The world's two highest-valued AI startups are adopting starkly contrasting strategies for business growth;
OpenAI anticipates its sales profit margin will remain lower than Anthropic’s over the next five years. However, its investments in AI technological infrastructure such as chips and data centers far exceed those of Anthropic, as it seeks to attract top research talent through more generous stock option incentives.
Documents obtained by industry insiders reveal that the financial conditions of two major artificial intelligence startups in Silicon Valley are showcasing markedly different development trajectories amid the current AI boom: Anthropic is on track to achieve profitability faster than its competitor, OpenAI.
Documents show that Anthropic, leveraging the robust capabilities of its Claude chatbot in areas like programming, has seen a steady rise in enterprise users and is projected to break even for the first time by 2028.
In contrast, OpenAI, which continues to aggressively 'burn cash,' expects its operational losses for that year to swell to approximately USD 74 billion—equivalent to three-quarters of its revenue at that time—mainly due to skyrocketing computing costs. Previously, OpenAI had estimated that its cash burn would be about 14 times that of Anthropic before achieving profitability in 2030 (two years later than Anthropic).
Different 'playbooks' of the two AI giants
These two financial roadmaps, disclosed to investors earlier this year, indicate that the world's two most highly valued AI startups are pursuing markedly different business growth strategies. OpenAI expects its sales profit margin to remain lower than Anthropic’s over the next five years, but its investments in AI technological infrastructure such as chips and data centers significantly surpass those of Anthropic, while it attempts to attract top research talent with more lucrative stock options.
This aggressive plan also reflects the grand vision of OpenAI CEO Sam Altman: to transform OpenAI into a tech giant worth trillions of dollars, dictate the pace of the artificial intelligence wave, and demonstrate an almost limitless risk appetite—a strategy that requires continuous financing to sustain the startup’s operations. If market confidence in AI technology or its profitability wavers, it could backfire.
In fact, concerns over whether the scale of AI spending and related infrastructure development can generate sufficient revenue recently prompted some investors to sell off shares in major U.S. tech companies.
The above financial data from OpenAI emerged prior to the company signing a series of new computing agreements with several cloud computing and chip giants—indicating its expenditures in the coming years may climb even further. As Sam Altman recently revealed on the X platform, these agreements will saddle OpenAI with committed expenditures of up to USD 1.4 trillion over the next eight years, sparking concerns among industry skeptics and some investors about its solvency.
By contrast, documents show that Anthropic is adopting a more cautious strategy, with its cost growth rate more closely aligned with revenue growth. The company is focused on increasing sales to enterprise clients, which currently account for approximately 80% of its total revenue, and has avoided costly explorations in image and video generation—fields that require significantly higher computational power. Additionally, Anthropic's AI models have gained relatively greater popularity among programmers.
Notably, Anthropic’s founding is also closely linked to OpenAI—the company was established four years ago by former Google researcher Dario Amodei, who left OpenAI after a disagreement with Altman. When ChatGPT went viral, this startup was initially caught off guard—the product gave OpenAI a massive user base and a significant lead in sales.
However, Anthropic soon identified its development path and shifted its focus to selling its Claude chatbot to enterprises. Its valuation has recently surged to $183 billion, while OpenAI’s current valuation stands at approximately $500 billion.
Which approach is better?
Currently, nearly all major investors in Silicon Valley hold stakes in both companies, anticipating participation in what could become the largest IPO in tech history. The world’s three largest cloud service giants have also closely tied their growth to these two startups:$Microsoft (MSFT.US)$is OpenAI’s largest cloud service provider,$Amazon (AMZN.US)$and$Alphabet-A (GOOGL.US)$provides cloud support for Anthropic.
Private companies often disclose rapidly growing revenue figures, but other financial details tend to remain confidential—because these numbers are usually far less impressive than the revenue figures. This practice is particularly common among AI developers, who are reluctant to reveal their staggering cash burn rates. However, some industry outlets have recently disclosed partial financial metrics of the two companies.
Documents show that OpenAI expects to consume $9 billion in cash this year, with sales reaching $13 billion; Anthropic, meanwhile, anticipates burning nearly $3 billion in cash on the basis of $42 billion in sales—both companies’ cash burn rates amount to approximately 70% of their revenues.
However, Anthropic’s operational efficiency will improve significantly thereafter. By 2026, Anthropic expects its cash burn to drop to about one-third of its revenue, while OpenAI’s burn rate during the same period will remain at 57%. In 2027, Anthropic’s cash burn rate is projected to further decline to 9%, while OpenAI’s will stay unchanged.
Of course, Anthropic continuing to 'burn cash' even as it moves toward profitability does not necessarily imply that OpenAI’s current trajectory has any significant issues. In fact, if product demand continues to surge, OpenAI’s substantial early investments (particularly in new chips and data centers) may yield substantial returns. The company recently launched the highly impactful video application Sora and the browser Atlas, while actively developing new consumer-grade hardware devices, a ChatGPT e-commerce advertising feature, and humanoid robots.
A spokesperson for OpenAI recently stated, 'The current demand for artificial intelligence has exceeded the supply capacity of existing computing resources. Every dollar we invest in AI infrastructure is being used to serve hundreds of millions of consumers, businesses, and developers who rely on ChatGPT to enhance productivity.'
Documents show that the company is spending nearly $100 billion to expand data center spare capacity in anticipation of unexpected demand that may arise from future products and research. Its reserved computing power for new artificial intelligence research far exceeds that of Anthropic.
Altman also recently posted on the X platform, 'We believe the risk of insufficient computing power facing OpenAI is far more pronounced and likely to occur than the risk of excess computing power.'
Editor/melody
