Anthropic's Filing Shows $16.3bn Revenue and $80bn Half-Year Loss

Anthropic booked $16.3bn in revenue and lost $80bn in the first half of 2026, according to its S-1 filing — a document that could break records for a public-market debut. Much of that loss came from a $72.4bn one-off charge. Investors have valued the five-year-old firm between $1.5tn and $2tn, though the filing itself lists no target valuation.

The numbers are stark on both sides of the ledger. In 2025, Anthropic brought in $4.585bn in revenue and lost roughly $33.9bn, per the filing. Founded in 2021, the company has posted explosive growth every year of its existence. It has also burned cash at a matching pace: securing computing power accounts for the vast majority of its costs.

Where does the money go?

Anthropic spent nearly $9bn on AI infrastructure in the first half of this year. That's up from $2.3bn in the same period last year — nearly a fourfold jump in twelve months.

Unlike OpenAI, Anthropic has made clear its customers are largely other businesses. "Frontier AI is most valuable in real-world settings, as deployment across enterprises is where economic impact is greatest," company executives wrote in the filing. The concentration cuts both ways. Anthropic said its top 5 customers account for 25% of its business, which the company itself flags as a risk.

What does Anthropic think AI can do?

Executives described the technology's potential in sweeping terms, calling it "'a country of geniuses in a datacenter' – millions of virtual agents with the ability to revolutionize every area of human endeavor: software, physics, math, biology, medicine, finance, economics, commerce, robotics, military and intelligence capabilities."

That mention of military capabilities is notable. Anthropic remains in a standoff with the Trump administration after refusing to loosen restrictions on how the Department of Defense could use Claude.

The filing highlights four opportunity categories:

• Acceleration of progress in biology and medicine
• Economic growth
• Spreading that growth globally
• Streamlining government services, including social services and medical programming

Anthropic said it already works with several pharmaceutical and biotechnology companies and has improved clinical trial processes, "cutting the time it takes to write clinical trial reports from months to minutes." Executives also claim biological discoveries made with Claude, including a "molecular machine" with the "potential to be developed into a novel biotechnology tool" — one cited use case being gene editing.

What about the safety caveats?

Anthropic breaks AI's societal risks into four categories: misuse by bad actors, loss of control, authoritarian abuse and economic disruption. The company details its guardrails, then concedes they may not hold. Its most capable models can provide "meaningful assistance to malicious actors seeking to develop chemical, biological, or other dangerous weapons," executives wrote. The safeguards, they add, "are incomplete, may be outdated in light of evolving threat models, have been and may in the future again be circumvented, or may fail to operate as intended."

The filing discloses what Anthropic believes is the first documented large-scale cyber-attack with little human intervention. The company says a Chinese state-sponsored hacking group manipulated a Claude code tool into attempting infiltration of roughly 30 global targets, including large technology companies, financial institutions, chemical manufacturers and government agencies.

On authoritarian abuse, Anthropic offered less of a defense. It said AI may help authoritarian governments "deepen repression" and build "high-tech surveillance states," and committed to working with "free societies to provide a counterweight to authoritarian abuse and to promote freedom and democracy." The firm fell out with the Trump administration after refusing to allow its models to be used for domestic surveillance or autonomous lethal weapons.

Why do the risks run 92 pages?

The filing lays out 92 pages of business risks. Anthropic ties much of its success to retaining top executives, outlasting Google and OpenAI, navigating AI and datacenter regulation, and surviving public fallout if its models are misused — say, to create bioweapons. Its enterprise focus means a departing top buyer could hit revenue hard.

Its structure as a public benefit corporation, which balances financial interests with a stated social good, may make short- and medium-term financial goals harder to achieve, the company said. It must report progress on public benefit goals the same way it reports financials.

CEO Dario Amodei has called for a pause on AI development to ensure firms act responsibly and avoid catastrophe. The filing adds that insurance for AI liabilities is still evolving: "We may be unable to obtain coverage for certain AI-related liabilities on commercially reasonable terms, or at all, leaving us exposed to potentially significant uninsured losses."

The company's AI infrastructure is concentrated in the US, leaving it exposed to regulatory changes, power grid constraints, natural disasters and local opposition. The filing specifically flags data center moratoriums already passed or under consideration. "We are expanding our data center capacity internationally as part of our efforts to diversify our infrastructure footprint and reduce concentration risks," it reads. Anthropic also cited US immigration policy — including Trump administration targeting of H-1B visas — as a threat to recruiting the talent it calls central to its value proposition.