Hardware

J.P. Morgan Flags Red Flags in AI Market

J.P. Morgan warns that 42 AI companies in the S&P 500 generated 65 to 80 percent of the index's profits, revenues, and investments since ChatGPT launched in 2022.

Close-up of a blue screen error shown on a data center control terminal.

Photo: panumas nikhomkhai / Pexels

J.P. Morgan has raised concerns about the AI market, citing signs of investor exuberance and potential risks. The bank notes that since the launch of ChatGPT in 2022, just 42 AI companies within the S&P 500 have driven roughly 65 to 80 percent of the entire index's profits, revenues, and investments. This concentration of gains has sparked warnings about the market's vulnerability to overvaluation and speculative bubbles. The bank draws parallels to the dotcom bubble, highlighting technical patterns in the semiconductor rally that mirror past speculative excesses. Hedge funds are heavily invested in chip and hardware stocks, margin lending on the Korean stock exchange is climbing, and retail traders are increasingly piling into semiconductor options. Leveraged chip ETFs have quintupled their influence on global stock markets since early 2024, with gains concentrated in a small number of companies. The ten largest US stocks now account for about 40 percent of the S&P 500's market cap, up from 17 percent in 2015. Despite this increase, the US still ranks among the markets with relatively low stock market concentration, with only India and Japan being less concentrated. Nvidia remains the dominant player in the AI accelerator market, but its share is expected to drop from 85 percent in 2023 to 75 percent by 2026. Custom chips from major cloud providers like Google's TPUs or Amazon's Trainium offer cost savings of 30 to 40 percent compared to Nvidia GPUs. Anthropic, for example, has committed to running its AI Claude on Amazon's Trainium for the next decade. J.P. Morgan also flags risks around revenue at leading AI labs like OpenAI and Anthropic, noting their rapid sales growth but massive compute costs and uncertain future profitability. Rising token prices could push companies to switch to cheaper open-source models, with early signs of this shift already emerging. Companies are shifting tasks to cheaper models, average token prices are falling, and Chinese open-source models are approaching top-tier performance at a fraction of the cost. Tech investment's share of economic growth is also rising, while free cash flow margins at major cloud providers are shrinking and their debt financing is growing. All told, J.P. Morgan says AI is creating multiple layers of concentration risk across markets, infrastructure, and the broader economy. NYU finance professor Aswath Damodaran has warned the same thing, saying an AI crash could hit harder than the dotcom bust.

Source: thedecoder

Key points

Since ChatGPT launched in 2022, 42 AI companies in the S&P 500 have driven roughly 65 to 80 percent of the entire index's profits, revenues, and investments.
The ten largest US stocks now account for about 40 percent of the S&P 500's market cap, up from 17 percent in 2015.
Nvidia's share of the AI accelerator market is expected to drop from 85 percent in 2023 to 75 percent by 2026.
Custom chips from major cloud providers like Google's TPUs or Amazon's Trainium cut operating costs by 30 to 40 percent compared to Nvidia GPUs.
Anthropic has committed to running its AI Claude on Amazon's Trainium for the next decade.
Companies are shifting tasks to cheaper models, average token prices are falling, and Chinese open-source models are approaching top-tier performance at a fraction of the cost.
Tech investment's share of economic growth is rising, while free cash flow margins at major cloud providers are shrinking and their debt financing is growing.

Source: The Decoder Read the original →

WRITTEN BY

Sam Bergstrom

AI Infrastructure & Hardware

Sam specializes in AI chips, data centers, and training infrastructure.