The Last Hundred Feet
On a mild fall afternoon in downtown San Francisco, Tony Xu — co-founder and CEO of a company valued north of $100 billion — was sprinting toward a white Audi SUV parked in a questionably legal spot on a narrow two-way side street in SoMa. He was not late for a board meeting. He was late on a delivery. Four orders from the same ghost kitchen, what Xu called "playing the game on extra hard mode." As he neared the car, he thrust his iPhone toward the passenger window. "The order's not ready yet but look, I want to show you something cool," he said. An alert had appeared:
Would you like to unassign from the order? The algorithm had determined that one tardy order was about to compromise three others.
Triage, automated and instantaneous, the kind of micro-optimization invisible to anyone who has never stood on a curb outside a takeout-only kitchen watching an ETA tick past the promise window.
This is the image that explains DoorDash: a billionaire CEO earning $19 for an hour of deliveries, geeking out over the logic that decides which burrito to abandon. Not the Super Bowl ad, not the market share chart, not the IPO pop. The algorithmic triage. Because DoorDash's essential insight — the one that separated it from a half-dozen well-funded competitors and transformed it from a Stanford class project called PaloAltoDelivery.com into a Fortune 500 company — is that the delivery wars would not be won by whoever spent the most on customer acquisition or signed the most exclusive restaurant deals. They would be won by whoever mastered the last hundred feet. The parking spot. The building entrance. The dessert most likely to be forgotten. The seconds between a Dasher's arrival and the moment food crosses a threshold.
"We are trying to build the catalog for the physical world," Xu told a reporter during that ride-along. "This repository of information does not exist on Google Maps. It doesn't exist on ChatGPT. We are compiling it all for the first time."
A bold claim. But consider: in 2024, DoorDash processed $80.2 billion in Marketplace Gross Order Value, generated $10.7 billion in revenue — up 24% year-over-year — and posted its first-ever annual profit. It controls roughly 60% of the U.S. food delivery market, more than double its closest competitor, Uber Eats. It operates in more than 40 countries. Its ad business crossed $1 billion in annualized revenue. And its CEO still does delivery shifts, still pulls mid-street U-turns that impress New Yorkers, still notices when the app fails to highlight desserts.
By the Numbers
DoorDash at a Glance
$80.2BTotal Marketplace GOV (2024)
$10.7BRevenue (2024, +24% YoY)
~60%U.S. food delivery market share
$108B+Market capitalization (mid-2025)
37M+Monthly active users
40+Countries of operation
$1.8B[Free](/mental-models/free) cash flow (2024)
$1B+Annualized advertising revenue
The Immigrant's Arithmetic
Tony Xu was born Xu Xun in Nanjing, China, and arrived in the United States at age five, speaking no English. His mother — a doctor in China — worked as a restaurant dishwasher in Illinois because her medical credentials didn't transfer. The family moved through a succession of restaurants where his mother eventually waitressed, then managed. Xu grew up watching the economics of small restaurants from the inside: the brutal margins, the dependence on foot traffic, the way a single bad Friday night could wreck a month's P&L. The experience left a specific imprint — not the vague Silicon Valley platitude about "empowering small business" but a visceral understanding of what it means when a restaurant can't reach customers beyond a three-block radius.
He went to UC Berkeley, then worked at McKinsey, then enrolled at Stanford's Graduate School of Business. There he met Andy Fang, a Stanford computer science undergraduate from Texas who would become DoorDash's head of engineering and who, in the company's mythology, once got so lost making deliveries that the founding team had to track him via Find My Friends. Stanley Tang, another Stanford undergraduate from Hong Kong, rounded out the founding group. The fourth co-founder, Evan Moore, would later depart, but in 2012 the four of them shared a problem statement that emerged not from a market-sizing exercise but from Xu's childhood: local businesses — restaurants especially — were technologically stranded, unable to offer delivery because the logistics were prohibitively complex.
The origin story is worth lingering on because it reveals the company's deepest structural commitment. DoorDash has always framed itself as a logistics company whose first product happened to be food delivery, not a food delivery company that built logistics. The distinction sounds like semantics. It is not. It dictated everything — the suburban strategy, the category expansion, the willingness to operate at negative margins for years while building density in markets competitors ignored.
Palo Alto Delivery and the Seed of Obsession
The company launched in January 2013 as PaloAltoDelivery.com — a domain acquired for less than $10. The "office" was Stanford student housing. The "fleet" was the founders' personal cars. The "dispatch system" was Google Voice and the Find My Friends app. Marketing consisted of fliers on dorm bulletin boards. The menu selection: a couple dozen restaurants in Palo Alto.
There was no consumer app, no Dasher app, no merchant tablet. No way to know when your order was approaching. The company operated fewer than six hours a day and didn't offer breakfast. Every person on the founding team performed deliveries, handled customer support, placed orders, created menus, managed restaurant accounts, and onboarded drivers. All of them. All of the things.
We each did all of the things because we deeply wanted to understand our customers, learn how to master our craft and how to prioritize our product roadmap.
Three months in — September 2013 — Stanford hosted its first home football game of the season, a noon kickoff that ended around 4 PM. The system was flooded with orders. Demand overwhelmed supply. The entire team went out to deliver. They were late on every order by more than an hour. And they were nearly out of money — Xu hadn't yet secured seed financing. Without deliberation, the co-founders decided to refund every single order from that night. The refunds consumed roughly 40% of the company's remaining runway. Then they stayed up all night baking cookies, writing personal apology notes, and delivering care packages to every disappointed customer by 5 AM.
The story has acquired mythic status inside DoorDash, invoked in every consequential decision since. But strip away the sentimentality and what remains is a unit-economics decision made in reverse: the founders chose to spend capital they didn't have to preserve a reputation they hadn't yet earned, betting that the long-term value of customer trust would exceed the short-term cost of near-insolvency. It was either profoundly wise or profoundly reckless. The distinction only became clear later.
The Suburban Thesis
The conventional wisdom in food delivery circa 2014–2017 was that the business was urban. Density equals efficiency. More restaurants per square mile means shorter delivery radii. More potential customers per block means lower customer acquisition costs. GrubHub dominated New York. Uber Eats leveraged Uber's existing driver network in major metros. Postmates was a San Francisco darling. The venture capital logic pointed toward the same dense ZIP codes that had made ride-hailing viable.
DoorDash zagged. While competitors fought over Manhattan and San Francisco and Chicago, DoorDash expanded into suburban markets and mid-tier cities — places like Cincinnati, Milwaukee, the Inland Empire, Boulder. By the time it turned four years old in mid-2017, it operated in more than 500 cities with 59,000 listed restaurants and over 100,000 Dashers.
The suburban thesis rested on three interlocking insights. First, suburban customers had larger average order values — families ordering dinner, not individuals grabbing a solo lunch — which improved unit economics per delivery. Second, competition was nonexistent in most of these markets, meaning customer acquisition costs were dramatically lower. Third, and most counterintuitively, the lower density actually created a stronger competitive position: once DoorDash built the driver network in a suburban market, no rational competitor would invest to replicate it because the market wouldn't support two platforms.
DoorDash's geographic expansion, 2013–2020
2013Launches in Palo Alto with a couple dozen restaurants.
2015Expands beyond the Bay Area into Southern California suburbs and select metros.
2017Reaches 500+ cities and 59,000 restaurant partners. Begins signing national chain partnerships (Wendy's, Taco Bell, Dunkin' Donuts, Cheesecake Factory).
2018Aggressively enters mid-tier and suburban markets, expanding from ~1,500 to ~6,000 locations. Poaches key Uber Eats executives.
2019Surpasses GrubHub to become #1 in U.S. food delivery market share for the first time.
2020COVID-19 lockdowns supercharge the suburban bet. Business more than triples. DoorDash captures 50%+ of the U.S. market.
The timing of the suburban push was as important as the strategy itself. In the late 2010s, Uber was in crisis — its culture was being publicly incinerated, Travis Kalanick had been ousted, and new CEO Dara Khosrowshahi was imposing financial discipline on a company that had treated cash like a renewable resource. Uber Eats, which might have been DoorDash's most dangerous competitor, was being pulled back from aggressive expansion as the parent company prioritized profitability and an IPO narrative. DoorDash struck while there was blood. It expanded from 1,500 locations to 6,000 in a single year, poached several Uber Eats executives, and planted flags in markets that Uber's new fiscal discipline wouldn't let it contest.
The Three-Sided Marketplace
Every marketplace is a chicken-and-egg problem. DoorDash's was a chicken-and-egg-and-egg problem: three distinct customer constituencies — consumers, merchants, and Dashers — each requiring the other two to exist before the platform had value for any of them.
DoorDash's early strategic vocabulary drew heavily from Amazon's playbook, and the parallel is not accidental. Brad Stone's
The Everything Store documents how
Jeff Bezos built Amazon's flywheel by obsessing over the customer experience while simultaneously building infrastructure that competitors couldn't replicate. Xu read the same playbook but applied it to a three-sided market where "the customer" was not one entity but three, often with conflicting interests. A consumer wants the lowest possible delivery fee. A Dasher wants the highest possible earnings per hour. A restaurant wants the lowest possible commission rate. The platform sits at the intersection, allocating value among them.
DoorDash's resolution — and this is the architectural choice that defines the company — was to treat all three constituencies as "customers" with equal standing. Every quarter, the entire leadership team spends a day working in a merchant's restaurant (the CFO, Ravi Inukonda, has served food in a Philadelphia kitchen) and another day performing deliveries as a Dasher. This is not a PR exercise. It is operationalized in a policy called WeDash: all U.S. salaried employees must complete four delivery shifts per year.
We have three customers: consumers, merchants, and Dashers. When we first started, our goal wasn't to build a food delivery business. The goal was to build a local commerce business.
The three-sided framing also explains DoorDash's product architecture. The consumer app is the visible layer. But beneath it sits the Dasher app — which now includes proprietary mapping data on optimal parking spots, building entrances, and real-time order readiness — and the merchant tools platform, which evolved from a simple order-receiving tablet into a full commerce suite that includes online ordering, delivery logistics, advertising, and now, with the proposed $1.2 billion acquisition of SevenRooms, reservation management and in-store
CRM.
When the World Stopped Going Out
COVID-19 arrived in March 2020 and did something no amount of venture capital could have accomplished: it made food delivery essential rather than convenient. Restaurants that had resisted delivery platforms — viewing the commission rates as parasitic — suddenly had no alternative. Consumers who had never used a delivery app discovered that it worked. And the suburbs, where DoorDash had invested years building driver density and merchant selection, became the center of American consumption as knowledge workers fled urban apartments for houses with spare bedrooms.
DoorDash's business more than tripled in 2020. The company had filed its S-1 with the SEC on November 13, 2020, and went public on December 9 at $102 per share, pricing well above expectations. On its first day of trading, the stock closed at $189.51 — an 86% pop that valued the company at approximately $72 billion. The IPO raised $3.4 billion. The timing was exquisite: the pandemic had compressed years of adoption curve into months, and DoorDash was the primary beneficiary.
But the IPO prospectus also revealed the business's fundamental tension. DoorDash had generated $2.9 billion in revenue in the first nine months of 2020, up from $587 million in all of 2019 — staggering growth that masked a cumulative net loss of $667 million. The company had never been profitable. The S-1 disclosed that "we have a history of net losses and we may not be able to achieve or maintain profitability in the future."
The three co-founders — Xu, Fang, and Tang — held all of the company's Class B shares, which carried 20 votes each compared to one vote for the Class A shares sold to the public. Under a voting agreement, Xu held irrevocable proxy over Fang's and Tang's shares. Upon completion of the IPO, the three founders collectively controlled approximately 69% of the company's voting power, a figure that would rise to 79% as equity awards vested. Tony Xu, in other words, controlled DoorDash in the way that
Mark Zuckerberg controls Meta — a comparison that would become more apt when Xu later joined Meta's board of directors.
The Wolt Gambit
If the suburban thesis was DoorDash's first major strategic bet, the acquisition of Wolt was its second. In November 2021, DoorDash announced it would acquire the Finnish delivery company Wolt in an all-stock deal valued at approximately $8 billion. The transaction closed in 2022.
Wolt, founded in Helsinki in 2014 by Miki Kuusi, was not a distressed asset or a failed competitor being scooped up for parts. It was the best-run delivery company in Europe — profitable in several markets, beloved by consumers in the Nordics, and expanding aggressively into Eastern Europe, the Middle East, and Japan. Kuusi, an entrepreneur with the analytical intensity of a chess grandmaster and the product sensibility of a Scandinavian designer, had built Wolt with the same obsessive attention to operational quality that characterized DoorDash.
The strategic logic was clear: DoorDash needed international scale to amortize its technology investments, and building from scratch in 25+ countries would take a decade and cost more than $8 billion in cumulative losses. Wolt offered an instant international platform with a culture that rhymed with DoorDash's own. The risk was equally clear: cross-border marketplace integrations have an abysmal track record, and the all-stock structure diluted DoorDash's existing shareholders in a company that had yet to prove it could generate profits in its home market.
By 2024, the Wolt integration appeared to be working. International Marketplace GOV was growing, and the combined entity operated in more than 30 countries. Then, in May 2025, DoorDash doubled down: a proposed $3.8 billion acquisition of Deliveroo, the UK-based delivery company, which would give DoorDash a top-three position in the UK and a combined presence in more than 40 countries.
🌍
Building a Global Footprint
DoorDash's international M&A strategy
| Acquisition | Year | Value | Strategic Rationale |
|---|
| Wolt | 2022 | ~$8B (all-stock) | Instant scale in 25+ countries; cultural alignment |
| Deliveroo (proposed) | 2025 | ~$3.8B | UK market entry; 40+ country footprint |
| SevenRooms (proposed) | 2025 | $1.2B | Hospitality software; in-store CRM and reservations |
| Symbiosys | 2025 | $175M | Ad tech; off-platform advertising capabilities |
Beyond the Restaurant
Four years into DoorDash's existence, Tony Xu wrote that "food is only the first piece of the puzzle." By 2024, the puzzle had expanded to include grocery, alcohol, convenience, retail, flowers, pet supplies, and package pickup. More than 99% of DoorDash's monthly consumers in the U.S. had access to a non-restaurant retailer on the platform. Twenty-five percent of total monthly active users were shopping new vertical categories — not just ordering from restaurant menus. The catalog spanned over 11 million grocery and retail products. Key retail launches included partnerships with Albertsons, Wegmans, Lowe's, Ulta Beauty, Sephora, Walgreens, and Walmart Canada.
The category expansion follows a logic as old as Amazon's: high-frequency use cases (people eat three times a day, or roughly 100 occasions a month) generate habitual engagement that can be channeled into lower-frequency, higher-margin categories. A consumer who opens DoorDash for pad thai on Tuesday discovers that she can also get her prescriptions, her groceries, and a bouquet of flowers. DashPass — the $9.99/month subscription that waives delivery fees — is the mechanism that converts occasional users into habitual ones. Roughly half of DoorDash's customers subscribe, and the company claims DashPass has saved consumers over $10 billion globally since its launch in 2018.
The grocery vertical deserves particular attention because it represents both the company's largest growth opportunity and its most dangerous competitive exposure. As of Q4 2023, over 7 million consumers were ordering grocery through DoorDash each month — 20% of the overall user base. But grocery delivery is a structurally different business from restaurant delivery: lower margins, heavier baskets, more SKU complexity, higher customer expectations for freshness and accuracy. And the competitive set includes not just Uber Eats and Instacart but Amazon Fresh, Walmart+, and every regional grocer building its own digital channel.
DashMart — DoorDash's own convenience and grocery micro-fulfillment operation — represents an even more aggressive bet. By operating its own stores within the app, DoorDash captures first-party margin rather than marketplace commission. But it also assumes inventory risk, lease obligations, and operational complexity that a pure marketplace model avoids. It is the classic platform dilemma: do you own the supply or just the demand?
The Machine Learning Engine No One Sees
The visible product — the consumer app, the red branding, the Dasher on the bicycle — obscures the invisible one. DoorDash's core competitive advantage is a machine learning-powered logistics engine that has been continuously trained on a decade of delivery data, and which makes thousands of real-time decisions per second: which Dasher to assign to which order, when to batch multiple orders on a single route, how to predict restaurant preparation times, where to position idle Dashers to minimize expected wait times, and — as Xu demonstrated on that San Francisco ride-along — when to unassign a Dasher from a late order to protect the on-time delivery of other orders in the batch.
The mapping layer alone is extraordinary. DoorDash has built proprietary data on the physical topology of the last hundred feet — which entrance to use in a corporate building, where to park near an apartment complex, which restaurants have dedicated pickup windows versus which require Dashers to wait in the regular line. None of this exists in Google Maps. None of it exists in any public dataset. It is compiled, delivery by delivery, annotation by annotation, by millions of Dashers completing millions of orders, the data flowing back into models that continuously refine the next delivery's routing and timing.
All of this data, we are trying to build the catalog for the physical world. This repository of information does not exist on Google Maps. It doesn't exist on ChatGPT. We are compiling it all for the first time.
The machine learning infrastructure serves a dual purpose. Operationally, it makes each delivery incrementally cheaper and faster — the kind of compounding improvement that, over a decade, creates an enormous gap between the incumbent and any new entrant. Strategically, it makes the platform transferable across categories: the same logistics engine that routes a pad thai order can route a grocery basket or a bouquet of orchids or a package pickup for UPS. The investment amortizes across every category DoorDash enters.
CFO Ravi Inukonda has described the system as "a very efficient logistics engine that's been powered by machine learning for the past 10 years." The understatement is characteristic. What he is describing is the company's moat — and unlike brand loyalty or exclusive contracts, it compounds.
The Advertising Layer
Every successful marketplace eventually discovers the same truth: the most profitable business you can build on top of a marketplace is advertising. Amazon proved it. Google proved it. And DoorDash, with deliberate caution, is proving it now.
DoorDash's advertising business crossed $1 billion in annualized revenue in 2024. The company doesn't break out ad-specific financials, but analysts estimate that ad revenue carries dramatically higher margins than the core delivery business — a dynamic familiar to anyone who has studied Amazon's "other revenue" line. For a company that generated $10.7 billion in total revenue in 2024, a billion-dollar ad business growing rapidly represents a meaningful margin expansion lever.
In June 2025, DoorDash acquired Symbiosys, an ad-tech startup, for $175 million. Symbiosys enables brands that advertise within the DoorDash app to extend their reach to DoorDash customers on other platforms across the web — essentially transforming DoorDash's first-party consumer data into a targeted advertising network that reaches beyond the marketplace itself.
But Xu has been publicly cautious about the ad business in a way that distinguishes DoorDash from companies that have let advertising degrade the consumer experience. "The most important thing of getting the product right is making sure that we can balance the needs of the advertiser with the needs of the consumer," he said on a 2025 earnings call. "A healthy marketplace always precedes and trumps an advertising business."
The tension is real. Every sponsored listing that displaces an organic result makes the consumer experience incrementally worse. Every ad dollar that improves the take rate also increases the incentive to serve the advertiser over the consumer. The companies that navigate this tension successfully — Google for two decades, Amazon in the 2020s — generate extraordinary profits. The companies that don't — see Facebook's news feed circa 2018 — erode the trust that made the platform valuable in the first place.
The Commerce Platform and the SevenRooms Bet
The most ambitious element of DoorDash's current strategy is the least visible to consumers. DoorDash Commerce Platform is a suite of software tools — first-party ordering, delivery management, marketing, and now, with the proposed $1.2 billion acquisition of SevenRooms, reservation and CRM — designed to make DoorDash indispensable to merchants not just as a delivery channel but as a technology platform for their entire business.
The strategic logic echoes Shopify's relationship with e-commerce: rather than merely aggregating demand (the marketplace model), DoorDash wants to provide the operating system that powers a merchant's digital presence — on the marketplace, on the merchant's own website, and increasingly, inside the physical store itself. SevenRooms, which provides reservation management, guest CRM, and marketing automation to restaurants, hotels, and hospitality businesses, extends this ambition into the dine-in experience.
Parisa Sadrzadeh, a former Amazon executive hired to lead DoorDash Commerce Platform, framed the opportunity with precision: "While DoorDash solved a massive problem for merchants during the pandemic … introducing a delivery capability to many who had never considered doing it before, another challenge ended up being, 'How do you grow my actual volume in my physical store, because those are my most profitable consumers?'"
This is the pivot from marketplace to platform, from demand aggregation to infrastructure. If DoorDash executes it, the company becomes not a delivery service that restaurants reluctantly tolerate but the software layer that restaurants cannot operate without. The distinction is the difference between a 15% commission on delivery orders and a SaaS-like relationship with recurring revenue that doesn't depend on consumer demand for delivery.
The risk: DoorDash is attempting to become Shopify and DoorDash simultaneously, and the organizational complexity of serving merchants as both a marketplace that charges commission and a software provider that charges subscription fees creates tensions that are easy to identify and hard to resolve. Merchants already resent commission rates. Asking them to also adopt DoorDash software requires a level of trust that years of commission disputes may have eroded.
A Speck of Dust
In early 2025, on a quarterly earnings call, an analyst asked the question that every investor in a dominant market-share leader eventually asks: Where does the growth come from now? Will it come from new customers or from getting existing customers to order more?
Xu's answer was revealing — both in substance and in tone: "If you took our oldest area of exploration, U.S. restaurants … we're still single-digit percentages of the U.S. restaurant industry sales. If you look globally, that number would be even smaller. We are a speck of dust in terms of how penetrated we are."
Whether this was genuine humility or calculated expectation-management, the math is directionally correct. Total U.S. restaurant industry sales exceed $1 trillion annually. DoorDash's U.S. Marketplace GOV, while enormous, represents a single-digit share of that total. The theoretical ceiling is so far above the current level that the relevant question is not whether DoorDash can grow but at what rate and at what margin.
The company's strategic answer to this question is expansion along four axes simultaneously: geography (the Wolt and Deliveroo acquisitions), category (grocery, retail, convenience, pharmacy), monetization (advertising, subscriptions, commerce platform software), and technology (AI voice ordering, autonomous delivery, advanced mapping). Each axis carries its own risks and capital requirements. Together, they represent a bet that DoorDash's logistics engine and consumer habit loop are generalizable enough to power a local commerce platform that extends far beyond its origins in restaurant delivery.
We still believe growing
GDP locally continues to represent the best way to create economic opportunity and lift everyone up. Put more simply, our mission to grow and empower local economies is as relevant today as it ever has been.
The company's 2024 shareholder letter enumerated four pillars of execution: improving product inputs (selection, affordability, delivery speed and accuracy, customer service), maintaining a high bar for capital allocation before injecting large amounts into new projects, improving unit economics alongside the core product experience, and continuously getting more efficient. It reads less like a vision statement than an operating manual — and the distinction is deliberate. Xu has said that DoorDash invests "significantly more over the past five years than any other period in our history," but within a framework that demands each investment earn its place.
DoorDash now serves 94 of the top 100 restaurants and 44 of the top 50 retailers in the U.S. Its catalog spans more than 11 million grocery and retail products. DashPass membership continues to grow. The company accepted SNAP/EBT payments from grocery partners at more than 4,000 U.S. locations, a quiet move that extends the platform's reach into demographics that most on-demand services ignore. In October 2025, DoorDash unveiled Dot, a five-foot, 350-pound autonomous delivery robot capable of hitting 20 miles per hour — a harbinger of a future where the last hundred feet might not require a human at all.
And Tony Xu still delivers. Four shifts a year, minimum, same as every salaried employee. In the shareholder letter, he noted that he completed DoorDash's first delivery more than twelve years ago — "a chicken pad thai and spring rolls." The domain cost less than $10. All the IKEA furniture in the founding apartment cost less than $2,000. The company is now worth more than $100 billion. And its CEO's clearest articulation of what it does is this: "If you're in technology and you are not making improvements, you are actually decaying. Until it's over all of a sudden."
The delivery arrives. The door closes. The algorithm is already optimizing the next one.
