The Cave and What It Cost to Open It
In the fiscal year ending March 2024, Alibaba Group Holding Limited reported revenue of RMB 941.2 billion — roughly $130 billion — making it one of the largest internet companies on earth by any measure. Its platforms facilitated over $1 trillion in gross merchandise volume annually. Its cloud division served millions of businesses across Asia. Its logistics network, Cainiao, shipped 55 million packages a day. And yet its market capitalization, which had peaked above $850 billion in October 2020, had cratered by more than half, at times trading below $200 billion — a valuation that implied the market believed the company's best days were archaeological. The paradox is the kind that defines an era: the most successful e-commerce company in the most populous country on earth, a firm that had once been valued more highly than Facebook, Amazon, or eBay, was being priced as though its competitive moat had become a liability, its government a jailer, its founder a cautionary tale. The story of Alibaba is not, in the end, a simple narrative of entrepreneurial triumph. It is a story about what happens when a company builds the commercial operating system for a civilization — and then discovers that the civilization's sovereign has opinions about who gets to run the operating system.
By the Numbers
The Alibaba Empire
$130BApproximate annual revenue (FY2024)
$1T+Annual GMV across platforms
~1.3BAnnual active consumers globally
$25BRaised in 2014 IPO — largest in history at the time
$850B+Peak market capitalization (Oct 2020)
~200,000Employees worldwide
55MPackages shipped daily via Cainiao
The English Teacher and the Sealed Cave
Jack Ma — born Ma Yun on September 10, 1964, in Hangzhou, a provincial city of temples and green hills two hours southwest of Shanghai — was not a technologist. He was not an engineer. He failed the national college entrance exam twice before being admitted to Hangzhou Normal University, where he studied English. He was rejected from roughly thirty jobs after graduation, including a position at KFC and one as a hotel waiter. He applied to Harvard ten times and was refused each time. What he possessed, in vast and sometimes bewildering quantities, was charisma, linguistic facility, and an almost pathological refusal to accept that the world's existing power structures were permanent.
His first encounter with the internet came in 1995, on a trip to Seattle arranged through an American friend. He typed "beer" into an early search engine and found results from Germany, Japan, America — but nothing from China. "I was Chinese and there was no Chinese," he would later tell audiences, turning the anecdote into an origin myth with the practiced rhythm of a man who had spent years teaching English to Chinese businesspeople. He returned to Hangzhou and launched China Pages, a quixotic attempt to build a directory of Chinese businesses online. The venture collided immediately with the Chinese government's ambivalence toward the internet, and it fizzled.
Ma spent a brief, miserable stint working for a government-affiliated internet venture in Beijing before quitting and returning to Hangzhou in early 1999. On February 21 of that year — a date as precisely mythologized in Alibaba's corporate canon as Apple's garage or Amazon's Bellevue office — Ma gathered seventeen friends in his small apartment. The video footage, later compiled in Porter Erisman's documentary
Crocodile in the Yangtze, shows Ma in a polo shirt, cross-legged, delivering what amounts to a startup sermon: "We need to learn the hardworking spirit of Silicon Valley. If we have that kind of 8 am to 5 pm spirit, then we should just go and do something else." The eighteen co-founders scraped together approximately $60,000. They named the company Alibaba — chosen, Ma claimed, because it was a name recognized globally, one that evoked the opening of a sealed cave full of riches. The metaphor was apt in ways he could not have known: what they were building would indeed open Chinese commerce to the world, but the cave had guardians, and the guardians would eventually demand tribute.
Survival as Strategy
The first version of Alibaba, launched in April 1999 as Alibaba.com, was a business-to-business marketplace — a platform connecting Chinese manufacturers with overseas buyers. You need quick-dry super-elastic fabric for a wristband? Five hundred million tiny screws? A used 747? It was there, or it would be. The model was simple and, in retrospect, brilliant: China's manufacturing economy was the largest in the world, but its factories had no efficient way to reach global buyers. Alibaba would be the interface.
The timing was both perfect and catastrophic. In October 1999, Goldman Sachs led a $5 million investment. In January 2000, Masayoshi Son's SoftBank invested $20 million — a bet that would eventually become one of the greatest venture capital returns in history. Ma hired aggressively, opened offices in Silicon Valley, London, and Hong Kong, and started spending like it was, well, 2000.
Then the dot-com bubble burst. By early 2001, Alibaba was burning through cash with no clear path to profitability. The company came, by multiple accounts, within months of collapse. Savio Kwan, recruited as COO, found a company where engineers slept in bunk beds in a stuffy room at the office. Kwan's first move was to eliminate the bunk beds — a small, symbolic act that marked the transition from scrappy survivalism to something approaching corporate discipline. Ma was forced to lay off staff in overseas offices and pull back to Hangzhou, concentrating the company's resources on the China market. He called it "Back to China, Back to the Coast, Back to the Center."
This retreat was the first of several defining strategic contractions — moments when Alibaba's leadership chose survival and focus over ambition and sprawl. The instinct to contract before expanding would become a recurring pattern, though the company would not always follow it.
We need to learn the hardworking spirit of Silicon Valley. If we have that kind of 8 am to 5 pm spirit, then we should just go and do something else.
The eBay War and the Invention of Taobao
By 2003, Alibaba.com had stabilized. The B2B marketplace was generating subscription revenue from Chinese suppliers who paid annual fees to list their products. But Jack Ma was watching eBay's aggressive push into China — in particular its 2003 acquisition of EachNet, the leading Chinese consumer auction site — with something between alarm and competitive fury.
Ma's response was a secret project. He gathered a small team and told them, in the retelling that has become corporate legend, to prepare for "Alibaba's biggest challenge yet." The result was Taobao — a consumer-to-consumer marketplace modeled loosely on eBay but designed specifically for the Chinese market. The name means "searching for treasure."
What followed was one of the great platform wars of the early internet era, and the playbook Ma deployed would echo through every subsequent Chinese tech battle. eBay, under then-CEO
Meg Whitman, had market share, brand recognition, and deep pockets. Taobao had none of these things. What it had was a willingness to be free.
Ma announced that Taobao would charge no listing fees and no transaction fees for three years. eBay China charged both. The logic was ruthless: if Taobao could bleed long enough, it could drain eBay's Chinese operation of the economics that justified its investment. eBay's problem was structural — it was a publicly traded American company that needed to report profitable growth in China to its shareholders. Taobao, funded by Alibaba's B2B subscription revenue and SoftBank's deep pockets, had no such constraint.
There was a product dimension too. eBay's China site was essentially a localized version of its American platform — clean, transactional, impersonal. Taobao built in real-time chat between buyers and sellers through a tool called Wangwang, recognizing that Chinese consumers, who lacked the institutional trust infrastructure of credit scores and established retail brands, needed to negotiate, ask questions, and build relationship-based trust before purchasing. eBay's American-designed platform assumed a trust infrastructure that didn't exist in China.
By 2006, Taobao had captured roughly 70% of China's consumer e-commerce market. eBay effectively retreated. Meg Whitman would later admit the defeat at a conference, and the episode became a defining case study in the perils of applying American platform logic to non-American markets. For Ma, it was vindication: the crocodile had learned to swim in the Yangtze.
How Taobao conquered China's C2C e-commerce market
2003eBay acquires EachNet, China's leading auction site, for $150M. Alibaba secretly launches Taobao.
2003Taobao announces zero listing fees and zero transaction fees for three years.
2004Alibaba launches Alipay, an escrow-based payment system, to solve trust deficit in online transactions.
2005Taobao surpasses eBay China in market share for the first time.
2006eBay effectively withdraws from China. Taobao holds ~70% C2C market share.
The Architecture of [Trust](/mental-models/trust)
The Taobao victory was not merely a pricing play. Buried within it was a deeper insight about what e-commerce required in a society without the institutional trust apparatus that Western markets took for granted. In the United States, consumers trusted credit card companies to reverse fraudulent charges, trusted the postal service to deliver packages, trusted brand names to guarantee quality. In China circa 2003, none of this existed at scale. Credit card penetration was minimal. The postal system was unreliable for commercial goods. Counterfeiting was rampant.
Alibaba's answer was Alipay, launched in 2004. The system functioned as an escrow service: buyers deposited payment with Alipay, which held the funds until the buyer confirmed receipt of the goods. Only then did Alipay release the money to the seller. It was, in essence, a trust machine — an institutional substitute built by a private company because the institutions themselves did not yet exist.
This was the foundational insight that would power Alibaba's expansion for the next decade: in a developing economy, the platform that builds the infrastructure of trust captures the economy. Alipay would grow from an e-commerce payment tool into a sprawling financial services platform handling over $519 billion in annual payments by 2013 — more than PayPal, more than any payment system in the world outside of Visa and Mastercard. It would eventually be spun out into Ant Financial (later Ant Group), a company that at its peak in 2020 would be valued at over $300 billion before its IPO was spectacularly halted by Chinese regulators.
The Alipay story also contains the seed of Alibaba's most consequential political crisis, but that comes later. For now, the mechanism is what matters: Alibaba didn't just build a marketplace. It built the financial plumbing that made the marketplace possible. And then it charged rent on the plumbing.
The Yahoo Deal and the Art of Sleeping with Crocodiles
In August 2005, Yahoo invested $1 billion in Alibaba in exchange for a 40% stake and transferred its China operations to the company. The deal was engineered largely by Joe Tsai, Alibaba's co-founder and then-CFO — a Yale-educated, Taiwanese-Canadian lawyer who had left a career in private equity to join Ma in Hangzhou in 1999, taking a salary of $50 a month. Tsai was the diplomatic counterweight to Ma's evangelist energy: precise where Ma was expansive, financially rigorous where Ma was visionary, fluent in the language of Western capital markets where Ma was fluent in the language of Chinese ambition.
The Yahoo deal was brilliant and dangerous. Brilliant because it gave Alibaba $1 billion in cash at a time when the company was still primarily a B2B marketplace generating modest revenue, and it came with Yahoo's China operations — search, email, news — that Alibaba could integrate or discard as it saw fit. Dangerous because Yahoo now owned 40% of the company, and Ma had effectively handed governance influence to an American corporation at a time when the Chinese internet was becoming increasingly sensitive to questions of foreign ownership and control.
The tension between foreign capital and Chinese sovereignty would define Alibaba's corporate structure for the next two decades. Every significant financing decision — the Yahoo deal, the SoftBank relationship, the VIE (variable interest entity) structure that allowed foreign investors to hold economic interests in what were legally Chinese-only businesses, the eventual secondary listing in Hong Kong in November 2019 that raised another $12.9 billion — was a negotiation between the demand for global capital and the imperative of maintaining Chinese operational control.
They are like an Amazon, an eBay and a PayPal.
The Tmall Bet and the Platform's Second Act
Taobao's dominance of consumer e-commerce created a new problem: the marketplace was drowning in counterfeits. Small sellers proliferated, but so did fake goods, gray-market electronics, and knockoff luxury items. International brands were reluctant to engage with a platform famous for being a bazaar of questionable authenticity.
In 2008, Alibaba launched Tmall (originally Taobao Mall), a separate platform designed specifically for branded goods. The distinction was critical. Taobao remained the open bazaar — anyone could sell, listings were free, and the Wild West energy persisted. Tmall was curated: brands like Nike, Apple, L'Oréal, Procter & Gamble, and
Walt Disney established virtual storefronts, paying Alibaba commissions on sales and annual service fees for the privilege of verified authenticity.
The architectural split — free, chaotic Taobao feeding traffic and consumer habit; premium, monetizable Tmall capturing the ad spending and commissions of established brands — became the economic engine of the entire Alibaba ecosystem. Taobao was the funnel. Tmall was the cash register. By 2013, the combined GMV across both platforms exceeded $248 billion — more than eBay and Amazon combined. In the 2013 calendar year, Alibaba reported net income of $3.56 billion on revenue of $7.95 billion. That translates into a profit margin of roughly 45% — a number that Google and eBay could only dream of, because Alibaba, unlike Amazon, held no inventory and operated no warehouses for its core marketplace business. It was pure platform, pure margin.
The economic model was essentially an advertising business wearing the costume of a marketplace. Merchants paid for placement, paid for keywords, paid for featured listings, paid for Tmall storefronts. The consumer paid nothing to browse. The transaction data fed back into Alibaba's advertising algorithms, which became more valuable as more merchants competed for placement, which attracted more consumers, which generated more data. The flywheel was spinning.
Singles' Day and the Manufacture of [Desire](/mental-models/desire)
On November 11, 2009 — 11/11, a date chosen for its visual symmetry and its tongue-in-cheek association with China's unofficial "Singles' Day" holiday — Alibaba launched a 24-hour shopping festival. The first year, 27 merchants participated. Sales totaled roughly RMB 50 million. It was a marketing experiment, nothing more.
By 2013, Singles' Day generated $5.8 billion in GMV in 24 hours — more than Black Friday and Cyber Monday combined. By 2019, the number had swelled to $38.4 billion in a single day. Alibaba turned it into a televised spectacle, complete with celebrity performances, countdown clocks, and a real-time GMV ticker that functioned as both marketing and market signal. The event was a demonstration of concentrated demand — proof that Alibaba could orchestrate the purchasing behavior of hundreds of millions of consumers in a single day, a feat of logistical and psychological engineering that no other company on earth could replicate.
But Singles' Day also became a trap. The annual number had to go up. When it eventually stopped going up — or when regulators questioned the methodology behind the reported figures — the event shifted from asset to liability, from proof of power to emblem of the pressure that scale exerts on narrative.
The $25 Billion Opening
On September 19, 2014, Alibaba Group Holding Limited began trading on the New York Stock Exchange under the ticker BABA. The IPO priced at $68 per share, at the top of an already-raised range, and raised $21.8 billion — the largest initial public offering in American history at that time, eclipsing Facebook's $16 billion offering in May 2012. When the underwriters exercised their overallotment option, the total haul climbed to approximately $25 billion. At the IPO price, Alibaba's market capitalization was roughly $168 billion — more than eBay, Twitter, and LinkedIn combined.
The roadshow had been a phenomenon. A lunch presentation at the Waldorf-Astoria in Manhattan drew more than 800 attendees. One hedge fund with $3 billion in assets placed a multibillion-dollar order, hoping to receive even a fraction of that amount. Alibaba's advisers at Goldman Sachs, Credit Suisse, Deutsche Bank,
J.P. Morgan, Morgan Stanley, and Citi deliberately weighted the allocation toward long-term institutional holders — Fidelity, BlackRock, T. Rowe Price — rather than hedge funds that might flip on the first day.
The IPO minted billionaires. Jack Ma's stake made him, briefly, the richest man in Asia. SoftBank's Masayoshi Son, whose original $20 million investment in 2000 had grown into a stake worth tens of billions, became Japan's richest person. Yahoo's 22% stake was valued at over $31 billion.
For Western investors, the Alibaba IPO was something else entirely: the best available proxy for China's growth. Online shopping in China was expected to grow at 27% annually, and Alibaba was the leader. The filing disclosed revenue growing at 46–56% year-over-year. Nearly 20% of purchases on Alibaba were already happening on mobile phones. The pitch was irresistible: the largest internet market in the world, growing faster than any other, dominated by a single platform with 45% profit margins.
The prospectus also contained a warning that investors largely ignored: "Alibaba's management and major shareholders will control the board, giving ordinary shareholders no power over the direction of the company." The company's governance relied on a "partnership" — a group of 27 insiders (later expanded) who held the exclusive right to nominate a majority of board members. This structure was not quite dual-class stock, but it achieved the same result: Ma and his inner circle controlled the company regardless of their economic ownership. Hong Kong's stock exchange, which required one-share-one-vote governance, had refused to accommodate the structure — which is why Alibaba listed in New York rather than its natural home market.
Alibaba's record-breaking public debut
May 2014Alibaba files F-1 registration statement with the SEC, disclosing intent to list in the U.S.
June 2014Alibaba selects NYSE over Nasdaq, partly influenced by Facebook's botched Nasdaq debut in 2012.
Sep 18, 2014IPO prices at $68/share, raising $21.8 billion — the largest U.S. IPO in history.
Sep 19, 2014Shares open at $92.70. Market cap reaches ~$231 billion on day one.
Nov 2019Alibaba completes secondary listing in Hong Kong, raising $12.9 billion.
The Empire Expands — and the Emperor Steps Back
The post-IPO years were a period of aggressive expansion. Alibaba invested in or acquired companies across logistics (Cainiao), entertainment (Youku Tudou), food delivery (Ele.me), brick-and-mortar retail (Sun Art, Intime), ride-hailing (a major Didi Chuxing shareholder), Southeast Asian e-commerce (Lazada), and artificial intelligence (SenseTime). By 2018, the Economist described Alibaba and Tencent as "China's most formidable investors," noting that ambitious startups felt compelled to accept capital from one or the other — or face being frozen out of the ecosystems that dominated Chinese digital life.
The investment strategy reflected a philosophy Ma had articulated repeatedly: Alibaba was not an e-commerce company but an infrastructure company. The goal was to build the commercial operating system for China — payments, logistics, cloud computing, data analytics, marketing — and then extend that system to the rest of Asia and eventually the world. Revenue grew from $8.5 billion in fiscal 2014 to nearly $23 billion in fiscal 2017, a 56% year-over-year increase. By fiscal 2020, it reached approximately $72 billion. Alibaba Cloud, launched in 2009, grew into the largest cloud computing provider in Asia.
Jack Ma, meanwhile, was gradually stepping back. In September 2018, he announced he would retire as executive chairman, handing the role to Daniel Zhang — the quiet, finance-trained executive who had conceived Singles' Day and managed Tmall's transformation into a revenue machine. Ma framed the transition as a sign of institutional maturity: "No company can rely on its founders forever." He devoted himself to philanthropy, education, and a growing public persona as a global business statesman — meeting world leaders, speaking at Davos, pledging to create one million American jobs in a meeting with President-elect
Donald Trump in January 2017.
The retirement seemed graceful. It was not.
The Speech, the Crackdown, and the Disappeared Founder
On October 24, 2020, Jack Ma delivered a speech at the Bund Finance Summit in Shanghai that would prove to be the most expensive monologue in corporate history. Standing before an audience of regulators and financial industry leaders, Ma attacked China's banking system as operating with a "pawnshop mentality," argued that Basel III capital requirements were ill-suited to digital finance, and implied that Chinese regulators were stifling innovation with antiquated thinking.
The speech was delivered twelve days before Ant Group's planned dual listing in Shanghai and Hong Kong — an IPO expected to raise over $34 billion at a valuation exceeding $300 billion, which would have been the largest public offering in world history. On November 3, 2020, Chinese regulators summoned Ma and Ant Group executives for a meeting. On November 5, the IPO was suspended.
What followed was a regulatory assault that redefined the relationship between China's technology giants and its government. Alibaba was hit with an antitrust fine of RMB 18.23 billion ($2.8 billion) in April 2021. Ant Group was forced into a comprehensive restructuring, converting itself into a financial holding company subject to bank-like capital requirements. Ma effectively disappeared from public life for months. His net worth, which had peaked above $60 billion, eventually fell by more than half.
The crackdown was not solely about Ma's speech — it was part of a broader regulatory campaign targeting Chinese tech companies across competition, data security, and "common prosperity" — but the speech was the catalyst, the moment when the founder's charisma, which had built an empire, became the liability that endangered it. Alibaba's stock, which had peaked above $300 per ADS in October 2020, began a decline that would eventually erase more than $500 billion in market capitalization.
Every great company is born in a winter.
Six Pieces of a Broken Mirror
In March 2023, newly installed CEO Daniel Zhang announced Alibaba's most radical structural change in its history: the company would split into six independent business groups — Taobao and Tmall Group (domestic commerce), Cloud
Intelligence Group, Alibaba International Digital Commerce Group, Cainiao Smart Logistics, Local Services Group, and Digital Media and Entertainment Group — each with its own CEO, board, and potential path to independent fundraising or IPO.
The restructuring was a concession to reality. Alibaba had become a conglomerate so vast that its parts were dragging down its whole. The market was valuing the company at less than the sum of its observable parts. The six-way split was designed to unlock value, sharpen accountability, and — perhaps most importantly — reduce the regulatory target on Alibaba's back. A company that was six companies was harder to accuse of monopolistic behavior than a single integrated behemoth.
But the execution proved chaotic. In June 2023, Daniel Zhang abruptly resigned as both group CEO and chairman, replaced by Eddie Wu as CEO and Joe Tsai as chairman. Wu — one of the original eighteen co-founders, the technologist who had built Taobao's monetization engine and Alipay's technical architecture — was a very different leader than Zhang. Where Zhang was a methodical operator who had risen through finance and strategy, Wu was an engineer-founder returning to direct the company he had helped build from Ma's apartment. Tsai, the Yale-educated dealmaker who had structured every major capital event in Alibaba's history, brought international credibility and financial discipline.
By December 2023, Wu had consolidated further, taking direct control of both the Taobao and Tmall Group and the Cloud Intelligence Group. The planned IPO of the cloud business was scrapped, partly due to U.S. restrictions on advanced chip exports to China. Several of the six groups' independent listing plans were quietly shelved. The grand restructuring was being restructured.
Eddie's leadership of both Alibaba Cloud and Taobao and Tmall Group will ensure total focus on, and significant and sustained investment in, our two core businesses of cloud computing and e-commerce.
The AI [Pivot](/mental-models/pivot) and the Long Game
By 2024 and into 2025, Alibaba had found a new narrative — or rather, the narrative had found it. Artificial intelligence, and specifically large language models, offered a strategic rationale for the cloud business that chip restrictions had nearly strangled. Alibaba's Qwen family of open-source AI models accumulated over 400 million downloads, making them among the most widely deployed AI models globally. The company pledged $53 billion in AI and cloud infrastructure investment over three years — a bet that open-source AI distribution could do for Alibaba Cloud what Android had done for Google: establish an ecosystem dependency that translated into cloud computing revenue.
The logic was characteristically Alibaba: give away the model, own the infrastructure. Developers trained on Qwen would build on Alibaba Cloud. Startups that adopted the open-source models would become Alibaba Cloud customers as they scaled. The strategy mirrored the original Taobao playbook — free at the application layer, monetize at the infrastructure layer — translated into a new technological era.
Whether it would work was genuinely uncertain. The Chinese AI market was intensely competitive — ByteDance, Baidu, Tencent, and a swarm of startups were all racing to build foundational models. U.S. chip export restrictions constrained access to the most advanced training hardware. And Alibaba's core e-commerce business faced relentless competition from Pinduoduo's Temu (which had overtaken Alibaba in market capitalization by late 2023), JD.com's logistics-driven model, and Douyin's (TikTok's Chinese parent) live-commerce juggernaut.
In January 2024, Jack Ma and Joe Tsai made a public show of confidence by purchasing approximately $200 million worth of Alibaba shares. The stock rallied. By early 2025, a broader AI-driven rally in Chinese tech stocks lifted Alibaba's valuation — the company gained $50 billion in market cap in a single surge tied to AI progress — but the structural questions remained. Alibaba was no longer a growth story. It was a turnaround story, a transformation story, and a geopolitical story, all at once.
The Apartment in Hangzhou
There is a detail from Porter Erisman's account of joining Alibaba in April 2000 that has the quality of a founding myth and the texture of lived experience. Walking through the company's new offices — shiny, recently upgraded from Ma's apartment — he discovered a room filled with twelve bunk beds. Engineers sleeping in shifts, round the clock, in a stuffy, dark room. The stench of bodies and ambition. When the new COO arrived, his first act was to remove the bunk beds. The company could not scale on sleeplessness.
Twenty-five years later, Alibaba employs roughly 200,000 people. Its platforms touch more than a billion consumers. Its AI models run on servers spread across Asia. Its founder, the former English teacher who couldn't get hired at KFC, sits somewhere outside the frame — present in the company's DNA, absent from its daily governance, his fortune halved, his influence a matter of speculation.
The bunk beds are long gone. The question is whether the spirit that filled them — the willingness to bet everything on an idea the world hadn't validated yet — survived the institutional success, the regulatory winter, and the long, unfinished process of figuring out what Alibaba becomes when it can no longer be what it was.
