Renowned tech analyst and investor Mary Meeker has warned that leading U.S. AI companies, including OpenAI, may soon find themselves falling behind to faster and cheaper international competitors, particularly from China.
In a recent interview with the Financial Times, Meeker, whose insights have long been regarded as a compass for Silicon Valley, said AI is poised to create unprecedented economic value - but warned that these riches will not be concentrated in North America alone.
"Many $10 trillion companies will be born, and likely not all of them will be centered in North America," Meeker said.
"What we're getting into is extraordinary wealth creation. Never before has a market of five billion users been so easily accessible."
Growing challenges for U.S. AI giants
In her latest presentation, which has a wide following among investors, business owners, and technology leaders, Meeker highlights the growing challenges facing leading US companies such as OpenAI, Google, and Anthropic. Despite their dominance in the early stages of developing large language models (LLMs), Meeker warns that the escalating costs of training models, coupled with the emergence of leaner and more specialized alternatives, could cost them their early lead.
Among the most prominent competitors: China's DeepSeek, which made headlines by releasing powerful language models at a much lower cost. According to Meeker's report, "the business model is still in flux," warning that the economics of generic end-to-end models are "closer to that of commodity industries with an investment-grade capital consumption rate."
High costs and elusive profits
The costs of developing advanced models have increased by 2,400% over the past eight years, making the road to profitability even narrower for the big boys. OpenAI, xAI, and Anthropic have total estimated annual revenues of $12 billion, having raised $95 billion in funding.
However, Meeker sees OpenAI's valuation to revenue ratio as "high and expensive." The valuation comes as competition is mounting not only from US startups building customized lightweight models, but also from international players offering low-cost solutions that are rapidly becoming more efficient.
The gap between America and China is narrowing
The FT reports includes a chart showing a comparison of the performance of AI models in the US and China between January 2024 and February 2025, highlighting how quickly Chinese platforms have closed the gap, especially in aspects of cost efficiency and model speed.
Meanwhile, tech giants like Microsoft, Nvidia, and Meta are investing billions of dollars in AI infrastructure. According to Meeker, the collective capital expenditures of the six largest U.S. tech companies exceeded $200 billion in 2024. This investment rush has inflated the valuations of language modeling companies, with the combined value of the "Big Three" - OpenAI, Anthropic, and xAI - reaching nearly $400 billion.
History repeats itself... But at a price
Meeker, once dubbed the "Queen of the Internet" for her early analysis at Morgan Stanley in the 1990s, described today's AI investment climate as reminiscent of major ups and downs in tech history. She likened the financial trajectory of AI companies to cash-intensive rebels like Amazon, Uber, and Tesla, which suffered early losses before gaining market dominance - but only after a series of collapses and liquidations.
She also had a clear warning for investors who have been swept up in the AI frenzy: "The golden rule in these euphoria periods is to invest only what you can afford to lose, and to adopt a portfolio approach," Meeker told the FT. "Putting all your eggs in one basket is a real risk, because everything seems to be going up - until it stops going up."
Conclusion
As costs escalate and geopolitical competition intensifies, Meeker's warnings are a sobering reminder that the race for AI dominance is still in its infancy - and the winners may not be the ones in the lead today.
comments