High prices for high-end artificial intelligence could mean a price war just ahead of two more huge initial public offerings. Enthusiasm for Anthropic and OpenAI’s IPOs scheduled for later this year could cool if the companies start trying to undercut each other on prices to retain customers. Following SpaceX’s successful IPO earlier this month at a valuation of $1.75 trillion, Anthropic and OpenAI debut as publicly traded companies, the next milestone in the economy’s transition to AI and automation. Therefore, investors are closely monitoring price trends between the two Frontier models. “OpenAI is talking about lowering its prices,” Paul Meeks, head of technology research at Freedom Capital Markets, told CNBC earlier this month. “That could change, especially as[OpenAI and Anthropic]ramp up for IPOs, which could trigger at least a temporary price war.” The longtime analyst and investor believes big spending on AI infrastructure is still two to three years away, “until 2028.” Anthropic announced a $65 billion capital raise in May, valuing the company at $965 billion, and OpenAI closed a $122 billion funding round in March, valuing the company at $852 billion. Both companies have filed documents with the Securities and Exchange Commission to begin the IPO process. For now, investors are trying to avoid potential AI price wars by sticking with hardware manufacturers and service suppliers. Investors appear confident in investing further up the AI supply chain, including cloud services, infrastructure, and chips. AI “supercars” or “workhorses” OpenAI and Anthropic are widely recognized as having cutting-edge AI algorithms. But other cheaper models are closing in on the field. Although cheaper models don’t have all the high-end features, researchers say a fraction of the price is sufficient for most uses. As AI moves from a cutting-edge technology to a commodity, the popularity of lower-cost options could weigh on valuations. Deutsche Bank’s Jim Reed wrote in a note to clients last week that Anthropic’s cutting-edge AI model, Claude Fable 5, “does the same job for most (perhaps 90%) of the day-to-day tasks[DeepSeek’s V4-Pro of China]does for almost 1.5% of the cost.” Cheap AI models from DeepSeek, ChatGPT, MiMo, and MiniMax all have a weighted average cost per task of less than $0.13, while the top three versions of Anthropic’s Claude cost between $1.80 and $2.75, according to metrics from analytics website Artificial Analysis. “Employers who just need a reliable workhorse, not a supercar, will increasingly ask whether it’s worth paying the Frontier premium,” Reed wrote. Price Sensitivity In fact, AI customers are becoming increasingly price sensitive. Uber Chief Technology Officer Praveen Nepali Naga said in April that the company had used up its entire 2026 AI budget in the first four months alone. Andrew McDonald, the company’s head of operations, said last month that spending on AI was “getting harder to justify.” According to Forbes, Microsoft has restricted employees’ access to large language models and canceled Claude Code licenses to control costs. Executives from MetaPlatform, Salesforce and DoorDash also discussed the need to reduce access to AI and ensure it is used productively, The Wall Street Journal reported in May. Ahead of a potential price war between chip and infrastructure frontier AI models, a safer AI strategy for retail investors may simply lie with hardware manufacturers. Regardless of competition at the algorithmic level, overall demand for AI is skyrocketing, and unit costs are falling as more people use AI. That means long-term investments in chipmakers and providers of computing power may prove more reliable in the short term. “Hyperscalers remain impressively profitable, with positive early returns on investment. We expect operating cash flow from the hyperscalers to exceed $900 billion in 2027,” JPMorgan senior analyst Tarek Hamid said in a letter to clients last week. JPMorgan predicts that over the next five years, funding for various types of AI-specific chips will exceed $3 trillion. Concerns about OpenAI Beyond pricing, tech investors have broader concerns about the customer segments of top AI companies. Dan Niles, founder of Niles Investment Management, said last week that he believes OpenAI is caught between Alphabet’s dominance in consumer AI and Anthropic’s position in enterprise AI. “Are we going to let five different people win in the AI space? No, I think Google will win in the consumer space…We have Anthropic in the enterprise space,” he said. “I think OpenAI is caught between those two people. So, to me, those two are the winners. I think OpenAI is in trouble because it’s caught between the other two.”
