OpenAI has dedicated over $1.4 trillion to expand its data center infrastructure over the next 8 years. A recent report from the Wall Street Journal, citing internal company documents, indicates that it is estimated to incur a loss of $74 billion in 2028 alone.
In contrast, Anthropic, an AI startup operating with much less fanfare than Sam Altman’s company, is projected to break even in the same year. The documents suggest divergent paths to profitability for the two companies. Anthropic, valued at nearly $200 billion after a recent funding round, has been quietly building a subscriber base of corporate clients.
According to a recent WSJ report, approximately 80% of Anthropic’s revenue comes from its business customers, totaling over 300,000 clients. OpenAI claims to have around one million enterprise subscribers and over seven million ChatGPT for Work “seats,” a figure disclosed shortly after Anthropic’s data became public.
However, as per the Journal’s findings, OpenAI’s profit margins on subscriptions are lower compared to those of Anthropic. Given this scenario, one might expect OpenAI to manage its funds conservatively until a clear path to profitability emerges.
Instead, the company is heavily investing in expanding data centers and acquiring chips for model training and operation. It is also spending money liberally by focusing on attracting users to its ecosystem with hopes of monetizing them later on.
For instance, Forbes reported earlier this week that Sora 2, OpenAI’s video generation model causing copyright concerns due to user-generated content, costs about $15 million daily. This translates to an annualized expense of $5 billion.
OpenAI’s strategy appears reminiscent of the ZIRP era in Big Tech when firms operated at a loss due to negligible interest rates. Companies were willing to lower prices significantly to attract a broad audience before becoming the dominant player and being able to switch gears towards profitability.
While OpenAI believes in the feasibility of its approach, it remains uncertain whether the path they are on will lead them there. Internal documents indicate that the company anticipates turning profitable by 2030 as revenues grow rapidly. This timeline extends beyond its current commitment to developing costly data centers.
On its journey towards achieving profitability, OpenAI is expected to spend approximately 14 times more than Anthropic. The risk for companies like Anthropic operating in the AI sector lies in OpenAI potentially affecting their operations negatively if it faces financial challenges.
Given OpenAI’s significant financial obligations and interconnections within the AI industry, its failure could trigger a chain reaction impacting various players in the sector. It comes as no surprise that OpenAI is seeking government assurances regarding its expenditure plans.