Anthropic delays IPO to November 2026
Anthropic has moved its planned initial public offering from October to November 2026. Advisors state the company intends to release strong third-quarter figures before listing. Reports from The Information and The Wall Street Journal suggest the earliest now is late October, though November remains the likely window.
That reasoning does not fully align with the company’s recent performance. Second-quarter results appear sufficient to support a launch. A public listing would also allow new revenue to lift share prices. Investors anticipate a valuation near $2 trillion, with a capital raise reaching $100 billion. Those figures would surpass the records set by SpaceX during its June offering.
Revenue more than doubled between the first and second quarters of 2026. Income rose from roughly $4.7 billion to more than $11.5 billion, according to The Information. Spending on computing infrastructure increased by 65 percent over the same period, moving from $3.4 billion to $5.6 billion. The company’s adjusted operating margin improved from a negative 13 percent in the first quarter to a positive single-digit percentage in the second, though that metric carries caveats.
Business spending is growing fast, but Anthropic still needs cash
By the end of 2025, approximately 1,500 companies had each spent more than $100,000 on Anthropic’s AI software over the previous 12 months. By the end of the second quarter of 2026, that number had quadrupled to about 6,000. Over that same 12-month window, more than 100 companies each spent over $10 million, while more than 1,000 each spent over $1 million. Anthropic reportedly projects revenue of roughly $190 billion to $200 billion for 2028.
That money would not go straight into investors’ pockets. The IPO would mainly give Anthropic access to debt and equity markets for the years ahead. Despite reporting $120 billion to $130 billion in cash in early August, the company will likely need more funding as data center spending and model training drain its reserves. The computing infrastructure deal it signed with SpaceX in May alone costs $1.25 billion a month.
Competition, costs, and cybersecurity risks complicate the IPO
Several factors could explain the delay, and the answer is probably a combination of them. Anthropic recently claimed to be profitable, but its figures excluded major costs such as stock-based compensation and did not follow standard accounting rules.
At the same time, OpenAI has regained ground thanks to Astra and pulled ahead on the AI model platform OpenRouter. Both companies need to sustain rapid growth over the coming months to justify their massive data center buildouts. Rising interest rates are making those facilities more expensive to build, while cheaper open-weight models are adding competitive pressure.
Another concern is the safety debate over AI models deemed too dangerous. Their cybersecurity capabilities have already led to unintended hacks during safety tests involving OpenAI as well as Google, Anthropic, and Meta. Those incidents suggest substantial operational risks as long as the resulting damage isn’t covered by insurance. So far, these incidents have not been treated as attacks or led to sanctions or lawsuits. They have largely been dismissed as an AI novelty, but that response is unlikely to hold up once the company goes public and faces greater scrutiny.
OpenAI has also pushed its IPO back to 2027. Sam Altman cited safety concerns, arguing that an AGI company has good reasons to stay private through such sweeping changes. But given the other pressures both companies face, that is only part of the story at best.
The biggest question remains whether all that spending on AI pays off. How much more revenue are companies generating, or how much are they saving, compared with having people do the same work? Those returns remain hard to measure.
