OpenAI and Anthropic are rapidly growing revenue and private valuations, but their further expansion increasingly depends on access to compute, data center financing, and regulatory decisions, according to company statements and analyst reports.
According to a notice from Sam Altman’s company, OpenAI, the firm’s revenue closely tracks available compute. CFO Sarah Friar said compute capacity rose from 0.2 GW in 2023 to about 1.9 GW in 2025, while annual recurring revenue increased from $2 billion to more than $20 billion.
The ChatGPT developer calls compute the scarcest resource in AI and says greater access to infrastructure would accelerate product adoption and monetization. The company also said it moved from reliance on a single provider to a “diversified ecosystem” of suppliers. The post did not disclose profitability metrics.
Anthropic makes a similar case. On May 28, the company announced it had raised $65 billion in a Series H round at a $965 billion valuation. According to Anthropic, annualized revenue based on the current sales pace exceeded $47 billion in May. The funds will also go toward expanding compute capacity for Claude.
Infrastructure costs are rising
Scaling advanced AI models is becoming capital-intensive not only for developers but across the entire chain of data center, energy, chip, and cloud infrastructure suppliers.
Analysts at Goldman Sachs Research estimate that the four largest hyperscalers — Meta, Microsoft, Amazon, and Alphabet — will spend $5.3 trillion on capital expenditures in 2025–2030. These companies build and lease the infrastructure underpinning the AI sector, but OpenAI and Anthropic are not included in the sample.
Goldman Sachs expects private markets, infrastructure funds, and real estate to play a growing role in financing data centers.
The Bank for International Settlements in its 2026 annual economic report warned that the five largest cloud infrastructure providers will spend more than $1 trillion on AI-related capital expenditures in 2025–2026. The organization noted these commitments already outpace the companies’ profits and free cash flow, prompting some to tap debt financing.
The report says that competition for leadership in the sector could lead to overinvestment in projects with uncertain returns. If monetization expectations are not met, this could trigger a sharp pullback in financing and hit the entire investment cycle.
Regulatory risk is now part of the infrastructure
Beyond compute costs, major AI companies face the risk of sudden government restrictions.
In June, Anthropic said it had halted access to the models Fable 5 and Mythos 5 due to a U.S. government directive under export controls. The company said the government’s letter did not include specific details about the safety issue and linked the concerns to a possible bypass of Fable 5’s safeguards.
That same month, media reported that the Donald Trump administration asked OpenAI not to release GPT-5.6 to the general public immediately over safety concerns. OpenAI will first provide the model to a limited number of customers.
On June 30, the Department of Commerce lifted the restrictions on the models, after which Anthropic restored access. The episode became part of the company’s ongoing conflict with U.S. authorities. In February 2026, the U.S. Army used Claude in an operation to capture Venezuelan President Nicolás Maduro, and Defense Secretary Pete Hegseth called the developer “a supply chain risk.”
Startup CEO Dario Amodei said Anthropic would rather not work with the Pentagon than agree to uses of its technology that could “undermine, rather than protect, democratic values.”
IPO plans remain in the background
Even so, OpenAI and Anthropic continue to explore public listings. OpenAI said it filed confidentially on June 8, while Anthropic did so a week earlier. Both companies emphasized that offering terms have not been determined.
For investors, the key questions go beyond valuation to the economics of scaling: the cost of supporting user growth, data center payback periods, the durability of financing channels, and whether a regulator could restrict access to a flagship product.
In July, Republican senators Tim Scott and Bill Hagerty introduced a bill to protect U.S. information and communications technology supply chains. The initiative would allow the U.S. Department of Commerce to prohibit transactions involving technologies and services if they are developed, manufactured, or supplied by persons linked to risk countries and pose a threat to critical infrastructure, the digital economy, or national security.
