OpenAI vs. Anthropic—How Do the Financials of the "Strongest AI" Actually Look?

Two of Silicon Valley's hottest AI unicorns are racing toward IPOs within the year, but a rarely disclosed set of financial data reveals a shared dilemma: the astronomical computing costs of training AI models are devouring the profitability of both companies. According to financial documents recently obtained by The Wall Street Journal, OpenAI expects its computing expenditure to reach $121 billion by 2028. Even if revenue nearly doubles by then, the net loss for that year will still be as high as $85 billion—a figure that would surpass the loss records of almost all publicly listed companies in history. Meanwhile, while Anthropic's projected spending is far lower than OpenAI's, even its most optimistic forecasts reflect a landscape of continuously rising computing costs. Additionally, according to a Bloomberg report on Tuesday, Anthropic's latest annualized revenue (Run Rate) has surpassed $30 billion, a significant leap from $9 billion at the end of 2025. The financial data from both companies jointly outline the reality of the AI arms race: while revenue is growing rapidly, the cost of training is expanding at an equally staggering pace, and the path to profitability remains long. For investors interested in participating in their IPOs, this financial picture displays both incredible growth potential and clearly marked risks. Revenue Surging, but Losses Are Equally Shocking Both OpenAI and Anthropic expect their revenue to more than double this year, with the primary driver being the accelerated adoption of AI tools by enterprise customers. OpenAI's revenue sources include consumer subscriptions, enterprise services, and new products (including hardware); Anthropic relies almost entirely on enterprise customers and includes sales generated through cloud partners in its revenue. However, behind the impressive revenue growth are equally shocking losses. OpenAI expects that even with significant revenue growth by 2028, it will still incur a net loss of $85 billion that year. The company is not expected to achieve overall break-even until 2030, while Anthropic expects to reach that milestone earlier. It is worth noting that both companies utilize a dual-track profit disclosure method: excluding "research computing" expenses, OpenAI is expected to achieve a small pre-tax operating profit this year, as is Anthropic under its most optimistic scenario; however, once training costs are factored in, both are suffering losses. The Computing Arms Race: Cost Is the Biggest Variable The runaway cost of training is the core pressure point in the financial structures of both companies. The computing power required to improve the intelligence level of each generation of models far exceeds that of the previous one, and both companies are currently releasing new model iterations at an unprecedented frequency. OpenAI expects its computing expenditure for AI research to reach $121 billion by 2028. In comparison, Anthropic's training expenditure is smaller in scale, but its financial forecasts similarly show a trend of continuously rising computing costs. Inference costs (the expenses incurred from processing user queries) also constitute a major burden. Currently, inference costs account for over 50% of revenue for both companies, though this proportion is expected to gradually decrease as technical efficiency improves. Paid users account for only a tiny fraction of ChatGPT's base, meaning a massive amount of inference costs cannot be covered by revenue. On the cash flow level, both companies will continue to consume large amounts of cash over the coming years, with IPO fundraising seen as a critical source of capital to sustain business operations. Anthropic’s Annualized Revenue Surpasses 30 Billion, Secures New Computing Allies According to Bloomberg, Anthropic's annualized revenue has surpassed $30 billion, more than tripling from $9 billion at the end of 2025. Currently, more than 1,000 enterprise customers spend over $1 million annually on the Anthropic platform, a figure that has doubled since February of this year. To support this growth momentum, Anthropic has signed major computing cooperation agreements with Broadcom and Google. According to a filing by Broadcom on Monday, the three parties will expand their strategic cooperation, granting Anthropic access to approximately 3.5 gigawatts of computing resources starting in 2027. Broadcom is developing chips based on Google's Tensor Processing Units (TPUs) to provide an alternative to Nvidia, and the parties have signed a long-term supply assurance agreement extending to 2031. Anthropic CFO Krishna Rao stated that the partnership with Broadcom and Google will help the company build the "computing foundation necessary to serve its significantly growing customer base." Following the announcement, Broadcom's stock price rose by as much as 3.6% in after-hours trading. Furthermore, to accommodate the potentially record-breaking IPO financing needs of both companies, Wall Street is seeking to push the limits of existing rules. Bankers are lobbying major index providers to relax inclusion criteria, and Nasdaq recently announced it will allow newly listed companies to join its indices more quickly, thereby accessing a broader pool of capital. OpenAI has stated that the company is currently prioritizing growth over profit, noting that while it could cut training expenditure, it expects such investments to yield substantial returns.