude Code, and Claude Design, and the complete primary sources appendix.

Subscribe now

All visuals for this article were produced in my Claude studio. If you’d like to be able to do the same, make sure you register for the Compound Content Studio, launching this Friday.

(If you’re an Operating Founder member, message me directly for details)

Grab Your Studio Seat Here

Read more

A personal note before the trilogy

This is the first article in a three-part Operating Stories arc on Uber. Article two looks at Uber as a multi-product logistics and commerce operator. Article three looks at the autonomy strategy as a textbook application of aggregation theory. I am publishing them as a deliberate arc (an Operating Stories tryptych) because the story of Uber under Dara Khosrowshahi is the most under-read operating turnaround of the last decade, and it rewards reading in order.

If you want the fast version of the whole trilogy in one paragraph, here it is.

Between 2017 and 2026, a single CEO and a disciplined team executed three turnarounds in series inside one company. A cost turnaround. A platform turnaround. A capital turnaround. Every one of them is stealable by an operator running a ten-person company or a ten-thousand-person company. This first piece is about the first one.

New to Operating by John Brewton?

Three reader favorites to start with:

• Operating Economics: The Optimization Company (Amazon)

• Operating by Warren and Charlie (Berkshire Hathaway)

• Operating Stories: Louis Vuitton and the Business of Collaboration

Subscribe now

A $20 billion swing that nobody credits to operating discipline

In 2019, Uber reported a $5.2 billion GAAP net loss on $14.1 billion in revenue.

In 2025, Uber reported $10.0 billion in free cash flow on $52.0 billion in revenue. Adjusted EBITDA reached $8.7 billion. Adjusted EBITDA margin on gross bookings reached 4.5 percent, up from 4.1 percent the prior year. Gross bookings for the full year crossed $193 billion. The company was in the S&P 500, which it had joined in December 2023, and returned roughly half of its free cash flow to shareholders through buybacks during that two-year window.

The swing from a $5.2 billion loss to $10 billion in cash generation over six calendar years represents roughly $15 billion in annual P&L improvement, within a company that was already producing more revenue than the 85th-largest country by GDP. Against 2017, when Dara Khosrowshahi took over an Uber bleeding $4.5 billion in operating losses, the cumulative swing between the 2017-2019 loss window and the 2023-2025 profit window is closer to $21 billion.

Twenty-one billion dollars is roughly the combined market capitalization of Hertz, Etsy, and Yelp. It is three times the price Verizon paid for AOL, about twice what Microsoft paid for LinkedIn, and more than the entire market value of the United States Postal Service’s implicit government guarantee. It is not a rounding error. It is the largest nominal P&L swing ever achieved by a consumer technology company within a single CEO's tenure.

The stock market’s response over the same window is the second surprise. The IPO priced in May 2019 at $45 per share and closed its first trading day at $41.90, down seven percent. By March 2020, in the depths of COVID, the stock had fallen to $13.71. The 2026 trading range sits above $67. The gain from IPO is around 50 percent. The gain from the COVID low is over 400%. Over the same window, the Nasdaq Composite roughly doubled. Uber, the company Wall Street spent most of 2019 and 2020 describing as structurally unprofitable, compounded at a pace that the word structurally was never designed to accommodate.

That is the opening.

The tension the article will work with is as follows: The conventional explanation for the turnaround is that Uber grew its way into profit. The financial record shows that the turnaround happened before the growth reacceleration, not because of it, and that growth was the consequence of the turnaround, not the cause.

The thesis

The Khosrowshahi-era Uber turnaround was not a growth story that finally achieved scale. It was a capital-allocation turnaround executed under the cover of a growth narrative. Uber did not grow its way to profit. It divested, deconsolidated, refused, and re-priced its way to profit. Then it grew.

The financial spine

I want to walk through the financial evidence carefully because the headlines have buried what actually happened.

Start with the pre-Khosrowshahi baseline. In 2017 Uber reported a $4.5 billion GAAP operating loss and a roughly $4.5 billion net loss on approximately $7.0 billion in revenue.

The company was carrying eight figures of monthly burn, a pending Waymo lawsuit, an internal cultural crisis, and a board fight that ended with the departure of the founding CEO. Dara Khosrowshahi took the job in August 2017.

His first full year in the seat, 2018, produced $11.3 billion of revenue and a $1.8 billion GAAP net loss. Operating loss was still above $3 billion.

The headline most often written about 2018 was that Uber’s losses were narrowing. That was true, and it was also beside the point.

The turnaround is visible in one image. Losses narrow from 2018 through 2022. GAAP operating profit arrives in 2023. Free cash flow separates from net income sharply in 2025.

The IPO arrived in May 2019. The filing priced at $45 per share, valuing the company at approximately $82 billion on a fully diluted basis. Day one closed at $41.90. The stock spent most of its first twelve months as a public company below $35. Inside the 2019 filings, Uber took a $5.2 billion GAAP net loss, of which a significant portion was stock-based compensation that Khosrowshahi described on the earnings call as a once-in-a-lifetime charge. The description was financially accurate. It was also a marker. The company had chosen to front-load the most painful accounting into the earliest post-IPO quarter. A management team planning to compound over a decade does not spread the pain. They take it.

Between 2020 and 2022 the company pushed through two things in parallel. The first was a set of divestitures that removed unprofitable or non-core businesses from the balance sheet. The second was a cost-structure overhaul that cut headcount by roughly 6,700 people in 2020 alone, trimmed variable driver incentives back toward sustainable unit economics, and flattened organizational hierarchy inside the core business. By the end of 2022, gross bookings had more than doubled from the 2020 trough, operating losses were inside one billion dollars, and adjusted EBITDA had crossed zero.

2023 is the inflection. First full-year GAAP operating profit: $1.8 billion. First full-year GAAP net income: $1.8 billion. First inclusion in the S&P 500, announced in December. Free cash flow of $1.1 billion. Revenue of $31.9 billion, up 17 percent year over year.

2024 consolidated the win. Revenue of $44.0 billion. Adjusted EBITDA of $1.8 billion at the annual level, expanding to a $2.5 billion quarterly run rate by Q4. GAAP net income of $4.4 billion, boosted in part by a deferred tax asset release but durable on an underlying basis. Free cash flow of $1.7 billion. Monthly active platform consumers crossed 170 million. Trips per monthly active user reached an all-time high in Q2 2024 and kept climbing.

2025 compounded it. Revenue of $52.02 billion, up 21 percent year over year. Adjusted EBITDA of $8.7 billion, up 35 percent year over year. Adjusted EBITDA margin on gross bookings of 4.5 percent. Gross bookings of $193 billion, up 20 percent. Free cash flow of $10.0 billion, up 42 percent year over year, and crossing 90 percent of adjusted EBITDA conversion. Mobility take rate held in the 27 to 30 percent range across quarters. Advertising crossed a $1.5 billion run rate at 60 percent year-over-year growth. Uber One membership passed 30 million subscribers.

Register (Free) Claude For Operators

Cash conversion is the real story. Free cash flow ran ahead of GAAP net income in 2025, signaling that the business is now structurally cash-generative rather than accounting-optimized.

Guidance for Q1 2026, filed alongside the Q4 2025 release in early February 2026, called for $52.0 to $53.5 billion in gross bookings, up 17 to 21 percent in constant currency, and adjusted EBITDA of $2.37 to $2.47 billion. Non-GAAP earnings per share at the midpoint implied 37 percent year-over-year growth. The Q1 2026 earnings call is scheduled for May 6, 2026. I will update the series with live Q1 numbers when they print.

Now I want to stop and pull one specific number out of the spine. Compare 2019 to 2024. Revenue grew from $14.1 billion to $44.0 billion, a factor of roughly 3.1. Headcount grew from approximately 22,000 to approximately 30,800, a factor of roughly 1.4. Revenue per employee grew from roughly $640,000 to roughly $1.43 million, a factor of 2.2. The 80 percent growth in gross bookings against a 10 percent growth in employees that Khosrowshahi cited on the 2024 investor day was not a favorable ratio. It was a structural re-engineering. The company built a cost curve that separates from its revenue curve, and operating leverage now drops through the whole P&L.

Revenue per employee more than doubled under the turnaround. Operating leverage is a hiring-plan concept before it is a financial one.

That is the spine. The losses narrowed, then they stopped, then they reversed, and now the cash conversion is widening faster than revenue is growing. The interesting question is how.

The mechanism: a capital-allocation turnaround hiding inside a growth story

The conventional explanation for Uber’s profitability is that the business finally scaled, so unit economics finally turned positive. That explanation is not wrong. It is incomplete. It misses the three things that actually produced the turnaround.

One. The divestiture program.

Between 2016 and 2022, Uber systematically removed businesses from its balance sheet that either could not be won or could not be made profitable under its cost of capital. Russia went to Yandex. Southeast Asia went to Grab. China went to Didi. Uber Eats India went to Zomato. Uber Advanced Technologies Group, the in-house autonomous vehicle effort, went to Aurora in December 2020 for equity and a board seat. Drizly, acquired in 2021 for $1.1 billion, was shut down in 2024 after the integration failed to deliver. Cornershop’s minority partners were consolidated and absorbed into the Uber Eats grocery surface rather than run as a separate brand.

The gross cash generated from those divestitures through Q1 2026 is estimated between $8 billion and $12 billion. Net of the losses taken on publicly traded stakes (Grab, Zomato, Aurora), the realized balance sheet gain was smaller, on the order of $2 billion to $3 billion. The cash mattered less than the removal. Every divested business had been consuming management attention, regulatory bandwidth, and capital that was now available for redeployment into mobility, delivery, advertising, and membership.

Most public-company CEOs do not sell. Selling is unglamorous. It triggers impairment charges. It is read by the press as retreat. The Khosrowshahi-era Uber sold aggressively from a position of weakness between 2018 and 2021, and then it kept selling from a position of strength between 2022 and 2024. Treat divestiture as a product. Operators who want to emulate the turnaround should start here.

Two. The unit-economic rebuild.

Uber rebuilt driver economics before it touched consumer pricing. This is backwards from most consumer companies. Most consumer companies cut product quality or raise prices on consumers first. Uber did the harder thing, which was to reconstruct the match between rider, driver, and platform incentive, and then push pricing through once the match was structurally sound.

The work happened inside the 2019 to 2022 window. Driver incentive variable spend was rationalized away from the aggressive match rates that had been used in 2016 and 2017 to build supply during the land-grab years. Surge pricing was re-tuned to reduce variance on the driver side while preserving the spike-day revenue capture on the consumer side. Mobility take rate moved from a mid-20s range compressed by incentive spend to a stable 27 to 29 percent range. In Q4 2025 mobility take rate reached 29.9 percent.

On the delivery side the path was different. Delivery take rate declined from roughly 20 percent in 2022 to an estimated 18 to 19 percent through 2025, as DoorDash competitive intensity forced operators to keep restaurant and merchant economics attractive. Uber did not fight that compression. It stacked a higher-margin revenue layer (advertising) and a higher-attach revenue layer (membership) on top, so that delivery could run at slightly thinner transaction margins while delivering rising profit per active user.

The two segments diverge on take rate and converge on contribution. Mobility compresses supply cost to protect rate. Delivery accepts rate compression and stacks margin on top through advertising and membership.

Three. The capital-return commitment.

At the February 2024 investor day, the company committed to returning roughly 50 percent of free cash flow to shareholders through buybacks and, in time, dividends. That commitment was made while free cash flow was still $1.7 billion, which made the commitment non-trivial. The 2025 number of $10 billion in free cash flow vindicated the commitment. The first year of buybacks took place under that policy and retired meaningful share count. Net share count now trends down over time rather than up.

The discipline matters more than the absolute number. Returning capital forces a company to treat free cash flow as a resource that has a cost. It breaks the internal habit of treating cash as a rainy-day buffer that management can spend on acquisitions of declining strategic relevance. Uber committed to the discipline before the result. That ordering is what separates an operating company from a treasury.

Historical Operating Parallels

The mechanics above map cleanly onto three of the most studied corporate turnarounds of the last thirty years. I want to walk through each one because the mapping is where the playbook becomes transferable.

Parallel one: Lou Gerstner at IBM, 1993-2002.

In April 1993 Lou Gerstner took over an IBM that had lost $8.1 billion the prior year on $64.5 billion of revenue. The analyst consensus, and it was a loud consensus, was that IBM had to be broken up into independent business units, each of which would have a chance to compete in its own market. Gerstner declined. He believed that IBM’s durable advantage was the ability to deliver integrated solutions to enterprise customers, and that breaking up the company would forfeit exactly the thing that made the whole worth more than the sum of its parts.

Gerstner’s operating playbook across the next nine years included three moves that are directly relevant to Uber. The first was decisive cost action. IBM reduced headcount by roughly 100,000 people between 1993 and 1994. The second was divestiture of non-core businesses. IBM sold the Federal Systems Company to Loral in 1994 for approximately $1.5 billion and walked away from a semiconductor strategy that was consuming capital without producing strategic return. The third, and the most important, was the construction of a higher-margin business (IBM Global Services) on top of the lower-margin legacy business (hardware).

The financial result. IBM’s net income moved from a $8.1 billion loss in 1992 to a $4.2 billion profit by 1994, a swing of $12.3 billion inside two years. Services became the majority revenue contributor by the end of the decade. The stock compounded at approximately 14 percent annually from Gerstner’s arrival through his departure in 2002, against an S&P 500 that returned approximately 11 percent over the same window.

The mapping to Uber is precise. Khosrowshahi, like Gerstner, declined the breakup narrative. Activists and analysts during 2019 and 2020 argued that Uber should be split into mobility and delivery, with mobility returned to shareholders as a cash cow and delivery sold or spun off. Khosrowshahi refused. He argued, correctly, that the two businesses shared a customer acquisition funnel and that over one third of Uber Eats first-time users arrived through the mobility app. The cost action mirrors Gerstner’s. The 6,700 layoffs in 2020 were the Uber equivalent of Gerstner’s 100,000 cuts. The divestiture program (Russia, Southeast Asia, China, ATG, Eats India) is directly analogous to Gerstner’s non-core asset sales. And the construction of the advertising and membership businesses on top of mobility and delivery is the direct structural analog of Global Services on top of IBM hardware.

Operators should take from the Gerstner comparison one specific principle. A legitimate operating turnaround does not require breaking up the company. It often requires refusing to. The integrated platform is frequently the source of the turnaround, not the obstacle to it.

The full Operating Stories deep dive continues below.

Paid subscribers get the remaining two historical parallels (Satya Nadella at Microsoft, Howard Schultz at Starbucks), the four operating principles extracted as a directive playbook, the broader implications section, the conclusion, and

The AI Prompt Pack with sixteen prompts that walk you through applying the playbook to your own business inside a working session with Claude, ChatGPT, Gemini or Perplexity,

The Operating Playbook Worksheet

The 22-page Deep Dive Financial and Historical Report.

Subscribe to continue reading →

Subscribe now

Read more

New to Operating?

Check out these reader favorites:

• Operating Economics: The Optimization Company (Amazon)

• Operating by Warren & Charlie

• Operating Stories: From Snowboards to $220B — The Tobias Lütke & Shopify Story

Subscribe now

What is Operating Stories?

This series is built differently. It comes in two speeds.

The primary article, above the fold, gives you the essentials of the story being told, the charts, and the four operating principles in under five minutes.

But the article is just the center of a larger system designed to meet you wherever you are in your operating journey. If you run a company doing $50 million, the Operating Audit will score your business against the same playbook that built Nike, On Running, and New Balance, and the Prompt Pack will turn your favorite AI tool into a working session partner that produces a 90-day action plan specific to your business.

If you are a solopreneur selling your first product out of a metaphorical car trunk, the same frameworks scale down to your level because the principles are the same ones Phil Knight used when his total capital was $500 and his distribution strategy was a Plymouth Valiant.

And if you are the kind of operator who wants to see the raw research, the five Research Briefs give you every number, every source, and every data table so you can build your own analysis on top of mine, or just gain a better understanding of a fascinating piece of business history.

The goal is not to hand you a newsletter you read once. It is to hand you a system you use until the principles become embedded in your thinking and operating strategy.

For the first week’s release of this new design, I’m making the Operating Audit and Prompt Pack available for free, here and here.

- j -

Start Here

In 1964, Phil Knight sold 1,300 pairs of Japanese running shoes out of a Plymouth Valiant and grossed $8,000. His total startup capital was $500. He kept his day job for five years. By 1980, Nike commanded 50% of the US athletic shoe market with $269 million in revenue. Sixteen years to acquire and 50% of the market. That’s the first operating lesson.

The playbook was three moves.

1. Product obsession: Bill Bowerman cut shoes apart with a band saw and poured rubber into a waffle iron.

2. Grassroots distribution: Knight sold at track meets because that is where runners were.

3. Capital discipline: 3% of revenue on marketing in 1977. Every dollar earned its place.

Sixty years later, two companies are running that same playbook. Against Nike.

On Running started in 2010 when a retired Swiss triathlete glued garden hose segments to shoe outsoles. He pitched the idea to Nike first. Nike said no. On hit $3 billion in revenue in 2025 with a 62.8% gross margin. Roger Federer invested $54 million for 3% equity in 2019. That stake is now worth $400-500 million.

New Balance has been privately owned since Jim Davis bought it for $100,000 in 1972. He turned down a blank-check acquisition offer from Warren Buffett. The company reached $9.2 billion in revenue in 2025.

Its fifth consecutive year of double-digit growth.

Both challengers grew because Nike opened the door. Between 2017 and 2025, Nike made three strategic mistakes that map precisely to the three moves of its own founding playbook.

1. It stopped obsessing over product. Nike replaced sport-category teams with demographic silos. The innovation pipeline thinned. Nike disappeared from the top five in US performance running specialty.

2. It abandoned grassroots distribution. Nike cut its wholesale partners from 30,000 to 40. It dropped DSW, Macy’s, Urban Outfitters, Zappos, and dozens more. Challengers walked into the vacated shelf space.

3. It traded capital discipline for franchise extraction. Nike flooded the market with Air Force 1, Jordan 1, and Dunk colorways until the cultural equity collapsed.

Revenue fell from $51.3 billion to a $46 billion run-rate. The stock dropped 56% over five years. More than $200 billion in market capitalization erased.

The company that invented the Challenger playbook became the incumbent that the playbook was designed to beat.

Subscribe now

The Challenger Playbook 4 Operating Principles

1. Sell to the community that confers credibility before you sell to anyone else. Knight sold at track meets. On launched through REI exclusives and run clubs. New Balance partnered with Teddy Santis and Aimé Leon Dore. Find the smallest, most influential community that cares about what you are building. Serve them first.

2. Product obsession is a founder’s job, not a department’s job. Bowerman’s waffle iron. Bernhard’s garden hose. Davis’s domestic factory investment. The product advantage was created by a founder with domain expertise making decisions that a professional management team would not make.

3. Capital constraints are a competitive weapon, not a disadvantage. Knight spent 3%. On gave Federer equity instead of a $300 million endorsement deal. New Balance stayed private for 52 years. Every constraint forced a better decision.

4. Distribution is brand strategy expressed operationally. Every distribution decision is a brand decision. Choose channels that tell the story you want told. Track meets told the story that Nike was for serious runners. DSW never told a story anyone remembers.

Score your business against the Challenger Playbook. Take these three questions to ChatGPT, Claude, or Gemini with your own business context:

1. Name the last three product changes you made based on direct feedback from your most demanding customers. If you cannot, your Product Obsession score is low.

2. Rank your distribution channels by credibility with your core customer. If your highest-volume channel is your lowest-credibility channel, your Distribution Credibility score needs work.

3. What percentage of revenue do you spend on paid customer acquisition? Knight spent 3%. If yours is above 15%, your Capital Discipline is renting attention instead of building equity.

The full Operating Stories Deep Dive continues below for paid subscribers, including the complete Nike origin story, the On Running and New Balance operating profiles, running boom market analysis, the Challenger’s Operating Audit worksheet, a 16-prompt AI working session pack, and three branded research briefs you can use for your own strategic planning.

- john -

Subscribe now

Read more

Start Here

Private equity never wrote us a check.

They did something far more valuable.

They told me everything that was wrong with our business.

For free. Repeatedly.

Over several dozen meetings across a period of years.

This is the story of R.G. Brewton, Inc.

My family’s industrial distribution company that I helped optimize, grow and scale more than 3X from $25 million in revenue before we sold to a strategic acquirer in August of 2021. It is also the story of how a string of private equity rejections became the most expensive education I never had to pay for.

Hope this helps.

- j -

The Business

R.G. Brewton was an MRO distributor. We sold indirect materials into manufacturing plants. Cutting tools, abrasives, coolants and lubricants, welding supplies, safety products. The stuff that keeps a factory floor running but never shows up in the finished product.

We were not a supply house taking phone orders. We installed and managed industrial vending and RFID systems inside our clients’ facilities. We controlled their indirect supply chains and managed their inventory. We we expert at creating documented cost savings programs. Every engagement required quantifiable proof (to the penny) of the spend we were reducing relative to production run rates year-over-year.

The model worked. Clients loved it.

But it had structural problems that would define the next decade of my career.

The Margin Trap

Beginning gross margins on new client engagement ran between 19% and 25%. Healthy enough. But as relationships grew and three-year agreements stacked renewal over renewal, margins compressed. The better we performed for a client, the more leverage they had at the negotiating table.

We were rewarded for excellence with tighter economics.

On top of that, we competed exclusively against companies five, ten or even twenty times our size (think Grainger and Fastenal). They were global players with purchasing power and brand recognition we could not match. Their presence put constant downward pressure on pricing.

Additionally, the business was capital intensive. Clients did not buy the vending hardware. We owned every machine. We financed every installation. That meant we needed a minimum of $750,000 to $1 million in projected spend just to justify engaging with a new customer. Add the inventory carrying costs for a business of that scale and you had a model that easily grew revenue faster than it grew earnings.

The PE Education

Between 2010 and 2012, private equity firms started calling. The thesis was straightforward:

• Industrial distribution was fragmented.

• Roll up regional players, build scale, compete with the globals.

• We fit the profile on paper.

But every conversation ended the same way:

• Our margins were too thin.

• Our sales cycle too long.

• We were too committed to the integrated supply model.

• Not enough diversification across a base of smaller, traditional buyers.

And the embedded labor component, full-time employees placed inside client facilities, made the cost structure heavier than PE wanted to underwrite with leverage.

I knew after the first meeting or two that PE was not going to invest, ever. But I took every meeting anyway. There were at least a dozen firms and family offices over the years that reached out to me.

Each one was a masterclass in what a sophisticated financial buyer sees when they open your books. Each one sharpened my understanding of the gap between where we were and where we needed to be. It also gave me an extraordinary, first-person masterclass in debt structuring and restructuring strategy. Super interesting stuff.

I used every meeting to build relationships. These were people who would eventually be doing deals in my space. If I could not be their portfolio company, I could be in their network. I could find ways to help them. They were also buying manufacturers who could become clients.

Side Note

Most people are cosplaying with technology they’ve opened twice.

84% of humans have never used AI.

0.3% pay for it. Actual adoption is a fraction of what the tools can do. And yet the feed is full of “AI-native founders” who couldn’t build an automated workflow if their revenue depended on it.

The gap between what the technology can do and what people actually do with it is the biggest business opportunity of the next decade.

That’s exactly what we work on inside the Operating Founder program:

• 4 1:1 (60 Minute) Strategic Operating Sessions w/ John

• Weekly Cohort Zoom Working Sessions

• Access to a private Skool community filled with all of the tools and frameworks we’re building along the way.

That’s less than two dollars a week to be in the room with operators who are quietly building the most efficient businesses on the internet while everyone else is arguing about which chatbot is better.

We’re at 187 members.

At 200, the program is closed.

The people inside aren’t posting about AI.

They’re building their futures with it.

Join the Operating Founder Cohort

The PE Playbook

The PE feedback gave me a clear operating thesis:

Revenue growth alone would not get us to an exit.

We had to expand EBITDA without the luxury of raising prices or cutting client-facing headcount.

Clients saw our on-site teams as the core of the value they were paying for. That was untouchable. So we went inside.

• We consolidated purchasing, sourcing, and inside sales into two streamlined groups.

• We aggressively automated billing, receivables, and payables management.

• We optimized our ERP systems and built a data warehouse that could house all of our ERP data alongside client inventory data. We also built internal applications to improve the integrity of all new data being entered into our systems.

• Then we layered analytics on top of the whole system to automate purchasing decisions, and client inventory and cost savings analysis.

The other critical move was people.

We used industrial and organizational psychology testing to understand the strongest skill sets on our team. Then we got everyone into the right seat on the bus (thanks Jim Collins).

But executing all of this inside a family business going through a transformation is something else entirely. Promoting from within and placing people where they could do their best work created stability. That stability is what made the next part possible.

We reduced headcount from 102 to 61.

Revenue more than tripled during the same period.

We were winning.

The Exit

Just before COVID, we connected with a strategic acquirer, as opposed to a financial buyer looking for a return on leverage. The buyer wanted what we had built:

• Our expertise in industrial vending.

• Our geographic footprint.

• Our customer relationships.

• And yes, the operational efficiencies we had spent years engineering.

COVID paused the conversation. They came back. Nine months of due diligence and negotiations later, the deal closed in August of 2021.

The efficiencies were the essential financial factor.

Without the EBITDA transformation, there is no deal.

Without a dozen PE meetings telling me exactly where the gaps were, there was no transformation and no deal.

Five Lessons for Operators & Builders

1. Every “no” is a free diagnostic: PE firms spend millions on due diligence. When they pass on your business, they are handing you a detailed assessment of your weaknesses at no cost. Most founders take the rejection personally and move on. Take it as a consulting engagement you never had to pay for.

2. Revenue growth without margin expansion is a trap: We more than doubled revenue and the business still was not sellable until the EBITDA profile changed. Acquirers buy earnings. Not revenue. Growth is seductive. Do not let it distract you from profitability.

3. Automate the back office. Protect the front line: Know which costs are load-bearing and which are redundant. We could not cut client-facing headcount because that was the value proposition. So we consolidated internally, built data infrastructure, and automated everything behind the scenes.

4. Put people in the right seat before you reduce headcount: IO psychology and talent assessment gave us the map. Getting the right people in the right roles and promoting from within created the organizational stability required to absorb a 40 percent reduction in headcount without losing institutional knowledge or morale. Sequencing matters. Optimize talent, then resize.

5. Know whether you are building for a financial buyer or a strategic one: PE could not buy us because of our margin structure and capital intensity. They couldn’t find a fast enough path to getting their money out. And then by the time a path could be discerned, the strategic deal was more attractive to us as the seller. A strategic acquirer wanted our expertise, geography, and customers. Understanding which buyer type your business actually serves changes every operational decision you make and most every element of how you negotiate your deal. Build accordingly.

Hope this one helps. Would love to hear more about the pieces you found most helpful, so I can create more work telling my personal business building stories.

- john -

Are you new to Operating?

You Might Enjoy These

(Access the entire library of articles and resources for $7.99 per month)

• 0 to 55,000 - The First 90 Days Playbook - What I’d Do Differently Knowing What I Know Now

• The Operating by John Brewton Resource Library - No Filler. All Killer.

• WINNING THE LOSER’S GAME: The Creator Economy and the Companies We All Need to Build

• Operating Stories: Hermès and the Manufacturing of Scarcity

• Operating Economics: Building Antifragile Companies

• How Three Decades Rewrote the Economics of Companies

Become a paid subscriber to access all the content ($7.99 / month) or get to work with us and become a member of the Operating Founder cohort for $99 (annual).

Subscribe now

John Brewton documents the history and future of operating companies at Operating by John Brewton. He is a graduate of Harvard University and began his career as a Phd. student in economics at the University of Chicago. After selling his family’s B2B industrial distribution company in 2021, he has been helping business owners, founders and investors optimize their operations ever since. He is the founder of 6A East Partners, a research and advisory firm asking the question: What is the future of companies? He still cringes at his early LinkedIn posts and loves making content each and everyday, despite the protestations of his beloved wife, Fabiola, at times.

She did everything right.

Best public school in the county.

Four years at a university her parents still mention at dinner parties.

An MBA she paid off in nine years.

She climbed the ladder in the exact sequence the system promised: analyst, associate, manager, senior manager, director.

Director of Strategic Operations. $242,000 a year base.

A standing desk. A title strong enough, and brand name respected enough to leave people talking in every room she entered and exited.

She managed eleven people.

Those eleven collectively managed seventy-three.

Her team produced strategy decks, operational reviews, cross-functional alignment documents, and quarterly business updates that were forwarded to people who forwarded them to other people who never opened them.

The work felt important in the way that all corporate work feels important, it had deadlines, it had stakeholders, it had a Slack channel with 74 members.

That was 2023. The vibes were fine.

Her CEO stood on a stage at an industry conference that year and said the words that every CEO said in 2023: “AI is a tool, not a threat.”

Everyone clapped.

He believed it.

She believed it.

The analysts in the back row, who would be cut first, believed it too.

It did not happen all at once…

It happened the way debt accumulates and margins decays…slowly, then everywhere.

Wall Street, which had spent multiple lifetimes turning public companies into financial instruments, optimized for quarterly earnings, discovered something better than offshoring and better than automation: a technology that could do the work of a knowledge worker for the cost of a software subscription. A $60-a-month subscription.

The feeling that we had been doing the work computers simply couldn’t do, yet, began to appear quite prescient.

The work didn’t have to be better. It had to be good enough.

When a CFO can replace a team of eleven with three AI agents and one prompt engineer named Tyler who is 24 and makes $68,000, the decision is easy to execute.

It is, after all, their fiduciary duty.

The stock went up eleven percent the quarter her department was “restructured.” The CEO posted about it on LinkedIn about “operational efficiency” and his new love affair with the word “lean” and all things 1990s Toyota.

Lean meant that real people with real mortgages were replaced by software that hallucinates sometimes but never asks for healthcare.

Nobel laureate Daron Acemoglu calls this the “anti-worker” path of AI development, technology designed not to extend human capability but to replace it, because replacing humans is how the builders of these models intend to monetize.

GDP rises.

Wages stagnate.

Corporate profits hit records while the labor market quietly collapses underneath.

Economists have a term for this. During the British Industrial Revolution, it took fifty years of rising output and falling wages before workers saw any benefit. They called it the Engels’ Pause.

We are living through a compressed version, the same divergence between capital and labor, but squeezed into a single decade and aimed not at factory workers but at the professional class that believed education was armor.

We were wrong.

(And oh my are things about to get messy)

She applied for seventy-four jobs in one year.

The applicant tracking systems parsed her twelve years of cross-functional leadership experience as “entry level.”

A recruiter asked if she was “open to contract work,” which is a phrase that means no benefits, no security, and no guarantee that next month exists.

She watched the World Economic Forum report land like a polite bomb: ninety-two million jobs displaced by 2030.

78 million new jobs to be created for skills she had not yet acquired.

She searched for these new roles on LinkedIn.

They were not there.

What was there: a promoted post from the company that laid her off, advertising a “Director of AI Transformation” role that was, functionally, her old job described in language she no longer qualified to use.

Ford’s CEO said AI would replace “literally half of all white-collar workers.”

JPMorgan told its managers to stop hiring.

Salesforce claimed AI was already handling fifty percent of its workload.

These were confessional, not predictive.

The companies were saying, out loud, on earnings calls, that they were done pretending a multitude of labor types had value they intended to pay for.

She attended a “reskilling webinar” hosted by the same company that had let her go. The webinar was forty-five minutes. It was AI-generated. She could tell because it used the phrase “unlock your potential” three times and never once acknowledged that her potential had been unlocked for twelve years before becoming obsolete.

Are you new to Operating?

You Might Enjoy These

(Access the entire library of articles and resources for $7.99 per month)

• 0 to 55,000 - The First 90 Days Playbook - What I’d Do Differently Knowing What I Know Now

• The Operating by John Brewton Resource Library - No Filler. All Killer.

• WINNING THE LOSER’S GAME: The Creator Economy and the Companies We All Need to Build

• Operating Stories: Hermès and the Manufacturing of Scarcity

• Operating Economics: Building Antifragile Companies

• How Three Decades Rewrote the Economics of Companies

Become a paid subscriber to access all the content ($7.99 / month) or get to work with us and become a member of the Operating Founder cohort for $99 (annual).

Subscribe now

The same financial system that turned companies into instruments, Wall Street’s relentless optimization of headcount, its worship of margins, its decades-long campaign to convert every human being on a payroll into a line item to be reduced, was quietly, almost accidentally, building the infrastructure for individuals to become their own economies.

Not companies in the old sense. Not LLCs with org charts and office leases. Something stranger. Something the system did not intend to create.

While Wall Street was busy making public companies inhospitable to humans, payment platforms like Stripe made it possible for a single person to accept money from anyone on earth in seconds.

Fractional equity platforms let individuals invest in assets that were once gated behind six-figure minimums.

Crypto, for all its chaos, proved that value could be stored, transferred, and programmed without a bank’s permission.

Social platforms gave a single voice the same distribution that used to require a media conglomerate. A phone camera replaced a production studio. An email list replaced a sales team. A Substack replaced a publishing house.

The rails were built. The same forces that made corporations hostile to workers made it structurally possible for workers to leave.

She was sitting at her kitchen table in month seven of unemployment when it happened. Not a revelation. Not a movie moment.

More like the moment your eyes adjust in a dark room and you realize you can see.

She was staring at the same tools that had replaced her, the AI writing assistant, the design generator, the analytics dashboard, the scheduling software, the video editor, and the arithmetic landed:

These tools did not just make her replaceable to a corporation.

They made the corporation replaceable to her.

She did not need eleven people. She did not need a VP’s approval or a $340,000 vendor platform or a cross-functional stakeholder review. She needed Substack, Stripe, a phone camera, and the accumulated expertise of a twelve-year career that no large language model actually possesses.

Expertise is judgment. It is the ability to walk into a room and know which problem is real and which one is a decoy.

No algorithm has that. Not yet. Maybe not ever. At least not for now.

Acemoglu draws the distinction clearly: AI that replaces humans versus AI that extends human capability.

The first path leads to displacement.

The second gives way to leverage, to opportunity.

She chose the second.

Because the alternative was another reskilling webinar.

Fourteen months later, she has 14,000 newsletter subscribers.

She makes more money than she ever made before.

She built a brand on the exact intersection of operations, strategy, and industry knowledge that no hiring panel could categorize and no applicant tracking system could parse.

The same breadth that made her “confusing” to corporate recruiters made her magnetic to an audience that was tired of specialists who understood one silo and generalists who understood none.

She accepts payments from nineteen countries via Stripe.

She films on her phone.

She edits with AI.

She publishes on platforms that did not exist five years ago to an audience that finds her through algorithms she does not control but also does not need permission to use.

She owns her distribution.

She owns her revenue.

She owns her time.

Nobody can restructure her. Nobody can lean her out. Nobody can send her a calendar invite with no subject line and HR on the copy.

The system that was supposed to employ her decided she was a cost to be optimized. So she optimized herself. She took the tools that eliminated her position and used them to eliminate her need for one.

She is a company of one, built on her own values, her own expertise, her own judgment, and her own voice.

She did not plan this. The economy did not plan this.

But here she is.

And here, increasingly, is everyone.

The old promise was simple: go to the best school, get the best job, work at the best company, and the institution will keep you safe.

That promise is broken.

Not bending.

Not wavering.

Broken.

The institution does not keep you safe. The institution keeps itself safe. You are a variable in someone else’s optimization function. You always were, and that’s okay.

The computers just couldn’t do the work, yet.

The fluorescent lights were on and the direct deposit hit every two weeks and the 401(k) match felt like love.

But the technology that broke the promise also broke the gate.

The same AI that made her disposable to a corporation made the corporation disposable to her. The same financial infrastructure that Wall Street built to extract value from companies is now available to individuals who want to build it.

The same internet that let a CEO post about “operational efficiency” on LinkedIn is the same internet that lets a woman with twelve years of expertise and a kitchen table reach more people in a week than her old company reached in a quarter.

The last employee is not a tragedy.

The last employee is the opportunity.

The last employee is the first founder.

And the tools have never been cheaper, the audience has never been closer, the capacity to scale has never been more within reach.

The last employee is the first founder.

We are all the first founder.

Companies are becoming tech stacks.

We are all becoming companies.

- j -

If you’re a first founder, I’d love for you come build with our Operating community.

Become and Operating Founder and for $99 get four 1:1 (60 minute) sessions with me to start building your founder strategy.

Become an Operating Founder

John Brewton documents the history and future of operating companies at Operating by John Brewton. He is a graduate of Harvard University and began his career as a Phd. student in economics at the University of Chicago. After selling his family’s B2B industrial distribution company in 2021, he has been helping business owners, founders and investors optimize their operations ever since. He is the founder of 6A East Partners, a research and advisory firm asking the question: What is the future of companies? He still cringes at his early LinkedIn posts and loves making content each and everyday, despite the protestations of his beloved wife, Fabiola, at times.

Operating Stories: A weekly series dissecting the strategies of iconic companies to help us build our brands and companies

Read more

Operating Stories Series

Read more

Read more

Subscribe now

Operating by John Brewton | McLean, VA | 2.8.26

Hermès is the clearest modern case of a company that turned its operations into its brand. So you can probably imagine the affinity I have for their work and products.

It manufactures scarcity, not through marketing tricks, but through single-artisan production, controlled capacity expansion, vertical integration, and a relationship-driven allocation system. Then it prices into the demand it creates.

If you want a working definition of luxury strategy, here it is: protect the process, protect the distribution, and let the waitlist do the advertising.

I hope you enjoy this week’s edition of Operating Stories. I have a sincere place in my hear for the company this week.

- john -

What Is Hermès’ Business Model?

Hermès is a craftsmanship-first, vertically integrated luxury house. It sells primarily through its own boutiques, keeps production deliberately below demand, and uses scarcity to maintain pricing power and secondary-market prestige.

The numbers speak for themselves. In 2024, Hermès generated €15.2 billion in revenue with €4.6 billion in net income and a recurring operating margin around 40.5%. Not too shabby. What is most extraordinary is that it achieves these numbers while actively refuseing to scale in many of the manners their competitor have chosen.

Read more

In 2026, I’m working directly with 100 creators building real businesses.

I want to bring operating strategy, competitive positioning, and financial planning to a community that’s fundamentally different from my typical industrial and technology clients.

For the first 100 creator founders: Four 60-minute 1:1 advisory sessions for $95.

Context: My standard engagement starts at $10K/month. This isn’t that. This is me learning from you while helping you build something sustainable.

Limited to 100 spots. Message me with questions or to claim yours or just click below to sign up (sign up requires browser access):

Become An Operating Founder

Start Here

In 2004, Tobias Lütke wanted to sell snowboards online. He had no credentials, no capital, no connections, and no work authorization in the country where he wanted to start a business. What he did have: two months of uninterrupted time and the ability to code in Ruby on Rails.

By 2006, other entrepreneurs didn’t want his snowboards, they wanted his software. He pivoted. That software became Shopify. Today it powers over 9 million businesses across 175 countries, has facilitated over $1 trillion in commerce, and is valued at approximately $220 billion.

This is not a story about luck or timing. It’s a story about how capital markets force strategic discipline on companies that forget their core competence. Between May 2015 and today, Shopify has filed 42 quarterly earnings reports. Those reports trace three distinct strategic regimes:

1. 2015–2019: Product-led land-grab (clean software story)

2. 2020–2022: Pandemic super-cycle followed by logistics over-reach

3. 2023–2025: Strategic refocus on software and profitable growth

The stock appreciated 4,600% from IPO to May 2020, crashed 70% in 2022, then recovered 200% from its October 2022 low. The company acquired logistics infrastructure for $2.1 billion in May 2022, then sold it less than one year later while laying off 20% of its workforce.

What you’re about to read is not a typical earnings recap. It’s an autopsy of how quarterly communications shape—and are shaped by—public market expectations. And it’s a field manual for operators who need to understand how the companies of the future are built, priced, and disciplined.

Let’s start where Lütke started: with what you already have.

Subscribe now

Principle #1: Start With What You Have

Section 1.1: The Capital You Don’t Count

Most people think starting capital means money. Lütke’s story suggests otherwise. When he built Snowdevil (his snowboard shop) in 2004, his balance sheet looked empty by conventional standards. But his capability inventory was rich:

• Technical capital: Proficiency in Ruby on Rails at a moment when the framework was nascent

• Time capital: Approximately two months of uninterrupted focus

• Problem capital: Direct experience with friction that made existing e-commerce platforms unusable for his use case

• Distribution capital: Willingness to share an imperfect solution publicly

The asymmetry is instructive. His limitations were real and significant—no venture funding, no co-founder with retail experience, work permit issues constraining employment options, no existing network in the technology industry, no MBA or formal business training.

Yet the advantage that mattered most—technical skill applied to a genuine problem he was motivated to solve—was something he already possessed before deciding to start a company.

Three distinct strategic regimes: hyper-growth (2015-2019), pandemic super-cycle (2020-2022), and disciplined profitable growth (2023-2025)

Subscribe now

The chart above shows Shopify’s revenue from IPO through today. Notice three distinct regimes:

• Era 1 (2015-2019): Growth rates of 47–95% annually on an expanding base

• Era 2 (2020-2022): Pandemic spike (86% growth in 2020), then deceleration to 21% in 2022

• Era 3 (2023-2025): Stabilized at 24–26% growth with dramatically improved profitability

Operating Takeaway: Redefine Your Starting Capital

Most operators systematically undercount their starting capital. They inventory what they lack (funding, network, credentials) and ignore what they possess (domain expertise, technical skills, time, specific knowledge of a friction point).

A more useful framework:

• List skills you already have rather than resources you lack

• Writing ability and subject expertise are real capital

• Design skills are real capital

• Specific knowledge of niche communities is real capital

• Time you control is real capital

The error lies in discounting these because they don’t appear on balance sheets. But Lütke’s story, and the 4,600% stock appreciation that followed, demonstrates that the most durable competitive advantages often come from deploying existing capabilities against problems you understand deeply.

The modern environment amplifies this advantage. In 2004, Lütke needed to build infrastructure from scratch. Today, platforms that didn’t exist then—Shopify itself, Substack, Gumroad, YouTube, Webflow—provide scaffolding for enterprises. Financial barriers to starting have declined. Ability to reach customers directly has improved. Cost of learning new skills has fallen.

The question isn’t whether you possess sufficient advantages. The question is whether you’re willing to start with what you have.

• The Sorting Year: What We Learned From 2025 & Where We’re Going in 2026

• Operating Leaders: How JP Morgan Chase Showed the Rest of Us What AI Transformations Can Achieve (Part One)

• Operating Economics: The Optimization Company

• Operating Economics: Building Antifragile Companies

• Start With The Work, Not The Why

• Sorry Is Not A Solution

• How Three Decades Rewrote the Economics of Companies

• We Are All Becoming Operators

Become An Operating Founder

Principle #2: Listen to Markets

Section 1.2: The Pivot That Markets Demanded

Lütke’s decision to pivot from retail (Snowdevil) to infrastructure (Shopify) in 2006 wasn’t about visionary foresight, it was about pattern recognition. I think it was about humility. He observed that his tool generated more demand than the snowboard business itself. Other entrepreneurs wanted his software. The market was redirecting him.

This moment represents a fork that determines many ventures’ trajectories. One path involves defending the original plan, insisting that retail was the “real” business and software was merely supporting infrastructure. The other path involves following market signals: recognizing the software was more valuable than retail operations and reorienting accordingly.

The market speaks through price, but also through less obvious signals. For Shopify in 2006, those signals appeared as qualitative observations:

• The question customers asked most frequently (not about snowboards, but about the software)

• The feature they used most consistently (the platform, not the product catalog)

• The problem they were most visibly struggling to solve (building online stores)

Similar examples appear across successful pivots:

• YouTube began as a dating platform; users uploaded non-dating videos

  • → pivoted to general video hosting

• Instagram originated as a check-in application (Burbn); users primarily shared photos

  • → stripped to photo-sharing only

• Slack emerged from a gaming company’s internal tool; outside teams wanted to use it

  • → spun out as standalone product

• Twitter was built on failed podcasting platform infrastructure (Odeo); short-status updates were the breakout feature

  • → became the core product

Operating Takeaway: Treat Demand Curves as Information Events

When market signals diverge from your original plan, that’s an information event—not a distraction. The instinct to defend the original vision is strong, especially when you’ve invested time explaining it to others. But markets reveal preferences through behavior, not stated intentions.

Your job as an operator is to remain attentive to:

• Usage patterns that differ from design intent: If customers use your product for an unintended purpose more frequently than the intended one, you’ve discovered a new problem to solve

• Questions that cluster around unexpected features: The question customers ask most often reveals their primary friction point

• Willingness to pay for byproducts: If a component you considered secondary generates more commercial interest than your core offering, the market is sending a signal about value

The companies that scale are often those that recognized these signals early and pivoted before capital constraints forced the issue. Lütke could have insisted that Snowdevil was the “real” business. Instead, he followed the signal. By Q4 2015 (first full quarter post-IPO), Shopify was already processing $2.8 billion in GMV and growing revenue at 99% year-over-year.

The market told him what it wanted. He listened.

Principle #3: Solve the Problem You Actually Experience"

Section 2.1: When Pull-Forward Demand Looks Like Permanent TAM Expansion

Between Q2 2020 and Q4 2021, Shopify experienced the most dramatic demand acceleration in its history. Q2 2020 revenue grew 97% year-over-year. GMV surged 119%. New stores created jumped 71% quarter-over-quarter. The stock reached an all-time high of approximately $350 (split-adjusted) in November 2021.

Management called it a “once in a generation” demand shift. Offline merchants, forced online by pandemic lockdowns, scrambled to build digital storefronts. Shopify was the obvious choice. The company introduced a 90-day free trial (March 21 – May 31, 2020) to capture the merchant influx, and merchants replaced 94% of lost point-of-sale GMV with online sales.

The pandemic pull-forward is unmistakable: from 119% growth in Q2 2020 to just 11% by Q2 2022

Read more

Check out these reader favorites:

• Operating Economics: The Optimization Company

• Operating Economics: Building Antifragile Companies

• Start With The Work, Not The Why

• Sorry Is Not A Solution

• How Three Decades Rewrote the Economics of Companies

• We Are All Becoming Operators

Subscribe now

Buffett and Munger did not build Berkshire by “finding cheap stocks.”

They built it by marrying capital to operators who could run exceptional businesses for decades with almost no interference.

That is the real meaning of “our favorite holding period is forever.”

How Berkshire Actually Used “Forever”

When Buffett talks about forever, he is referring to a specific structure: find a business with durable economics, pair it with an exceptional manager, and then give them the time, autonomy, and capital they need. The compounding comes from the operator–business pairing, not from clever trading.​

This philosophy is codified in Berkshire’s Owner’s Manual, a document Buffett first wrote in 1996 and has updated periodically since. The manual lays out 15 principles, but the core thesis is simple: Berkshire is a collection of businesses run by autonomous operators who think like owners, supported by a parent company that provides capital and stays out of their way.​

The practical application of this philosophy shows up in Berkshire’s Acquisition Criteria, which has remained largely unchanged for decades. The criteria explicitly state: “We need management in place (we can’t supply it).” Berkshire does not buy turnarounds. It does not install heroic new CEOs. It buys businesses where the operator is already exceptional and wants to keep running the company.​

Four acquisitions show the pattern clearly: See’s Candies, GEICO, Nebraska Furniture Mart and BNSF Railroad. In each case, the long holding period was earned by management behavior and operating results, not granted in advance as a virtue signal.

Compounding on a Small Base

Berkshire bought See’s in 1972 for approximately $25 million, when it was generating roughly $4 million in pretax profit on about $8 million of tangible capital. That looked expensive on traditional value screens. What changed Buffett’s mind was the realization that See’s had genuine pricing power: customers would accept price increases well above inflation without switching because the product and brand were embedded in habit and emotion.​

The key to the acquisition was Chuck Huggins, a 20-year See’s veteran who became president. In a 1972 letter to Huggins that has since become legendary among Berkshire students, Buffett laid out his operating philosophy in plain language.​

The instructions were remarkably simple: never sacrifice quality for profit, keep production local, protect the brand’s reputation. No growth hacks. No rapid scaling. No pressure to hit quarterly targets. Just protect the moat and let pricing power compound.

Buffett’s letter included a revealing comparison to Coors beer: “I always had the suspicion that 99% was in the telling and 1% was in the drinking.” This was a remarkably honest admission that brand storytelling—not just product quality—creates pricing power. Huggins understood this intuitively and spent the next three decades protecting See’s positioning while methodically raising prices.​

The results over 35 years:

• $25 million invested → $2 billion+ in cumulative pretax earnings sent to Berkshire

• Only $32 million of additional capital reinvested

• Price per pound: $1.95 → $11.60+ (roughly 7-8% annual increases, far above inflation)

• Revenue: $30 million → $430 million

• Effective returns: 8,000%+ on original investment​

“See’s gave birth to multiple new streams of cash for us,” Buffett later wrote. The cash from See’s funded acquisitions of other businesses, which generated more cash, funded more acquisitions. The compounding was not just within See’s—it was across Berkshire’s entire portfolio.

What made it work was not a turnaround or a brilliant strategic pivot. It was a disciplined exploitation of an existing moat by an operator who understood the business deeply and had no desire to “grow” it into mediocrity. Volume grew about 4% annually. Prices grew at about an 11% annual rate, representing slow growth but fast compounding.​

Huggins ran See’s until 2005, over three decades of continuous operation under the same philosophy. Buffett called See’s his “prototype of a dream business” and credited it with teaching him that paying up for quality, when paired with exceptional management—was superior to buying mediocre businesses at a low price.​

Berkshire completed its purchase of GEICO in 1995, paying approximately $2.3 billion for the half of the company it did not already own. At that point, GEICO had a small, single-digit share of the U.S. auto market but a clear structural edge: a direct-to-consumer, low-cost model that allowed it to spend more on marketing while still underpricing many of its competitors.​

GEICO: Scaling a Cost Advantage

Berkshire’s relationship with GEICO began in 1951 when Buffett, then a 20-year-old Columbia student, visited the company’s headquarters on a Saturday and spent four hours learning about the business from Lorimer Davidson. Buffett invested a significant portion of his net worth and later sold for a profit, but GEICO stayed lodged in his memory as an example of structural competitive advantage.​

In 1976, GEICO nearly went bankrupt. Buffett began repurchasing shares, eventually accumulating a 50% stake by 1980. In 1995, Berkshire paid $2.3 billion to acquire the remaining 49% it did not already own.​

At that point, GEICO was the seventh-largest auto insurer in the United States with roughly 2.5% market share. The business had a clear structural advantage: a direct-to-consumer model that eliminated agent commissions, allowing it to undercut competitors on price while still earning underwriting profits.​

The operator who would scale this advantage was Tony Nicely, who became CEO in 1993. Nicely had joined GEICO in 1961 as a 19-year-old and worked his way up through the company over the course of three decades. He understood the cost structure intimately and had a clear thesis: GEICO’s advantage would compound with scale, because fixed costs could be spread across more policies. At the same time, the direct model kept variable costs low.​

The key passage from Buffett’s 1995 Shareholder Letter explained the strategic logic: “In business, I look for economic castles protected by unbreachable moats.” GEICO’s moat was its cost structure. Nicely’s job was to widen it.​

The results over Tony Nicely’s 25-year tenure:

• Market share: 2.5% → 14%+ (now second-largest auto insurer in U.S.)

• Policies: ~2 million → 18 million+

• Premiums earned: grew at 9.1% compound annual rate

• Combined ratio: 92.2% average (vs. 100%+ industry average)

• Cumulative underwriting profit: $9.7 billion+

• Float: $3 billion → $12.5 billion+​

The combined ratio is the key operating metric. A combined ratio below 100% means the company is earning an underwriting profit; it collects more in premiums than it pays out in claims and expenses. Most auto insurers hover around 100% or above, meaning they break even on underwriting and make money by investing the float. GEICO consistently earned underwriting profits, allowing it to invest the float.​

Buffett’s 2018 assessment of Nicely’s tenure was unequivocal: “Tony’s management of GEICO has increased Berkshire’s intrinsic value by more than $50 billion.”​

The operating edge came from relentless focus on the direct model. GEICO’s expense ratio remained around 18%, while the industry average was 25%. That 7-point advantage funded both aggressive advertising (GEICO spent over $1 billion annually on marketing by the 2010s) and lower prices for customers. Scale begat more scale.​

Nicely retired in 2018 after 57 years at the company. Competitors have since copied the direct-to-consumer model he scaled, but GEICO’s head start and accumulated brand awareness have proven difficult to replicate.

The Handshake Deal

In 1983, Buffett bought 90% of Nebraska Furniture Mart for $60 million. The acquisition contract was 1.25 pages long.​

There was no audit. No verification of receivables. No inventory check. Buffett did not even visit the store before making the offer. He trusted the word of the 89-year-old woman who ran it.​

Rose Blumkin—”Mrs. B”—had emigrated from Russia in 1917, speaking no English. She started Nebraska Furniture Mart in 1937 with $500 in savings and built it into the largest furniture store in North America by following a simple formula: buy in bulk, operate on thin margins, never lie to customers.​

Her credo, repeated constantly to employees and customers alike: “Sell cheap, tell the truth, don’t cheat nobody.”​

Buffett’s 1983 Shareholder Letter devoted considerable space to explaining why he paid a premium for NFM without conducting due diligence. The answer was Mrs. B herself. Buffett had watched her operate for years, seen her competitive behavior up close (she had repeatedly beaten competitors who tried to undercut her), and concluded that her character was the due diligence.​

The economics of NFM were driven by purchasing power and fixed-cost leverage. The Omaha store did more volume than any furniture retailer in the country, which meant suppliers gave preferential pricing. The high volume also meant fixed costs (rent, management, systems) were spread across more revenue. Competitors with smaller stores could not match NFM’s prices without losing money.​

The results over four decades:

• Original 90% stake: $60 million

• Current structure: stores in Omaha, Kansas City, and Dallas/Fort Worth

• Omaha and Kansas City stores: each does roughly $450 million in annual revenue

• Dallas store: did $500 million+ in its first partial year (opened 2015)

• Total estimated revenue: $1.2-1.6 billion annually

• Still family-managed under the Blumkin family​

Mrs. B kept working 12-14 hours a day, seven days a week, until she was 103 years old. She briefly “retired” at 95 and opened a competing store across the street from NFM, forcing Berkshire to buy her out a second time. Her relentlessness was both the source of NFM’s competitive advantage and the reason Buffett wanted to own it.​

The lesson from Nebraska Furniture Mart is that management assessment is not primarily about analyzing org charts or incentive structures. It is about character. If you need extensive due diligence to trust the operator, you are probably buying the wrong business.

Patient Capital in Heavy Industry

Berkshire’s 2010 acquisition of BNSF Railway looked like the opposite of See’s Candies: asset-heavy, cyclical, capital-intensive, and heavily regulated. The price—approximately $44 billion including debt—was Berkshire’s largest acquisition ever.​

The attraction was not a quick efficiency fix. It was the chance to own a strategic infrastructure asset run by a management team already executing well and willing to reinvest at scale.

Buffett called it “an all-in wager on the economic future of the United States.” Railroads move coal, grain, consumer goods, and intermodal containers more efficiently than trucks over long distances. If the U.S. economy grows, rail volumes grow. If fuel costs rise, rail’s efficiency advantage over trucking increases.​

Matt Rose, who had been CEO since 2000, stayed on after the acquisition. In a 2015 interview, Rose described the relationship with Buffett: “Warren is my boss, and it’s a happy marriage. He lets us run the railroad.”​

The Berkshire ownership model for BNSF has been consistent: provide patient capital for long-payback infrastructure investments that public market investors might pressure management to defer. Since the acquisition, BNSF has invested approximately $3-4 billion annually in track, rolling stock, and capacity upgrades.​

From Berkshire’s 2009 Shareholder Letter, written just after announcing the deal: “Our country’s future prosperity depends on its having an efficient and well-maintained rail system... Berkshire is making a large and long-term commitment to BNSF.”​

The results over 15 years:

• Annual capital investment: $3.8 billion planned for 2025 alone​

• Expansion investment over past 5 years: $2.6 billion+

• Q2 2025 pretax earnings: $2 billion (up 11.5% year-over-year)​

• Cumulative operating cash since acquisition: $90 billion+​

• Current estimated value: significantly above original purchase price

The “improvement” at BNSF is less about fixing a broken operator and more about combining a capable management team with an owner willing to fund investments others might avoid. Public market railroads face pressure to return capital to shareholders. BNSF, under Berkshire, can prioritize network capacity and service quality over quarterly earnings.

Rose retired in 2020 after two decades as CEO, having led BNSF through its transition from public company to Berkshire subsidiary. The operating philosophy remained unchanged: invest in the network, maintain service quality, think in decades rather than quarters.

The “improvement” here is less about fixing a broken operator and more about combining a capable management team with an owner willing to fund long-term, high-return investments that others might avoid.​

The Pattern: What All Four Share

Across candy, insurance, furniture retail, and railroads, Buffett and Munger did the same thing:

Insisted on proven management before buying. Berkshire’s acquisition criteria explicitly state: “We need management in place (we can’t supply it).” In every case—Huggins at See’s, Nicely at GEICO, Mrs. B at NFM, Rose at BNSF—the operator was already running the business successfully before Berkshire arrived. There were no turnaround projects, no imported executives, no 100-day plans.​

Selected for character, not just competence. The 1.25-page NFM contract and the absence of due diligence illustrate the point: if you trust the operator’s character, you do not need elaborate protections. If you do not trust their character, no contract will save you.​

Delegated almost completely. Buffett’s description of his management style: “We delegate almost to the point of abdication.” The only things subsidiary managers must clear with Omaha: changes to post-retirement benefits, significant acquisitions, and unusually large capital expenditures. Everything else is autonomous.​

Protected reputation above profits. Every two years for more than 25 years, Buffett sent a one-page memo to all Berkshire managers with the same message: “The top priority—trumping everything else, including profits—is that all of us continue to zealously guard Berkshire’s reputation. We can afford to lose money—even a lot of money. But we can’t afford to lose reputation—even a shred of reputation.”​

Held for decades. See’s has been owned for 53 years. GEICO for 30 years since full acquisition (49 years since Buffett resumed buying). Nebraska Furniture Mart for 42 years. BNSF for 15 years and counting. “Forever” is not a slogan; it is a consequence of finding combinations of business model, market position, and management where time strengthens the economics instead of eroding them.

What This Means for Operators

The Berkshire model is not primarily an investment framework. It is an operating framework. The lesson is not “buy and hold good stocks.” The lesson is that exceptional long-term outcomes come from:

1. Businesses with durable competitive advantages (moats)

2. Operators with integrity, competence, and owner mentality

3. Organizational structures that give operators autonomy

4. Capital allocation that prioritizes reinvestment over extraction

5. Time horizons measured in decades, not quarters

Most modern “operating advice” diverges sharply from this. The internet is flooded with systems for exponential growth, funnels and playbooks, and obsession with top-line metrics at all costs.

Buffett and Munger focused on the craft of managing an enduring enterprise. That is the same target for Operating by John Brewton: not “how to jack MRR next month,” but how real companies have actually been operated—strategy, incentives, capital allocation, and organizational design—and what that implies for becoming a better operator and capital allocator today.

- john -

Subscribe now

Appendix: Primary Sources and Further Reading

If you want to read Buffett and Munger in their own words on these acquisitions and management principles, here is a curated library of the documents that matter most.

The Letters That Shaped Berkshire’s Operating Philosophy

See’s Candies

• 1972 Letter to Chuck Huggins — Buffett’s original instructions on protecting the See’s brand and culture

• 2007 Shareholder Letter — Detailed retrospective on See’s pricing power and returns after 35 years

Nebraska Furniture Mart

• 1983 Shareholder Letter — The acquisition announcement and Buffett’s explanation of why he trusted Mrs. B

• NFM Offer Letter and 1946 Financials — The legendary 1.25-page handshake deal

GEICO

• 1995 Shareholder Letter — Full acquisition rationale and Tony Nicely assessment

BNSF Railway

• 2009 Shareholder Letter — Acquisition announcement and “all-in wager on the American economy” thesis

Latest Thinking

• 2024 Shareholder Letter — Most recent letter reflecting on 60 years of compounding

Foundational Documents

How Berkshire Operates

• Owner’s Manual — Buffett’s 15 principles for how Berkshire operates and treats shareholders

• Acquisition Criteria — The official (and remarkably brief) criteria for what Berkshire looks for in businesses

Management Philosophy

• Biennial Memo to Managers — The “reputation over profits” memo Buffett sent every two years for 25+ years

Complete Archive

• All Shareholder Letters (1965-2024) — 60 years of unfiltered thinking from Buffett

Secondary Sources Worth Reading

See’s Candies

• Financial Teardown of See’s Candies — Detailed quantitative analysis of See’s returns and economics

GEICO

• GEICO In Numbers — Historical data on GEICO’s growth under Berkshire ownership

Nebraska Furniture Mart

• Rose Blumkin Biography — Deep profile of Mrs. B’s life and operating philosophy

BNSF Railway

• Revisiting the BNSF Acquisition — Analysis of the deal three years after closing

• Matt Rose on Working for Buffett — BNSF CEO describes the Berkshire relationship

I hope you enjoy reading these as much as I have.

- j -

Earlier this week, I published “The Optimization Company” a comprehensive analysis of Amazon’s two-year transformation strategy, from Q4 2023’s margin expansion breakthrough through Q2 2025’s capital-intensity inflection point that sent the stock down 8.3% despite crushing earnings. That piece documented what Amazon is doing and why Wall Street is skeptical. This exclusive deep-dive reveals the strategic DNA behind Jassy’s moves: Andy Grove’s doctrine of “Only the Paranoid Survive.” Grove transformed Intel by exiting the memory business from a position of strength, not crisis, exactly what Jassy is doing now. For paid subscribers who want to understand how great operators recognize and navigate strategic inflection points before markets force their hand, this historical parallel offers invaluable lessons. Jassy isn’t reacting to problems; he’s channeling Grove’s playbook for preemptive transformation.

[Read the main article here]

Read more

This deep-dive is part of Operating’s multi-part series analyzing Amazon’s transformation strategy. If you haven’t yet read the anchor piece—“The Optimization Company: Amazon’s Algorithmic Reckoning”—start there for the full context on Amazon’s $118 billion capital strategy, earnings trajectory from Q4 2023 through Q2 2025, and the systematic decoupling of growth from employment that has Wall Street confused but operators riveted. That comprehensive analysis examined what Amazon is doing. This exclusive piece examines the historical precedent: how Andy Jassy’s current playbook directly channels Andrew Carnegie’s industrial dominance strategy from the 1870s to the 1900s. For paid subscribers seeking to understand not just Amazon’s moves but the 150-year pattern of capital intensity creating competitive moats, this is essential reading.

[Read the main article here]

Read more

In my flagship analysis, “The Optimization Company: Amazon’s Algorithmic Reckoning,” I broke down Amazon’s recent financial performance and strategic transformation, operating margins quintupling from 2.4% to 11.5%, free cash flow compressing 60% as capex surges, and a deliberate workforce reduction strategy targeting 600,000+ avoided hires by 2033. That article examined Amazon’s earnings and economics in depth. This paid subscriber exclusive takes you deeper into the organizational architecture, enabling transformation at this scale. Andy Jassy isn’t just copying Jeff Bezos, he’s executing Alfred Sloan’s century-old playbook for “coordinated decentralization” that made GM the world’s largest corporation. For operators building or analyzing complex enterprises, understanding how structure itself becomes a competitive advantage is critical. This piece shows you exactly how Amazon’s segment model mirrors, and modernizes, Sloan’s revolution.

[Read the main article here]

Read more

New to Operating by John Brewton?

Check out these reader favorites:

• Start With The Work, Not The Why

• Sorry Is Not A Solution

• The Problem of AI-Transformation for Large Companies

• We Are All Becoming Operators

• Operating Economics: The Lump of Labor Fallacy

• Operating Economics: Building Antifragile Companies

Subscribe now

Earning’s Reports

On the morning of July 31st, 2025, Amazon reported quarterly earnings that on paper represented triumph: revenue up 13% year-over-year to $167.7 billion, earnings per share of $1.68 crushing the consensus estimate by 26%, and operating income reaching $19.2 billion. Yet by market close the following day, the stock had plummeted 8.3%, its worst post-earnings reaction in years. Wall Street’s verdict was unambiguous: the company’s guidance for the subsequent quarter implied that Amazon’s unprecedented capital expenditure, despite driving revenue growth, was compressing margins without delivering the operating leverage investors had been promised.​

Just three months later, Amazon has announced plans to eliminate 14,000 corporate positions, part of an expected reduction of up to 30,000 jobs, representing roughly 10% of its white-collar workforce. The timing, two days before third-quarter earnings, is hardly coincidental. It echoes CEO Andy Jassy’s explicit, strategic commitment in June that “in the next few years, we expect that [AI] will reduce our total corporate workforce.”

Beneath these headlines is an instructional transformation story. Internal documents obtained by The New York Times revealed that Amazon expects to avoid hiring more than 600,000 workers by 2033, even as it projects sales volumes to double, achieving this through automation of 75% of its warehouse operations. Combined with over $100 billion in annual capital expenditure (40% higher than 2024), the company is executing one of the most ambitious labor-to-capital substitution strategies in corporate history.​​

The financial results of the past two years tell the story of this transition with unusual clarity. It is a cost optimizing blueprint for a fundamentally different kind of corporation: one where growth and headcount are not as closely aligned, where capital intensity becomes the primary competitive driver, and where traditional assumptions about the relationship between firm success and job creation appear to be diverging.

Amazon's earnings trajectory offers an invaluable case study for strategists and operators like me, most concerned with understanding the future of how companies operate successfully. The question is no longer whether technology enables companies to grow while shrinking their workforce (Amazon has demonstrated that it does, as has Meta in recent times) but rather what this portends for the future of companies, markets, and successful operating execution.

From Cost-Cutting to Capital Intensity

The inflection point came in the fourth quarter of 2023. After years of pandemic-era overhiring followed by painful retrenchment, Amazon reported operating income of $13.2 billion and net income of $10.6 billion, more than triple the prior year’s profit. The stock surged 9.5% in a single day. Investors were witnessing the fruits of Andy Jassy’s unrelenting efficiency campaign: substantial layoffs, the regionalization of the fulfillment network into eight autonomous zones, and an obsessive focus on cost per unit shipped.​

Throughout 2024, this margin expansion story continued. Operating margin, which had languished at 2.4% in 2022, reached 11.3% by year’s end, the highest sustained level in company history. Each quarterly earnings call reinforced the narrative: 13% revenue growth in Q1, 10% in Q2, 11% in Q3, with operating income soaring 219%, 88%, and 56% respectively. Amazon Web Services, the company’s profit engine, stabilized at 17-19% growth after several quarters of deceleration, while advertising revenue climbed 20-26% quarterly.​

Systematic Margin Transformation: Amazon achieved operating margin expansion across every business segment from 2022 through 2025. Total company margins quintupled while AWS maintained 32-40% profitability, North America retail turned structurally profitable above 7%, and International finally crossed into positive territory, all amid record capital spending.

Wall Street’s response, however, was bifurcated. Despite consistent earnings beats, the stock fell after four of the eight quarters from Q4 2023 through Q2 2025. The market had begun focusing less on backward-looking results and more on forward guidance, particularly operating income projections. When Q2 2025 guidance implied margins below expectations, despite the massive revenue beat, analysts concluded that Amazon’s escalating capital spending was beginning to overwhelm its operational efficiency gains.​

The Paradox of Success: Amazon’s operating margins climbed steadily from 7.8% to 11.5% between Q4 2023 and Q2 2025, yet stock reactions remained volatile—falling sharply after beats when guidance disappointed.

The numbers bore this out. Capital expenditures hit $31.4 billion in Q2 2025 alone, running at a $125 billion annual rate. Free cash flow, which had reached $50 billion on a trailing twelve-month basis in early 2024, compressed to roughly $20 billion by mid-2025, a 60% decline. Amazon was explicitly trading near-term cash generation for long-term structural transformation.​

The strategy's duality makes it striking. The capital deployment flows into two parallel but complementary initiatives: AWS data center expansion to address AI infrastructure demand and warehouse robotics to systematically eliminate human labor from physical operations. The former aims to capture the “once-in-a-lifetime” artificial intelligence opportunity. The latter pursues something far more radical: proving that a business processing billions of physical items annually can operate with a fraction of its current workforce.​

The Economic Architecture of Automated Operations

Amazon is an operation being reimagined from first principles. At their Shreveport, Louisiana, fulfillment center, which opened in 2024, employment is already 25% below what it would have been without automation. By year two, the target is 50% below the traditional staffing model. The Stone Mountain, Georgia, facility (built only five years ago) is being retrofitted at significant expense to process 10% more volume with 1,200 fewer workers, a 30% headcount reduction.​

The economics are compelling. Amazon projects savings of 30 cents per unit processed through automation, a modest figure on its face, but one that translates to billions in annual cost avoidance when multiplied across the company’s volume. A $10 billion robotics investment through 2027 is expected to generate $12.6 billion in labor cost savings, representing a 26% return. With warehouse wages averaging $19.50 per hour and typical tenure measured in months, not years, the payback period for robotic systems amortized over five to seven years pencils out to roughly two to three years, an internal rate of return that few corporate investments can match.​​

The Economic Logic of Automation: Amazon’s $10 billion warehouse robotics investment through 2027 generates a 26% return on investment through labor cost avoidance—30 cents saved per item processed, 600,000+ jobs avoided by 2033, and a payback period of just 2-3 years.

They’re not experimenting. Amazon has surpassed one million deployed robots in its operations network. The company’s acquisition of Kiva Systems in 2012 for $775 million, considered expensive at the time, now appears prescient. Those early investments created the foundation for increasingly sophisticated systems: robotic arms that can handle items of varying shapes and weights, autonomous mobile robots that transport inventory without human guidance, and AI-powered systems that optimize routing and picking sequences in real-time.​​

Critically, these are not tools that augment human labor but rather systems designed to eliminate it. Where traditional fulfillment centers employed 2,000 workers, new automated facilities target 800, with the majority of the remaining positions either targeting specialized technicians or temporary workers for peak periods.​

Daron Acemoglu, the MIT economist whose research on automation and employment has become foundational, observed to The New York Times:

“Nobody else has the same incentive as Amazon to find the way to automate... Once they work out how to do this profitably, it will spread to others, too”.

His prediction: Amazon will become “a net job destroyer, not a net job creator.” This stands in direct contradiction to classical economic theories that assume historical waves of automation create more jobs than they eliminate.​

The distinction lies in intentionality. Previous industrial automation was often a byproduct of productivity improvements. Amazon’s strategy is explicitly designed to avoid employment growth even as volumes surge. The company’s projection of doubling sales by 2033 without workforce expansion implies labor productivity (measured as revenue per employee) must double over eight years, a sustained rate without modern precedent.​

The Decoupling of Growth and Employment: Amazon projects doubling sales volume by 2033 while reducing total workforce from 1.6 million to approximately 1 million. The widening gap represents unprecedented labor productivity in modern retail.

Operating History: How Carnegie Used Capital to Change an Industry

Wall Street’s Education: When Beats Become Disappointments

The tension between Amazon’s strategic vision and market expectations crystallized most clearly in the second quarter of 2025. The results themselves were exceptional: revenue growth accelerated to 13%, AWS revenue reached a $123 billion annual run rate (up 17.5%), advertising climbed 23%, and EPS of $1.68 demolished estimates. Brian Olsavsky, Amazon’s CFO, emphasized that AWS infrastructure constraints, not demand weakness, were limiting growth, and that massive capital deployment would alleviate these bottlenecks in subsequent quarters.​

Yet Q3 operating income guidance of $15.5 billion to $20.5 billion, with a midpoint of $18 billion, fell short of the Street’s $19.5 billion expectation. Analysts immediately grasped the implication: Amazon’s operating margin was compressing from 11.5% in Q2 to roughly 10-11% in Q3, despite all the efficiency initiatives. The capital intensity required to build data centers, install GPU clusters, and retrofit warehouses was overwhelming the cost savings from labor reduction, at least in the near term.​

The market reaction was direct. The stock fell 8.8% over two trading sessions, wiping out nearly $180 billion in market capitalization. Analysts who had maintained “Strong Buy” ratings scrambled to explain the divergence. RBC Capital Markets called it “an uncharacteristic setback” and predicted AWS would reaccelerate once new capacity came online. Wedbush Securities was less forgiving, warning that “AWS’s image as an AI leader may be taking a hit” as Microsoft Azure and Google Cloud posted growth rates nearly double Amazon’s.​

This pattern, beats followed by selloff, has become characteristic of Amazon’s relationship with investors. The company beat earnings estimates in all eight quarters from Q4 2023 through Q2 2025, yet the stock fell after four of these reports. What changed was the market’s recognition that Amazon had entered a fundamentally different phase. The margin expansion era of 2023-2024, driven by cutting costs from the existing business, had given way to a capital deployment era where billions flow out before returns materialize.​

The echo of this moment reaches back nearly three decades. In his 1997 letter to shareholders, Amazon’s first as a public company, Jeff Bezos established what he called an “It’s All About the Long Term” philosophy that would define the company’s decision-making for decades. He wrote explicitly:

“We will continue to make investment decisions in light of long-term market leadership considerations rather than short-term profitability considerations or short-term Wall Street reactions.”

More provocatively, when asked about profitability in a 1997 New York Times interview, Bezos declared: “We are not profitable. We could be. It would be the easiest thing in the world to be profitable. It would also be the dumbest. We are taking what might be profits and reinvesting them in the future of the business.” He was asking investors to accept years, potentially a decade, of losses in exchange for market leadership and scale that would eventually generate substantial returns.​

Bezos was so committed to this compact that he attached the 1997 letter to every subsequent annual report through 2020, ensuring each new cohort of shareholders understood the terms of engagement. In his 2003 letter, he wrote: “Long-term thinking is both a requirement and an outcome of true ownership. Owners are different from tenants.” The philosophy enabled Amazon to pursue the “Get Big Fast” strategy that prioritized growth over profitability, building fulfillment infrastructure and expanding into new categories while operating at losses that would have destroyed most companies.​

The parallel to AWS’s own history is instructive. Amazon Web Services required years of heavy investment before becoming profitable, ultimately generating 32-40% operating margins that now contribute over $50 billion annually in operating income despite representing only 18% of revenue. The bet paid off spectacularly: AWS transformed from an internal cost center to a business generating more operating income than Amazon’s entire North American retail operation. Management is betting that robotics and AI will follow a similar J-curve: near-term margin pressure in exchange for a structurally superior cost base by 2027-2028.​

The J-Curve Investment Cycle: Amazon’s capital expenditures surged 40% to $118-125 billion annually in 2025, driven by AWS infrastructure and warehouse automation, while free cash flow compressed from $50 billion to $20 billion. Management is explicitly trading near-term cash generation for long-term structural transformation, echoing the AWS buildout years.

Yet there is a crucial difference that explains Wall Street’s ambivalence. AWS built entirely new revenue streams in the cloud infrastructure market that grew from nearly zero to over $200 billion globally in less than two decades. Warehouse automation, by contrast, reduces costs on existing revenue. If successful, it improves profitability without expanding the top line. The savings from eliminating 600,000 positions over eight years, even at $12.6 billion cumulatively, pale in comparison to the $100+ billion in annual revenue AWS now generates.​

This makes the current strategy simultaneously lower-risk (Amazon retains its core retail business regardless of automation outcomes) and lower-upside (cost savings, while substantial, cannot match the transformative impact of a new $100 billion+ revenue stream). Where Bezos asked 1997 shareholders to tolerate losses while building a new category, Jassy is asking 2025 shareholders to tolerate margin compression while optimizing an existing one. The former bet on market creation; the latter bets on operational excellence.

What remains consistent across both eras is Amazon’s willingness to absorb near-term investor pain for long-term structural advantage. Bezos built this tolerance into the company’s DNA, writing in 1997: “When forced to choose between optimizing the appearance of our GAAP accounting and maximizing the present value of future cash flows, we’ll take the cash flows.” The question facing today’s investors is whether they still possess the patience that Bezos demanded, and whether the automation bet, unlike AWS, offers rewards commensurate with that patience.

Operating History: Only the Paranoid Survive: Grove, Inflection Points, and Acting from Strength

Four Strategic Operating Principles

Amazon’s transformation, viewed through the lens of its financial results and operational decision making, yields four principles that are applicable beyond the company itself, and essential for all operators to understand.

One

Model the capital-labor substitution frontier with precision. Amazon’s automation decisions are not driven primarily by technological feasibility, most warehouse tasks have been technically automatable for years, but by detailed financial modeling. Internal documents show exact break-even calculations:

• At what wage level does robotics become economically superior?

• What is the payback period for specific automation investments?

• How do these economics vary by facility type, throughput volume, and labor market conditions?​

This approach treats automation as a capital allocation decision, not an engineering challenge. The implications are significant: automation accelerates not when technology improves but when the factor price ratio shifts. Post-pandemic wage pressure and labor scarcity tipped numerous tasks past the automation threshold, accelerating deployment timelines that had existed in theory for years. Strategic operators should similarly map their own substitution frontiers, identifying which roles become automatable at which cost structures, rather than waiting for technology to “mature.”

Two

Sequence investments for compound returns. Amazon’s phasing is deliberate: regionalization of fulfillment (2021-2023) to reduce shipping distances and improve speed, followed by robotics deployment (2024-2027) to amplify those gains, followed by AI coordination systems (2025-2028) to optimize the automated network. Each wave creates infrastructure for the next while maintaining adequate profitability to fund subsequent phases.​​

The alternative, attempting simultaneous transformation, would have collapsed operating margins below levels Wall Street would tolerate, potentially constraining the company’s ability to raise capital or defend against activist investors. By maintaining 10-11% operating margins throughout the transition, Amazon preserves strategic flexibility. This sequencing discipline, more than any individual technology, distinguishes successful large-scale transformations from value-destroying restructurings.​

Three

Third, exploit platform economics to decouple growth from headcount. Amazon’s ability to project doubling sales without workforce expansion depends critically on its platform model. Growth increasingly comes not from Amazon’s direct operations but from third-party sellers (now 62% of units sold), AWS customers, and advertisers. The company earns fees, commissions, and service charges on economic activity it enables but does not directly execute.​

Third-party seller services generated $156 billion in 2024, growing 11% quarterly in 2025, while advertising reached $63 billion annually with 20%+ growth. Both carry significantly higher margins than first-party retail. Combined, they represent a structural shift from Amazon as retailer (employing people to sell things) to Amazon as infrastructure provider (collecting rent on others selling things). This platform transition is what makes the automation strategy economically viable: revenues grow through the ecosystem while direct employment shrinks through automation.​

For strategists in other industries, the principle generalizes: businesses that can capture value from activity they orchestrate rather than execute directly can scale with minimal marginal labor. This explains Big Tech’s valuation premiums, AWS, Google Cloud, and Meta’s advertising platform achieve revenue-per-employee ratios 3-5x traditional businesses. Amazon’s bet is that platform economics can be imported into physical operations through sufficient automation.

Operating History: Alfred Sloan’s Lessons: Structure for Scale

Four

Deploy capital intensity as a competitive moat. Amazon’s $100 billion+ annual capex is simultaneously a competitive weapon and a barrier to entry. Walmart, Target, and other retailers cannot match this investment level, creating a time-based advantage in automation deployment. By the time competitors respond, Amazon will have accumulated years of learning curve advantages, superior robotics software, and economies of scale in hardware procurement.​

This represents an inversion of lean startup orthodoxy. In mature, capital-intensive industries, the optimal strategy may be maximum investment velocity, spending to the edge of capital market tolerance to build irreproducible infrastructure advantages. Amazon’s stock underperformance in 2025 relative to other Magnificent Seven companies (up just 3.9% versus the S&P 500’s 16.5%) suggests the company is testing those limits. But management is explicitly betting that long-term competitive position justifies near-term multiple compression.​

The strategic question for competitors becomes: Can they afford not to match Amazon’s capital intensity? If Amazon achieves its automation targets while rivals remain labor-intensive, the cost structure differential could prove insurmountable. This creates a potential arms race dynamic where all major retailers must pursue aggressive automation regardless of financial pressure, simply to avoid being rendered uncompetitive. The industry-wide implications could dwarf Amazon’s individual impact.

The Broader Reckoning

Amazon’s trajectory forces confrontation with a question that has lurked at the edges of economic discourse for decades: What happens when productivity growth and employment growth fully decouple at the firm level?

The post-war economic consensus held that technological progress raised productivity, which increased profits, which funded expansion, which created employment. It was a virtuous cycle where growth and jobs moved together. Amazon and others are engineering the opposite: technological progress that raises productivity, which increases margins, which funds further automation, which explicitly avoids employment. The cycle is still virtuous for shareholders and consumers, but it systematically excludes labor from the value distribution.

The academic literature provides frameworks for understanding this. Acemoglu and Pascual Restrepo’s seminal work on automation identifies two countervailing forces: the displacement effect (technology directly substitutes for labor) and the productivity effect (cost savings enable output expansion that creates new labor demand). The traditional assumption is that productivity effects dominate at the firm level, even if displacement creates negative externalities at the societal level.​

Amazon challenges this. Management’s explicit goal is displacement without corresponding productivity-driven job creation. The savings from automation flow to shareholders via buybacks (Amazon recently authorized $20 billion), to consumers via lower prices that defend market share, and to capital reinvestment in more automation.

If this model proliferates, the macroeconomic implications are profound. Amazon remains the second-largest private employer in America. If the top 100 US employers adopted similar labor-intensity reduction targets, and competitive pressure suggests many will, the aggregate employment impact could reach 5-10 million fewer jobs by 2035 under current growth assumptions. Walmart has already committed to flat headcount despite growth. JPMorgan Chase circulated an internal memo emphasizing “a strong bias against hiring,” and Goldman Sachs instructed managers to “constrain headcount growth.”​

The question is not whether this is technologically feasible; Amazon is proving that it is, but how it can become socially sustainable. Labor’s share of national income has declined from roughly 65% in the 1970s to under 60% today. If automation accelerates this trend toward 50% or below, the distributional consequences become severe: a shrinking portion of economic output flows to workers, while an expanding share accrues to capital.​

Yet constraining automation through policy faces formidable obstacles. Proposals like automation taxes (to capture productivity gains for redistribution), displaced worker funds (requiring companies to finance retraining), or labor board representation (giving workers a voice in automation decisions, as in Germany) all reduce private returns to automation, potentially slowing adoption and making the implementing jurisdiction less competitive globally. This creates race-to-the-bottom dynamics: Amazon can pursue aggressive automation because U.S. corporate governance imposes minimal stakeholder obligations. Jurisdictions that regulate heavily risk losing their corporate tax base.​

Operating by John Brewton goes deep on what it takes to build, scale, and optimize modern companies. This Amazon analysis is just the beginning.

What Paid Subscribers Get:

📬 3x Weekly Operating Notes – Short, actionable insights on corporate strategy, operating excellence, and business history

📊 Epic Operating Resource Articles – Deep-dive case studies and frameworks (like the Carnegie, Sloan, and Grove comparisons) available exclusively to paid members

💬 Paid Subs-Only Chat & Operating Working Group – Direct access to John and a community of operators tackling real challenges

🎙️ Quarterly Live Operating Q&A – Community working sessions where we break down current strategies and answer your questions

Operating Founders get even more:

🤝 Quarterly 1:1 Operating Sessions with John – Four 30-minute private sessions per year to work through your specific operating challenges

Plans:

Monthly: $17/month

Annual: $95/year (save 54%)

Operating Founder: $550/year (includes 1:1 sessions)

Most importantly, thank you for reading! I’d love for you to join our growing community and help you solve your biggest operating conundrums in any way I can.

Conclusion: The Optimization Company

Amazon’s evolution over the past two years represents more than quarterly earnings optimization or tactical cost management. It marks a fundamental inflection point in the history of corporate operations, comparable in significance to Alfred Sloan’s divisional structure at General Motors in the 1920s, Toyota’s just-in-time manufacturing in the 1970s, or Dell’s direct-to-consumer supply chain innovation in the 1990s.

What distinguishes Amazon’s current transformation is its systematic application of capital as the primary lever for operational advantage. The company is demonstrating, with remarkable transparency in its financial results, that mature businesses can achieve simultaneous revenue growth and workforce reduction through deliberate substitution of technology for labor. The mathematics are straightforward: $10 billion in robotics investment generating $12.6 billion in labor savings represents a 26% return on a specific class of capital deployment. The strategic insight lies in recognizing that this return exceeds most traditional growth investments while also creating competitive barriers through accumulated learning curves and infrastructure advantages.​

The earnings trajectory from Q4 2023 through Q2 2025 provides an unusually clear window into the mechanics of this transition. Operating margins expanding from 7.8% to 11.5% while the company simultaneously invests over $100 billion annually demonstrates that the Bezos-era principle of “long-term thinking over short-term profits” remains intact, just applied to a different objective. Where early Amazon sacrificed profitability to build market dominance, today’s Amazon sacrifices near-term cash flow to build operational dominance through automation and infrastructure density.​ For an operator like me, it’s a wonder to behold.

This approach represents what might be termed the capital-intensive operations strategy: the recognition that in mature industries with established demand, competitive advantage derives not from discovering new markets but from achieving structural cost positions that rivals cannot replicate. Amazon’s $118 billion in projected 2025 capital expenditure creates a time-based moat, competitors who match this spending three years later will face an Amazon that has already optimized its systems, accumulated operational data, and moved further down the automation learning curve.​

The parallel to Amazon Web Services is instructive not just for the J-curve investment pattern, but for what it reveals about strategic sequencing. AWS succeeded because Amazon built expertise in large-scale infrastructure operations for internal use, then commercialized that capability. The current automation strategy follows identical logic: develop sophisticated robotics coordination for warehouses, then potentially license those systems to the thousands of logistics operators who cannot afford comparable R&D investments. The $775 million Kiva Systems acquisition in 2012 appears, in retrospect, as foundational as the decision to build AWS infrastructure that could be externalized.​

The business history significance lies in how this strategy could reshape entire sectors. Walmart’s commitment to flat headcount, the logistics industry’s accelerating automation, and corporate services firms deploying AI agents all suggest Amazon is establishing a template rather than an outlier approach. If the efficiency gains materialize as projected, margins reaching 15% by 2028, revenue per employee doubling, the competitive pressure on industry peers to adopt similar strategies becomes overwhelming.​

What Amazon will report on October 30th matters less than what the company has already demonstrated: that modern enterprise strategy increasingly centers on the optimization of capital deployment across both digital infrastructure (AWS, AI) and physical operations (robotics, automation) simultaneously. The earnings calls provide a real-time record of management navigating this dual transformation, making mistakes (Q2 2025’s margin miss), adjusting (the 14,000 job cuts to preserve targets), and maintaining strategic conviction despite market skepticism.​​

For students of corporate strategy and business history, Amazon’s current phase offers a laboratory for understanding how companies evolve when traditional growth constraints (market size, distribution reach) give way to operational constraints (labor availability, capital efficiency). The company is, in effect, answering a question that has lurked in business literature since the first industrial revolution:

What happens when technology enables a firm to grow output while shrinking inputs, and management possesses both the capital and the institutional patience to execute that transformation systematically?

The answer, as revealed through eight quarters of earnings results and internal strategic documents, is that such firms can achieve sustained margin expansion, defend competitive position through capital-intensive moats, and fundamentally alter the economics of their industries. Whether this represents the optimal path for all businesses is sector-dependent. Whether it succeeds for Amazon specifically awaits the completion of this investment cycle around 2027-2028. But that it represents a coherent, historically significant operating strategy is beyond dispute.

Future business historians examining the 2020s will likely identify this period as the moment when capital-labor substitution transitioned from an academic concept or manufacturing-sector reality to a systematic strategy deployed by the world’s largest consumer-facing companies. Amazon’s earnings, read carefully, provide the chronicle of that transition, complete with financial results, management commentary, and market reactions that reveal both the promise and the friction inherent in such profound operational transformation.

For practitioners building and operating companies in the decades ahead, Amazon’s approach offers a beautiful strategic framework: a detailed case study in how patient capital, operational discipline, technological infrastructure investment, and systematic factor substitution can be combined to achieve competitive positioning that may prove difficult to erode. Understanding this strategy, its mechanics, its economics, its sequencing, and its trade-offs becomes essential not because every company should replicate it, but because every operator must comprehend the competitive dynamics it creates.

The lesson, ultimately, is that corporate strategy and operating excellence remain inseparable. Amazon’s financial performance over the past two years demonstrates that the greatest strategic advantages still derive not from clever positioning or marketing innovation, but from superior execution of complex operational transformations that require years of sustained investment and unwavering management focus. That has been true since the days of Carnegie Steel and Ford Motor Company. It remains true in the age of algorithmic operations and artificial intelligence.

The tools change.

The strategic principles endure.

John Brewton writes about business strategy, economic history, and the future of companies at Operating by John Brewton. This analysis draws on Amazon’s public earnings reports, internal documents obtained by investigative journalists, academic research on automation economics, and two years of quarterly earnings calls and analyst reactions.

If you’d like to work together, I’ve carved out some time to work 1:1 each month with a few top-notch Founders and Operators. You can find the details here.

John Brewton documents the history and future of operating companies at Operating by John Brewton. He is a graduate of Harvard University and began his career as a Phd. student in economics at the University of Chicago. After selling his family’s B2B industrial distribution company in 2021, he has been helping business owners, founders and investors optimize their operations ever since. He is the founder of 6A East Partners, a research and advisory firm asking the question: What is the future of companies? He still cringes at his early LinkedIn posts and loves making content each and everyday, despite the protestations of his beloved wife, Fabiola, at times.

Hello, Substack! I’m experimenting with a new format: long-form visuals built around economic history, data storytelling, and strategic insight to articulate the perspective I usually write within as an artic…

Read more

In 2012, Tesla was a niche electric car manufacturer burning through cash with fewer than 3,000 employees. Today, it's simultaneously an automotive company, energy storage provider, solar panel manufacturer, autonomous driving software developer, and charging network operator, worth more than most traditional automakers combined. Tesla demonstrated how …

Read more

He didn’t build a company. He didn’t scale a startup. He didn’t even run a team.

But in 1937, a 27-year-old economist named Ronald Coase gave the world something more enduring: the reason companies exist.

He wrote one essay, “The Nature of the Firm,” that changed ho…

Read more