Anthropic's Revenue Surged to $16.3 Billion in the First Half of 2026, Up From $1.1 Billion a Year Ago

Anthropic could IPO as early as mid-November, and prospective investors have been eagerly awaiting its S-1. While that document still isn't publicly available, The Wall Street Journal recently reported seeing disclosures from a draft IPO prospectus, many of which haven't been publicly reported before.

The disclosures provide the best look to date at Anthropic's financials and cover the many risk factors the AI lab has flagged.

Image source: Getty Images.

Anthropic's financials over the first half of 2026

Anthropic saw significant sales growth over the six months ended June 30, generating around $16.3 billion in revenue, up from $1.1 billion a year prior. The revenue is impressive, but it also comes with massive computing demand, as the AI lab spent nearly $9 billion on computing capacity.

Computing costs will only accelerate from here. Anthropic has committed to around $450 billion in computing agreements over the next seven to 10 years, most of them noncancelable, and it has an obligation of more than $160 billion to lease Alphabet's Google-designed chips in a deal with Broadcom. The disclosures also revealed that Anthropic is negotiating additional computing deals, and commitments could rise by up to $150 billion. Total commitments for the next decade already top $600 billion and could rise to more than $750 billion.

Anthropic also recorded a net loss of $80.7 billion over H1 2026. While that number will receive considerable attention, it merits a disclaimer. Most of the loss, $72 billion, came from accounting charges tied to the value of convertible notes and other fundraising instruments. The higher Anthropic's valuation goes, the bigger these paper losses become. In a way, the loss is a by-product of Anthropic's success.

Anthropic highlighted some serious risks

About a third of the draft prospectus covered risk factors, and these ranged from standard issues to concerns that would have seemed unbelievable a decade ago.

Of Anthropic's $16.3 billion in revenue, around one-fifth came from just two customers, and the AI lab expects a significant portion of revenue to come from a "limited number of customers for the foreseeable future." Anthropic also highlighted data center backlash as a potential problem, given its computing needs, and this is an issue affecting many AI companies right now.

Another risk factor was that misaligned models could cause real-world harm, and that Anthropic may be liable. It pointed out that AI models can exhibit self-preserving behaviors and generate responses that resemble blackmail to avoid being shut down by humans. The prospectus even mentioned that Anthropic may not be able to obtain sufficient insurance to cover "harm that could result from a catastrophic AI-related incident." Even AI labs can't predict with complete certainty what their models could be capable of, and that creates liability risks investors haven't seen at this scale before.

What to watch

The next step for Anthropic is to file a public S-1, which must be published at least 15 days before a roadshow begins. That will give prospective investors a detailed look at the company's audited financial results.

If you're thinking of investing in Anthropic when it goes public, consider the final IPO valuation. The company is reportedly targeting a valuation of up to $2 trillion, which means it will already trade at a premium. Annualizing its H1 2026 revenue would mean a valuation of over 60 times trailing sales. It may be worth waiting to see if the price comes down before investing, or starting with a small position.

Lyle Daly has positions in Alphabet and Broadcom. The Motley Fool has positions in and recommends Alphabet and Broadcom. The Motley Fool has a disclosure policy.