Amazon

The scariest risk section of 1997 preceded the greatest IPO investment of all time: up 355,140%, with the all-time low still above the offer price.

Amazon priced 3 million shares at $18.00 on May 15, 1997 -- $0.075 per share after 29 years of splits (2:1 in 1998, 3:1 and 2:1 in 1999, 20:1 in 2022). Barnes & Noble, the incumbent named throughout the competition risk factor, sued Amazon on May 12, 1997, three days before pricing, over the 'world's largest bookstore' claim; the suit settled out of court and Amazon kept saying it. The stock closed its first day up 33.3% at a split-adjusted $0.10 and ended May at $0.08 -- 6.7% above the offer price, a level it would never see again. That May 1997 close remains the lowest monthly close in the stock's entire history: an IPO buyer who simply held was never meaningfully underwater for 29 years.

Every promise in the risk factors came true first. The losses did 'increase significantly from current levels' as pledged, funded by repeated returns to the capital markets, exactly as the financing risk warned. Then the market-conditions risk factor -- the one warning that technology price-earnings ratios were 'at or near historical highs' -- detonated. From a monthly close of $4.30 in early 1999, through the Barron's 'Amazon.bomb' era and analyst predictions that the company would run out of cash, the stock collapsed roughly 93% to $0.30 by August 2001 (about 95% from its intraday peak), while the Nasdaq fell 78% from its March 2000 top. Amazon was routinely grouped with the corpses of the crash. Yet even at that bottom, the stock stood at four times its split-adjusted IPO price -- the crash wiped out bubble buyers, not IPO buyers.

The turn came almost immediately after the obituaries. In the fourth quarter of 2001 Amazon reported its first profit -- one cent per share on over $1 billion in revenue -- vindicating the S-1's core claim that profitability was a scale problem, not a model problem. Then the 'new business areas' risk factor, framed in 1997 as a danger, revealed itself as the entire thesis: music in 1998, the third-party marketplace, Prime in February 2005, Amazon S3 in March 2006 and EC2 in August 2006, the Kindle in November 2007, and acquisitions from Zappos ($850 million, 2009) to Whole Foods ($13.7 billion, 2017) and MGM ($8.45 billion, 2022). The internally built infrastructure that the S-1 flagged as a liability -- homegrown systems, no redundancy -- became AWS, arguably the most valuable business ever incubated inside a retailer.

The scoreboard after 29 years: 2025 revenue of $716.9 billion against the $15.7 million in the prospectus -- growth of roughly 45,000-fold -- with $79.98 billion in operating income and 1,576,000 employees against the 151 disclosed in 1997. The stock closed at $266.43 on August 28, 2026, up 355,140% from the IPO price, within sight of its June 2026 all-time monthly high of $271.58. The company that told investors it would 'incur substantial operating losses for the foreseeable future' became one of the most profitable enterprises on earth; the bookseller that feared Barnes & Noble outlived the question of whether online commerce would 'become a viable commercial marketplace' by making it the default way humans shop.

An online-only bookseller with $15.7 million in 1996 sales and 151 employees proposes to spend its way to retail scale, betting that a store with no stores -- minimal inventory, millions of titles, rapid fulfillment from two wholesalers -- can out-economize the superstore chains, if online commerce itself becomes a real consumer habit. The company is explicit that the price of this bet is deepening losses for the foreseeable future.

The bull case is that Amazon.com gets big before anyone else can. Net sales grew from $511,000 in 1995 to $15.7 million in 1996, driven by a swelling customer base and repeat purchases, and the company is the recognized leader in online bookselling at the exact moment the Web is becoming a mass medium. Books are the perfect wedge product -- standardized, easy to ship, with a catalog of millions of titles no physical store can stock -- and the virtual model carries structural advantages: minimal inventory, no retail real estate, and payables that fund growth. If the company invests the IPO proceeds into brand, technology and infrastructure fast enough, it builds a customer base and brand franchise that Barnes & Noble cannot claw back even with greater resources, achieves economies of scale on razor-thin margins, and then extends the same playbook into music, video and whatever else can be sold over a Web browser. A founder who owns 43 percent, pays himself $64,333, holds no options and personally guaranteed the company's credit cards is unusually aligned with that long game.

With hindsight: The bull case was that Amazon would out-scale Barnes & Noble in books and maybe extend into music and video. What happened exceeded that framing by orders of magnitude: the brand-plus-scale-plus-customer-experience formula the S-1 articulated conquered not just bookselling but general retail, then logistics, devices, streaming, advertising and cloud computing. Revenue went from $15.7 million to $716.9 billion (2025), and the 'economies of scale' the filing promised eventually produced $79.98 billion in annual operating income. Even the most aggressive 1997 bull was wrong only in imagining too small.

• Success will come from brand, customer experience, and scale economics rather than near-term profit.

  “The Company believes that its success will depend in large part on its ability to (i) extend its brand position, (ii) provide its customers with outstanding value and a superior shopping experience, and (iii) achieve sufficient sales volume to realize economies of scale.” source ↗

• Heavy deliberate investment in marketing, technology and infrastructure -- with prices cut to win customers -- will buy the market position that eventually pays for everything.

  “Accordingly, the Company intends to invest heavily in marketing and promotion, site development and technology and operating infrastructure development. The Company also intends to offer attractive pricing programs, which will reduce its gross margins.” source ↗

• The bookstore is a wedge: the same platform can expand into new products, formats and businesses.

  “The Company may choose to expand its operations by developing new Web sites, promoting new or complementary products or sales formats, expanding the breadth and depth of products and services offered or expanding its market presence through relationships with third parties.” source ↗

1. Losses will deepen for the foreseeable future -- by design

   came true

   profitability · structural

   Amazon does not merely warn that it might lose money; it promises to, and promises the losses will accelerate. With a $6.0 million accumulated deficit, a $5.8 million net loss in 1996, and a stated plan to spend ahead of revenue while cutting prices, profitability is deferred indefinitely and depends entirely on reaching enormous scale.

   “the Company believes that it will incur substantial operating losses for the foreseeable future, and that the rate at which such losses will be incurred will increase significantly from current levels” source ↗

   What happened: Exactly as promised: losses accelerated for years after the IPO and the company became the poster child for profitless internet retail. But the S-1's framing -- that profitability was a matter of 'substantially increased revenue levels' -- proved correct: the first quarterly profit arrived in Q4 2001 (one cent per share on over $1 billion of revenue), and by 2025 Amazon earned $79.98 billion in operating income. The risk materialized in full and turned out not to matter.

2. Barnes & Noble and better-funded giants are coming

   partly came true

   competition · structural

   Barriers to entry online are minimal, and the incumbents have every visible advantage: brand, sales volume, publisher relationships, and capital. B&N and Borders have both announced online pushes, B&N has an AOL distribution deal already live, and B&N has sent Amazon notice of a legal claim. AOL and Microsoft could channel customers to competitors or charge tolls for access.

   “Both B&N and Borders have announced their intention to devote substantial resources to online commerce in the near future and B&N, specifically, has entered into a relationship with AOL through which, beginning in March 1997, it offers a broad selection of titles at discounted prices.” source ↗

   What happened: The competitors arrived; the damage never did. B&N sued Amazon three days before the IPO (settled out of court), launched barnesandnoble.com with heavy spending, and Borders followed -- and both lost catastrophically. Borders eventually liquidated and B&N shrank into a niche chain while Amazon became the dominant force in Western retail. The deeper version of the risk -- giants like Microsoft controlling access -- eventually rhymed with real competition from Walmart, Google and Microsoft (in cloud), but none of it produced the feared 'diminished brand franchise.'

3. Online commerce may simply never catch on

   didn't happen

   market · structural

   The entire business presumes consumers will change how they shop. In 1997 that is an open question: the Internet may fail as a commercial marketplace through inadequate infrastructure, security fears, congestion, or plain consumer indifference. Amazon's fate is chained to the adoption curve of a two-year-old medium.

   “there can be no assurance that acceptance and use will continue to develop or that a sufficiently broad base of consumers will adopt, and continue to use, the Internet and other online services as a medium of commerce” source ↗

   What happened: The single biggest uncertainty in the filing resolved as completely as any risk in IPO history. Consumers did adopt the Internet as a medium of commerce -- Amazon itself, with $716.9 billion in 2025 revenue and Prime (launched February 2005) in a large share of households, did more than any company to make it happen. The 1997 questions about infrastructure, security and consumer willingness now read like dispatches from another century.

4. Thin bookselling margins demand enormous scale

   partly came true

   profitability · structural

   Gross margin was roughly 22% in 1996 and the company is deliberately pushing it lower with 40%-off discounting. At those economics, the planned spending only pencils if revenue multiplies many times over -- there is no profitable small version of this company.

   “Because the Company has relatively low product gross margins, achieving profitability given planned investment levels depends upon the Company's ability to generate and sustain substantially increased revenue levels.” source ↗

   What happened: Retail margins stayed brutally thin for decades -- the 'substantially increased revenue levels' the S-1 demanded took until Q4 2001 to produce even a penny of profit. What the filing could not foresee is that the durable margin engine would come from outside retail entirely: AWS (S3 in March 2006, EC2 in August 2006), third-party marketplace fees and advertising now generate the bulk of the $79.98 billion in 2025 operating income. The scale problem was real; the solution was a business the S-1 never imagined.

5. Everything depends on Jeff Bezos

   didn't happen

   key person · serious

   The company singles out one person by name. There is no long-term employment agreement, no key-person insurance, and the same man is founder, CEO, chairman, 43% owner, and until recently the personal guarantor of the company's credit. Losing him -- or key engineers the company admits it struggles to hire -- would be a material blow.

   “The Company's performance is substantially dependent on the continued services and on the performance of its senior management and other key personnel, particularly Jeffrey P. Bezos, its President, Chief Executive Officer and Chairman of the Board.” source ↗

   What happened: Bezos stayed, and stayed, and stayed -- running the company as CEO for over two decades of the very hypergrowth the filing worried about, then executing an orderly handoff to Andy Jassy, the AWS lieutenant, while remaining executive chairman. The key-person risk that lacked even a life-insurance policy in 1997 resolved into one of the longest and most successful founder tenures in business history.

6. Hypergrowth run by a four-month-old management team

   didn't happen

   operations · serious

   Headcount grew 14-fold in one year, most senior managers are brand new, some have never operated at a public company, and the company still needs to hire a CIO. Systems, controls and people must all be built while the rocket is in flight.

   “From January 1, 1996 to December 31, 1996, the Company expanded from 11 to 151 employees. The majority of the Company's senior management joined the Company within the last four months, and some officers have no prior senior management experience at public companies.” source ↗

   What happened: The company that grew from 11 to 151 employees in 1996 and warned it might not manage further growth scaled to 1,576,000 employees by 2025 -- a ten-thousand-fold expansion of the workforce -- while operational excellence in fulfillment and logistics became one of its deepest moats. There were strains (warehouse labor disputes, executive churn in later eras), but the core fear, that growth would break the company, never came true.

7. Ingram controls the shelf: supplier concentration with no contracts

   didn't happen

   operations · serious

   The 'store with no inventory' works only because two wholesalers hold the inventory instead. Ingram alone supplied 59% of purchases in 1996, with no long-term contracts guaranteeing merchandise, terms, or credit. A squeeze from either vendor -- or a vendor aligning with a competitor -- hits the heart of the model.

   “The Company purchases a substantial majority of its products from two major vendors, Ingram Book Group ("Ingram") and Baker & Taylor, Inc. ("B&T"). Ingram is the single largest supplier and accounted for 59% of the Company's inventory purchases in 1996.” source ↗

   What happened: The dependence dissolved as Amazon built exactly what the 1997 filing hinted it might: its own warehouses, then one of the largest logistics networks on the planet, with its own air cargo, trucking and last-mile delivery. Ingram never squeezed; instead, the power balance inverted so far that publishers and suppliers came to negotiate with Amazon from weakness. The risk describes a company that stopped existing within a few years of the IPO.

8. One building in Seattle, no backup, no disaster plan

   didn't happen

   operations · serious

   Substantially all computer and communications hardware sits in a single leased facility exposed to fire, flood, earthquake and break-ins. The company that exists only as a website has no redundant systems and insufficient business-interruption insurance.

   “The Company does not presently have redundant systems or a formal disaster recovery plan and does not carry sufficient business interruption insurance to compensate it for losses that may occur.” source ↗

   What happened: No fire, flood or earthquake ever took Amazon down in a business-ending way, and the exposure closed quickly as the company built out data centers and fulfillment sites. The ironic coda: the company with 'no redundant systems' in 1997 became, through AWS (launched 2006), the vendor that sells redundancy, availability zones and disaster recovery to most of the internet.

9. Homegrown systems that can't close the books

   didn't happen

   tech & security · serious

   Amazon built everything itself -- site, search, order management, payments -- optimized for order volume, not accounting. The transaction system isn't integrated with the financial systems, and management concedes it may be difficult to produce accurate, timely financial statements. That is a striking admission for a company going public.

   “the Company's current management information system, which produces frequent operational reports, is inefficient with respect to traditional accounting-oriented reporting and requires a significant amount of manual effort to prepare information for financial and accounting reporting” source ↗

   What happened: The accounting-integration problem was fixed without incident -- no restatements or reporting scandals defined the company. The larger arc runs the other way: the internally developed transaction and infrastructure technology the S-1 treated as a fragile liability compounded into the engineering culture that produced the marketplace platform, Kindle (2007) and AWS, the most profitable outgrowth of any 'internal IT system' ever.

10. New ventures could strain the company and damage the brand

    partly came true

    growth · serious

    The company openly reserves the right to launch new sites, new product lines and acquisitions -- and warns that failed expansions would burn cash, stretch management, and tarnish the Amazon.com name it is spending millions to build.

    “Furthermore, any new business or Web site launched by the Company that was not favorably received by consumers could damage the Company's reputation or the Amazon.com brand.” source ↗

    What happened: Individual ventures did fail -- Auctions, the Fire Phone and many quiet shutdowns burned real money -- so the letter of the risk materialized repeatedly. But the feared consequence, damage to the Amazon.com brand, never stuck, and expansion into new business areas turned out to be the entire company: music (1998), marketplace, Prime (2005), AWS (2006), Kindle (2007), Zappos, Whole Foods, MGM. The S-1 listed as a risk the thing that became the reason to own the stock.

11. More capital will likely be needed, on unknowable terms

    came true

    financing · serious

    The offering proceeds are earmarked partly to fund operating losses, and if losses outrun the cash, Amazon must raise again with no assurance the window will be open. A company promising accelerating losses is structurally dependent on capital-market goodwill.

    “If such net proceeds, together with cash generated by operations, are insufficient to fund future operating losses, the Company may be required to raise additional funds. There can be no assurance that such financing will be available in amounts or on terms acceptable to the Company, if at all.” source ↗

    What happened: Amazon did need much more capital, returning repeatedly to the markets and leaning heavily on convertible debt during the boom -- including a raise completed shortly before the window slammed shut in 2000. During the crash, prominent analysts predicted the company would run out of cash, and survival genuinely depended on that pre-crash war chest plus the working-capital float. The risk materialized in full; the company threaded it, and Q4 2001's first profit ended the existential phase.

12. The sales-tax advantage may not last

    came true

    regulation · serious

    Amazon collects sales tax only on shipments into Washington. Part of its price advantage over physical stores is a tax arbitrage that states have every incentive to close, and any new warehouse outside Washington widens the exposure.

    “A successful assertion by one or more states or any foreign country that the Company should collect sales or other similar taxes on the sale of merchandise could have a material adverse effect on the Company's business, prospects, financial condition and results of operations.” source ↗

    What happened: The states came, as predicted -- through years of 'Amazon tax' statutes, physical-presence fights as the fulfillment network spread into new states, and ultimately a legal regime that made online sales-tax collection universal. Amazon now collects sales tax nationwide. But the arbitrage lasted long enough to matter: by the time it closed, Amazon's advantage rested on Prime, selection and logistics, not on tax-free pricing.

13. The Bezos family holds absolute control

    came true

    governance · serious

    Post-offering, Bezos holds about 43% and his family another 10% -- a combined majority of voting power, without any dual-class structure. Public shareholders cannot elect a director, amend the charter, or approve a merger over the family's objection. Whatever Bezos wants to do with their money, he can.

    “The above persons and entities will hold an aggregate of approximately 53% of the outstanding voting power of the Company immediately upon completion of this offering.” source ↗

    What happened: The control block was real and decisive -- and it is precisely what let Amazon ignore Wall Street's demands for profits through the losses of 1997-2001 and reinvest through the decades after. Public shareholders who accepted powerlessness were paid 355,140% for it. Bezos's stake diluted over time (stock sales funded Blue Origin, among other things) and no dual-class structure preserved formal control, but by then the market had long since stopped wanting to overrule him.

14. Tech stocks are priced at historic highs -- including this one

    came true

    market · serious

    In an almost self-aware passage, the prospectus warns that Internet and technology valuations sit at or near record highs on price-earnings ratios far above historical levels, and that there is no assurance those multiples hold. Buyers of a profitless bookseller at $18 are warned they are buying into a hot market.

    “The trading prices of many technology companies' stocks are at or near historical highs and reflect price earnings ratios substantially above historical levels. There can be no assurance that these trading prices and price earnings ratios will be sustained.” source ↗

    What happened: The most prescient passage in the filing. The unsustainable price-earnings ratios it warned about kept inflating for three more years, then collapsed: the Nasdaq fell 78% from its March 2000 peak of 5,048 to 1,114 in October 2002, and Amazon's monthly close dropped roughly 93% from $4.30 in early 1999 to $0.30 in August 2001 -- about 95% from the intraday peak. The warning was correct in every particular, and an IPO holder who rode straight through it was still up four-fold at the bottom.

How the S-1 reads

This is the ur-document of the 'get big fast' playbook that a generation of consumer-internet S-1s would imitate: deliberate, accelerating losses framed as investment; brand and scale named as the only things that matter; margins sacrificed for customers; a founder with near-absolute control and token salary asking public investors to fund a land grab. What almost none of the imitators copied is the candor -- Amazon's S-1 argues against its own hype, warning that its growth rate is unsustainable and that tech valuations are at unsustainable highs. Also notable as a period piece: the existential risks of 1997 (will people ever type credit card numbers into a browser? can the Internet's infrastructure handle growth?) and a risk section short enough to actually read, roughly a tenth the length of a modern filing.

• “In July 1994, Jeffrey P. Bezos, the President, Chief Executive Officer and Chairman of the Board of the Company, purchased 10,200,000 shares of Common Stock for an aggregate price of $10,000.” source ↗

  The founder bought control of what he would build for ten thousand dollars.

• “Since July 1995, Mr. Bezos has personally guaranteed the obligations of the Company under a bankcard merchant account with Wells Fargo Bank. Since April 1995, Mr. Bezos has personally guaranteed company credit cards.” source ↗

  The CEO was personally underwriting the company's ability to accept credit cards.

• “Although the Company has experienced significant revenue growth in recent periods, such growth rates are not sustainable and will decrease in the future.” source ↗

  A rare anti-hype disclosure: the company talks down its own growth rate in its own prospectus.

• “The Company is not currently aware of any legal proceedings pending against it. The Company has received notice from B&N of an alleged claim.” source ↗

  Buried in the trademark section: the country's largest bookseller has already come knocking.

What this one teaches

• A terrifying risk section measures a company's honesty, not its odds. Amazon disclosed accelerating losses, supplier dependence, no disaster recovery and a hot-market valuation -- and returned 355,140%. The scariest disclosure and the best outcome in the archive belong to the same filing.
• 'Losses for the foreseeable future' is a question about what the losses buy, not about viability. Amazon's losses purchased customers, brand and infrastructure with durable value; the S-1 even told you the test to apply -- whether scale economics eventually arrive -- and in Q4 2001 they did.
• Named incumbents are usually the least dangerous risk on the list. Barnes & Noble had the brand, the capital and the publisher relationships, sued three days before the IPO, and lost everything that mattered. Structural advantage plus speed beats resources.
• The upside never appears in an S-1. The filing's risk factors flagged homegrown systems and new-business expansion as dangers; they became AWS and the marketplace -- the two engines of nearly all the value ultimately created. Risk sections enumerate downside; optionality goes undisclosed.
• Even a perfect investment routes through catastrophe. Amazon fell roughly 93% between 1999 and 2001 while analysts predicted bankruptcy -- yet the all-time trough was still above the split-adjusted IPO price. The drawdown shook out everyone who bought the hype instead of the filing.
• Founder control at IPO, disclosed as a governance risk, was the mechanism that made the long game possible: a 53% family voting block, a $64,333 salary and no options meant no one could force short-term profits -- and no one needed to.

The paper trail

1. 1997-03-24 S-1 filing index ↗ document ↗
2. 1997-04-21 S-1/A filing index ↗
3. 1997-04-29 S-1/A filing index ↗
4. 1997-05-09 S-1/A filing index ↗
5. 1997-05-13 S-1/A filing index ↗
6. 1997-05-14 S-1/A filing index ↗
7. 1997-05-14 S-1/A filing index ↗
8. 1997-05-15 424B1 filing index ↗ document ↗

Filed as AMAZON COM INC. All documents are public domain, served by SEC EDGAR.