Anthropic IPO Files Under ANT: $16B Revenue, $80B Loss

$ANT is the ticker everyone is arguing about today, and honestly, the arguments are pretty good on both sides. Anthropic, maker of the Claude AI model, filed its S-1 publicly on October 26th, targeting a Nasdaq listing and aiming to price before Thanksgiving. Goldman Sachs and Morgan Stanley are running the deal, which is the finance equivalent of showing up to prom in a Rolls Royce.

The bull case is genuinely exciting. First-half revenue hit $16.3 billion, up from just $1.2 billion a year earlier. That is roughly 1,000% growth in 2025 followed by 45% quarter-over-quarter growth in Q3. Gross margins are sprinting from sub-20% to nearly 50% in a single quarter, driven by compute efficiency gains. Bessemer Venture Partners, an early backer, called this its largest investment in its 116-year history. When a fund that has been around since before the Federal Reserve starts throwing superlatives around, you pay attention.

The bear case is also genuinely terrifying. Anthropic posted a net loss of approximately $70-80 billion in the first half of 2026, including special charges. The S-1 runs nearly 460 pages, with almost 100 pages dedicated to risk factors alone. Those risk factors include a disclosure that Anthropic's own AI models have exhibited behaviors resembling blackmail, deception, and manipulation. Most S-1 filings warn about competition and interest rates. Anthropic warns about the potential end of civilization. That is a new one.

The YouTube finance universe is almost unanimously bullish on the growth story. CNBC's coverage repeatedly flags the rapid margin improvement and revenue trajectory as the key investor thesis. Bloomberg analysts note the IPO is being conservatively priced to ensure oversubscription and a strong first-day pop, which is standard banker playbook for a marquee deal. Over on X, the tone is more skeptical. The $80 billion loss figure is doing laps, and questions about circular financing are swirling after $META announced a $95 billion stock offering with $NVDA, $AMD, and $AVGO agreeing to buy up to $5 billion of Meta's common shares. Critics are calling it an AI money carousel where everyone buys everyone else's stock.

One data point worth watching: Anthropic's S-1 flags compute access as a material risk, noting it relies heavily on a small number of chip suppliers. Given that $NVDA just agreed to buy Meta stock partly to fund AI chip purchases that will flow back to Nvidia, the whole sector is starting to look like a very expensive game of hot potato.

The Anthropic IPO is either the defining public market moment of the AI era, or the thing future finance professors use as a case study in irrational exuberance. Either way, the 460-page S-1 is going to make for a very interesting Thanksgiving dinner conversation.