
On the morning of November 19, Anthropic priced its IPO at $83, above the top of its marketed range, raising approximately $81.3 billion and setting a new global IPO record. Shares opened at $111.50, up 34%, and briefly reached $115.10. The Nasdaq Composite subsequently climbed more than 2%, which the market interpreted as a strong signal that the “AI trade remains healthy.” Jake Behan, head of capital markets at Direxion, captured the prevailing mood: first came Nvidia’s stronger-than-expected results, followed by traders’ willingness to chase Anthropic above its offer price. “The market isn’t ready to give up on this trade.”

But after the opening, events failed to follow the script of renewed faith in AI. Anthropic retreated steadily from around $115, closing at $97.17 and trimming its gain to 17%. The broader market reversed almost simultaneously: the Nasdaq swung from a gain of more than 2% to a loss of 2.25%, while approximately $1.3 trillion in market value was wiped off the S&P 500 that day—a figure the article describes as closely matching the amount raised in Anthropic’s IPO. All 11 S&P 500 sectors declined, more than 90% of its constituents closed lower, and the VIX briefly climbed above 29, its highest intraday level since March. Why did a successful IPO become a catalyst for such a broad market sell-off?
The IPO Itself Was Draining Market Liquidity
The $81.3 billion public offering, combined with approximately $26 billion in private placements to strategic investors, brought total fundraising above $100 billion. Reports from trading desks suggested that market participants sold gold, copper, cryptocurrencies and AI proxy assets such as Nvidia to fund their subscriptions. Market moves were consistent with this account: gold fell below $3,800, copper plunged approximately 5%, and cryptocurrencies declined across the board.
This was less a question of whether to buy Anthropic than an arithmetic problem of where the money would come from. When the market needs to assemble more than $100 billion for a single asset on the same day, investors sell assets that are easiest to turn into cash, regardless of whether they have turned bearish on them. AI stocks happened to meet two conditions: strong liquidity and substantial accumulated gains available to realise.
On the same day, short seller Hunterbrook released a report targeting AMD, alleging that it had provided chip firmware updates to blacklisted companies. AMD fell more than 9%, dragging down the broader semiconductor sector. This separate negative development intensified the AI sell-off against the liquidity vacuum created by the IPO. But it could not, by itself, explain the broader market’s reversal from a 2% gain to a 2% loss. What truly accelerated the decline was the market’s technical structure.
Negative Gamma and Options Expiry Mechanically Amplified the Decline
Nomura strategist Charlie McElligott had already warned that the combination of Nvidia’s earnings, Anthropic’s listing and monthly options expiry had pushed market makers’ negative gamma exposure to its highest level since 2019. When market makers are in a negative gamma position, falling prices force them to sell to hedge their exposure. That selling drives prices lower still, creating a self-reinforcing cycle. Heavy flows into 0DTE, or zero-days-to-expiration, options that afternoon further overwhelmed support from longer-dated positions.
This meant that when Anthropic retreated from its high and triggered the first wave of selling, market makers’ hedging activity replicated and amplified the pressure. A significant portion of the broader market’s decline was therefore generated mechanically by positioning and technical hedging dynamics, rather than by new fundamental information.
Goldman Sachs analyst Rich Privorotsky offered another perspective in a warning issued that morning: institutions’ net exposure to technology stocks stood at approximately 18%, well below the sector’s 29% index weighting and close to the most underweight positioning in five years. In theory, this created the potential for a positioning-driven squeeze higher as investors rebuilt their exposure, provided the market avoided a sharp decline. However, the mechanical selling pressure generated by negative gamma and options expiry overwhelmed that potential buying force.
More Than Half of Institutions Received No Shares, While Retail Investors Took a Larger Allocation
Less visible than the liquidity drain and negative gamma was Anthropic’s allocation structure. According to The Wall Street Journal and other media reports, approximately 1,200 institutional investors submitted orders, but more than half received no allocation. Around 15% of the offering was allocated to retail investors. The geographical distribution was heavily skewed towards the United States, with many European orders receiving no shares. A more consequential interpretation came from ZeroHedge: either the underwriters were so confident in demand that they deliberately restricted institutional allocations, or institutions were unwilling to buy at $83 and the shares were ultimately placed with less price-sensitive retail investors. The market leaned towards the latter explanation.
This ownership structure shaped the stock’s behaviour after listing. Institutions without shares had little incentive to support the price after the opening or buy on dips. Retail investors held a relatively large share of the tradable stock, and their trading behaviour was more susceptible to sentiment and short-term price movements. Once the opening surge triggered profit-taking, the absence of institutional buying support caused prices to lose their footing quickly. Anthropic’s closing price of $97.17 was approximately 13% below its opening price of $111.50 and around 16% below its intraday high of $115.10. The size of that retreat itself reflected the fragility of the ownership structure.
The Fed Also Signalled Tighter Scrutiny of AI Debt That Day
Minutes from the Federal Reserve’s October meeting were released on November 19. They showed that almost all officials believed further interest rate increases were still needed. More significantly, the Fed explicitly identified the opacity of AI-related debt, off-balance-sheet lease commitments and circular financing arrangements as financial stability concerns in an official document for the first time. Staff warned that if returns on AI investment fell short of expectations, off-balance-sheet lease commitments and circular financing arrangements among AI companies could amplify losses. Most participants observed that AI-related debt was growing rapidly, with a substantial portion raised through complex and opaque financing structures.
The implications were far more specific than simply rising expectations of further rate increases. Previously, the market had valued AI through questions of demand strength, order volumes and growth rates. The Fed minutes introduced another dimension: whether the financing behind that demand was transparent, whether off-balance-sheet commitments would amplify losses if returns disappointed, and whether circular financing was creating artificial revenue.
Anthropic’s own prospectus provided a telling illustration of these concerns: revenue of $33.2 billion in the first nine months, an operating loss of $1.26 billion, a net loss of $104 billion—primarily from changes in the fair value of convertible notes and derivative liabilities—and computing capacity purchase commitments totalling $900.9 billion. With the Fed identifying opaque AI debt as a financial stability issue in an official document, the market faced a systemic question extending beyond whether Anthropic could become profitable: could the financing model supporting the entire AI infrastructure sector withstand formal scrutiny in a higher-for-longer interest rate environment?