$97.17 ▲ 17.07%
Updated: 23 hours ago
ANT
What it does. Anthropic, PBC is an AI research and deployment company best known for Claude, a family of large language models sold as a consumer assistant, a developer API, and an enterprise platform. Its product surface has widened fast in 2026 — coding agents (Claude Code), enterprise security tooling, and a recent push into productivity software (sheets, meetings, whiteboards) that puts it in direct contact with Microsoft and Google workspace franchises.
How it makes money. Three main streams: consumer and team subscriptions, usage-based API revenue (sold directly and through Amazon Bedrock, Google Cloud and Microsoft Foundry), and large enterprise contracts. Revenue is consumption-driven, which scales beautifully when customers deploy agents at volume but is also re-priced downward every time a cheaper model generation lands.
Business quality. The franchise is genuinely strong — a leading position in coding and enterprise agents, deep hyperscaler distribution, and strategic investors including Nvidia and Google — but the economics are not yet a business in the conventional sense. Reported 2026 revenue is in the mid-teens of billions, against a disclosed net loss of roughly $104bn for the first nine months of 2026, a figure dominated by compute commitments and IPO-related charges rather than ordinary operations. The negative forward multiple (-844.96x) simply confirms the market is underwriting a profit stream that does not exist yet.
The listing itself. Anthropic debuted on the Nasdaq on 19 November 2026 under ticker ANT in the largest IPO in US history, pricing at $83 and raising roughly $81bn. Shares opened at $111.50, faded through the session, and the stored quote shows $92.60 after hours — up 17.07%, but below the 52-week low of $92.72 and well off the $115.10 high set on day one. Market cap stands at $1.27 trillion.
Notable challenges. Capital intensity is the central issue: frontier training and inference require multi-year compute commitments that must be funded regardless of demand. Competition is relentless and well-capitalised, model-level differentiation is short-lived, and the cloud partners who distribute Claude also compete with it. Safety and regulation add a tail risk few listed companies carry — the company's own amended S-1A flagged that Claude may soon meet its internal threshold for recursive self-improvement, an unusual disclosure that cuts both ways with investors and regulators.