The clearest signal that the AI investment arms race is nearing a peak comes not from the headlines, but from the balance sheets.
According to Christopher Wood, global head of equity strategy at Jefferies, the scale of spending by US hyperscalers has reached a point where they are consuming an increasingly larger share of their cash flow, particularly on chips and memory. Based on the company’s latest forecast, capex as a percentage of operating cash flow for the top four U.S. hyperscalers increased from 41% in 2023 to a projected 92% in 2026.
A significant portion of this is dedicated solely to memory, which is expected to account for around 30% of total investments, implying that around 28% of operating cash flow will be absorbed by memory investments this year, it said in its report. Greed and fear report.
This growing intensity of investment highlights a more fundamental issue: monetization. A recent report from Jefferies led by Edison Lee highlights that the challenges related to AI business models remain underestimated. The increasing cost of remaining competitive, driven by higher compute, memory and energy requirements, suggests that sustainable profitability for pure players in the AI model remains distant.
The wood aligns with this view. His base case is that AI may ultimately resemble a capital-intensive industry like airlines, rather than the high-margin, winner-takes-all dynamics seen in the Internet age.
Despite this, the current phase of spending shows no signs of slowing down. Large technology companies continue to implement aggressive investment plans. Microsoft plans to spend $190 billion this year, with about $25 billion attributed to rising component costs. Alphabet and Meta both raised their 2026 investment forecasts to $180-190 billion and $125-145 billion, respectively, while Amazon kept its forecast at $200 billion.
Of these, investor concerns appear most pronounced in the case of Meta, which does not enjoy the same direct cloud benefits from AI spending as peers like Alphabet, Microsoft, and Amazon.
For now, the game of picks and shovels remains intact, supported by continued spending and limited investor reluctance on returns.
However, the first signs of tension are beginning to surface. A recent report noted that OpenAI fell short of its internal goals for user and revenue growth, including its goal of reaching 1 billion weekly active users for ChatGPT by the end of last year. The company also reportedly missed several monthly revenue targets in 2026, while facing increased competition.
Market share trends reflect this change. Over the past 12 months through March, Gemini’s share of web traffic in the generative AI market increased sharply, from 6% to 25.5%, while ChatGPT’s share declined from 77.4% to 56.7%, according to SimilarWeb data.
At the same time, concerns have been raised about funding structures within the ecosystem, where partners such as Nvidia and Oracle fund OpenAI, which in turn uses that capital to buy compute from them.
Competition is also intensifying. Anthropic reported in early April that its annualized revenue run rate had surpassed $30 billion, up from around $9 billion at the end of 2025, now surpassing OpenAI’s reported revenue run rate of more than $25 billion in February.
Overall, the picture that emerges is one of escalating investment, growing competitive pressure and unresolved questions about returns. The spending cycle continues, but the pressures it puts on cash flow and the uncertainty surrounding monetization are becoming increasingly difficult to ignore.
(Disclaimer: The recommendations, suggestions, views and opinions expressed by the experts are their own. These do not represent the views of The Economic Times.)
