Anthropic flags US government views as revenue risk in IPO filing

Anthropic / Wikimedia Commons (Public domain)

Anthropic flags US government views as revenue risk in IPO filing

The AI company's S-1 warns that federal restrictions and export controls on its top models could cause material revenue losses

Anthropic wants to go public. Its paperwork says one of its biggest risks is the government of the country it calls home.

The AI company’s S-1 filing for its initial public offering warns that actions by US authorities could hurt its revenue. That is an unusual line item for a company whose largest market is the United States.

What the filing actually says

The prospectus points to a February 2026 directive from the Trump administration. It asked federal agencies to stop using Anthropic’s AI models.

That directive arrived alongside a Department of Defense designation tied to supply-chain risks.

Things escalated in June 2026. The Commerce Department placed export controls on two of Anthropic’s advanced models, Mythos 5 and Fable 5.

Commerce cited security concerns and worries that foreign actors could misuse the models. Under the new rules, foreign access required licenses.

The practical result was severe. Anthropic had to temporarily disable those models worldwide.

The company’s filing does not sugarcoat the stakes. Anthropic acknowledged the measures could lead to potential:

Advertisement

“material revenue losses or business disruptions”

The numbers behind the warning

The risk matters because the business underneath it has grown very fast. Anthropic’s revenue reached nearly $4.6 billion in 2025, a 12-fold increase.

The run-rate figures climbed even faster. The company cited run-rate revenue of $14 billion in February 2026, a figure later surpassed at $47 billion by May 2026.

Geography is where the government risk gets sharper. The US market contributed approximately two-thirds of Anthropic’s sales.

Customer concentration is another pressure point. Two unnamed customers each accounted for 12% of revenue in 2025.

Then there is how Anthropic gets paid. About 83% of its income came from consumption-based revenue.

That detail links directly to the export controls. When Mythos 5 and Fable 5 went dark globally, usage on those models stopped, and so did the meter.

Small contracts, big shadow

Direct government business is a rounding error on Anthropic’s income statement. Government contracts represent less than 1% of annual revenue.

Among the notable 2026 direct awards: $200 million from the Department of Defense and approximately $19,000 from the State Department.

Anthropic’s filing flags concern about reputational damage. It also points to possible ripple effects on commercial partnerships.

Ambition on a tight rope

Anthropic is pursuing these plans at a large scale. The company is eyeing a valuation of around $2 trillion.

It is also carrying forward infrastructure commitments of roughly $518 billion. About 80% of those commitments are non-cancelable.

What this means for investors and the AI sector

For prospective IPO buyers, the filing offers an unusual tradeoff. The growth figures are the kind that sell offerings, while the risk factors describe a company whose top models have already been forced offline once by its own government.

Concentration adds a second layer of fragility. Two customers at 12% each, plus heavy reliance on consumption billing, mean shocks travel quickly to the income statement.

Government contracts account for less than 1% of revenue, yet government action may be the single most consequential variable in the prospectus.

Whether Commerce grants licenses that let Mythos 5 and Fable 5 serve foreign users again will shape near-term revenue. The DOD designation matters too, as any change to it could ease reputational pressure on commercial deals.

Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.
Anthropic flags US government views as revenue risk in IPO filing
Anthropic flags US government views as revenue risk in IPO filing

The AI company's S-1 warns that federal restrictions and export controls on its top models could cause material revenue losses

Anthropic / Wikimedia Commons (Public domain)

Anthropic wants to go public. Its paperwork says one of its biggest risks is the government of the country it calls home.

The AI company’s S-1 filing for its initial public offering warns that actions by US authorities could hurt its revenue. That is an unusual line item for a company whose largest market is the United States.

What the filing actually says

The prospectus points to a February 2026 directive from the Trump administration. It asked federal agencies to stop using Anthropic’s AI models.

That directive arrived alongside a Department of Defense designation tied to supply-chain risks.

Things escalated in June 2026. The Commerce Department placed export controls on two of Anthropic’s advanced models, Mythos 5 and Fable 5.

Commerce cited security concerns and worries that foreign actors could misuse the models. Under the new rules, foreign access required licenses.

The practical result was severe. Anthropic had to temporarily disable those models worldwide.

The company’s filing does not sugarcoat the stakes. Anthropic acknowledged the measures could lead to potential:

Advertisement

“material revenue losses or business disruptions”

The numbers behind the warning

The risk matters because the business underneath it has grown very fast. Anthropic’s revenue reached nearly $4.6 billion in 2025, a 12-fold increase.

The run-rate figures climbed even faster. The company cited run-rate revenue of $14 billion in February 2026, a figure later surpassed at $47 billion by May 2026.

Geography is where the government risk gets sharper. The US market contributed approximately two-thirds of Anthropic’s sales.

Customer concentration is another pressure point. Two unnamed customers each accounted for 12% of revenue in 2025.

Then there is how Anthropic gets paid. About 83% of its income came from consumption-based revenue.

That detail links directly to the export controls. When Mythos 5 and Fable 5 went dark globally, usage on those models stopped, and so did the meter.

Small contracts, big shadow

Direct government business is a rounding error on Anthropic’s income statement. Government contracts represent less than 1% of annual revenue.

Among the notable 2026 direct awards: $200 million from the Department of Defense and approximately $19,000 from the State Department.

Anthropic’s filing flags concern about reputational damage. It also points to possible ripple effects on commercial partnerships.

Ambition on a tight rope

Anthropic is pursuing these plans at a large scale. The company is eyeing a valuation of around $2 trillion.

It is also carrying forward infrastructure commitments of roughly $518 billion. About 80% of those commitments are non-cancelable.

What this means for investors and the AI sector

For prospective IPO buyers, the filing offers an unusual tradeoff. The growth figures are the kind that sell offerings, while the risk factors describe a company whose top models have already been forced offline once by its own government.

Concentration adds a second layer of fragility. Two customers at 12% each, plus heavy reliance on consumption billing, mean shocks travel quickly to the income statement.

Government contracts account for less than 1% of revenue, yet government action may be the single most consequential variable in the prospectus.

Whether Commerce grants licenses that let Mythos 5 and Fable 5 serve foreign users again will shape near-term revenue. The DOD designation matters too, as any change to it could ease reputational pressure on commercial deals.

Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.