OpenAI and Anthropic's growth is tied to infrastructure and regulations, with rising costs and regulatory risks impacting future plans.
OpenAI and Anthropic are experiencing rapid revenue growth and increased valuations, yet their future success is increasingly contingent upon securing adequate computing power, funding for data centers, and navigating regulatory landscapes. This was underscored by recent statements from both companies and various analyst assessments.
OpenAI's revenue is closely tied to its computing capacity, as highlighted by CFO Sara Fryer, who noted a substantial increase in capacity from 0.2 GW in 2023 to around 1.9 GW by 2025, alongside a rise in annual recurring revenue from $2 billion to over $20 billion. The company posits that computing is the most critical resource in AI and aims to enhance product adoption through improved infrastructure access. Furthermore, OpenAI has diversified its supplier base, moving away from dependence on a single provider.
Anthropic has announced similar growth trajectories, recently securing $65 billion in a Series H funding round, bringing its valuation to approximately $965 billion. The company reported that its annual revenue surpassed $47 billion in May, with plans to use the new funds for expanding computing capacity for its AI model, Claude.
Rising Infrastructure Costs
The endeavor to scale advanced AI models is becoming increasingly capital-intensive, affecting not just the developers but the entire data center supply chain, including energy and chip providers. Goldman Sachs Research estimates that the four leading hyperscalers—Meta, Microsoft, Amazon, and Alphabet—will collectively invest $5.3 trillion on capital expenditures from 2025 to 2030. However, OpenAI and Anthropic are not part of this major group.
Goldman Sachs anticipates that private markets, infrastructure funds, and real estate will play a more prominent role in financing data centers in the coming years. Additionally, the Bank for International Settlements has projected that the top five cloud infrastructure providers will spend over $1 trillion on AI-related capital expenditures in 2025-2026. This level of investment is already outpacing profits and free cash flow, prompting some companies to seek debt financing.
Competition for dominance in the AI sector could lead to over-investment in initiatives with uncertain returns, which may adversely affect funding and the overall investment cycle if revenue expectations are not met.
Regulatory Risks in Focus
Alongside rising costs, major AI firms face potential regulatory hurdles. In June, Anthropic temporarily suspended access to its Fable 5 and Mythos 5 models due to a U.S. government directive related to export controls, which lacked specific details on the security issues raised. Concurrently, it was reported that the Trump administration had advised OpenAI against releasing its GPT-5.6 model due to security concerns, leading the company to initially limit access to select clients.
On June 30, the Department of Commerce lifted restrictions on the models, allowing Anthropic to resume access. This incident highlights an ongoing tension between the company and U.S. authorities, which has been further complicated by the military's use of Claude in operations. CEO Dario Amodei expressed a reluctance to collaborate with the Pentagon, emphasizing a commitment to democratic values over military applications of their technology.
Future IPO Considerations
Despite these challenges, both OpenAI and Anthropic are eyeing public offerings. OpenAI filed a confidential registration statement on June 8, while Anthropic did so a week earlier. Both firms have indicated that details regarding the offerings remain to be determined.
Investors will be particularly focused on not just the companies' valuations but also the economics of scaling, including user growth costs, the return on investment for data centers, funding sustainability, and potential regulatory restrictions on flagship products. Additionally, a legislative proposal introduced by Republican Senators Tim Scott and Bill Hagerty aims to bolster U.S. information technology supply chains by allowing the Department of Commerce to restrict transactions involving technologies from entities linked to high-risk countries that threaten national security.