AI trade stagnates as token spending falls 49% and compute costs cool

  • Jurrien Timmer said the AI trade has produced no gains for more than three months.
  • 49%: Silicon Data LLM token expenditure index fell over 50 days, Fidelity data show.
  • Global memory chip shortages are expected to continue until 2028.

The artificial intelligence trade has delivered little return for more than three months, according to Jurrien Timmer, Director of Global Macro at Fidelity Investments. Fidelity data show the Silicon Data LLM token expenditure index fell 49% over 50 days, while rental rates for Nvidia H100 and A100 GPUs declined after earlier gains of 56% and 16%, respectively. Timmer said token spending and GPU lease rates are flat to down, leaving memory, or DRAM, as the main rising cost in the AI buildout. A global memory chip shortage is expected to keep pressure on costs until 2028. Timmer described the investment cycle as inflationary, with an uncertain return for companies spending trillions on computing capacity. Corporate demand for debt and equity is rising at a $3.3 trillion pace, while U.S. hyperscalers are projected to spend about $916 billion on capital expenditures over the next 12 months. Combined 2027 capital spending by Alphabet, Amazon, Microsoft, Meta Platforms, Oracle and Space Exploration Technologies is expected to exceed $1 trillion. Timmer nevertheless said AI could eventually change daily life and create a productivity surge that raises the economy’s non-inflationary speed limit. Ark Invest’s Cathie Wood projects the technology could have a greater economic impact than the Industrial Revolution, with real GDP growth reaching 7% or more within five years. The Invesco QQQ Trust closed Tuesday up 0.93% at $747.46 and gained 0.12% in extended trading. The iShares PHLX SOX Semiconductor Sector Index Fund rose 2.56% to $572.78 and added 0.15% after hours. Benzinga’s Edge rankings gave QQQ a Momentum score in the 78th percentile, with positive short-, medium- and long-term price trends.

本网站上的信息是使用AI生成的,我们无法保证其准确性。 请仅作为参考信息使用。