AI Investment Trade Has Stalled

Capital expenditure for artificial intelligence remains high, yet the market has seen stagnant growth for three months.

Updated on Sept. 23, 2026 in Artificial Intelligence

Isometric editorial illustration showing a complex industrial array of cooling pipes and modular conduits, representing AI infrastructure investment. — Fidelity Investments strategist Jurrien Timmer described the current AI trade as dead money as capital expenditure growth for artificial intelligence stalls. AI Illustration. Upload story photo >

Live Poll

Do you believe now is a good time to invest in companies heavily focused on artificial intelligence?

Fidelity Investments strategist Jurrien Timmer has characterized the current artificial intelligence trade as dead money, noting that the sector has remained flat for over three months. This assessment arrives as the Silicon Data LLM token expenditure index fell 49% on a 50-day basis.

Why it matters

The current AI buildout remains highly inflationary with no clear consensus on returns for investing companies. Hyperscalers continue to pour massive capital into the technology, even as indicators suggest a cooling in previous momentum.

The Silicon Data LLM token expenditure index has seen a 49% decline, while GPU rental rates for H100 and A100 chips have dropped after previously peaking. U.S. hyperscalers are currently on pace to deploy $916 billion in capital expenditures over the next 12 months.

The players

Fidelity Investments

A global asset management firm providing investment strategy and market research.

Jurrien Timmer

A lead strategist at Fidelity Investments who analyzes macroeconomic trends and equity markets.

Invesco QQQ Trust

An exchange-traded fund that tracks the Nasdaq-100 index and reflects tech sector performance.

The details

Hyperscalers, or massive data center operators, are financing a capital-intensive buildout of artificial intelligence infrastructure while facing a persistent global shortage of memory chips. This supply constraint pushes up procurement costs, compounding the fiscal pressure of a $3.3 trillion total corporate demand for capital. The shift in expenditure indices suggests that while physical investment continues, the rapid growth phase seen earlier in the year has tempered.

September 22, 2026: The Invesco QQQ Trust and SOXX index fund closed higher.
Next 12 months: U.S. hyperscalers are projected to spend $916 billion in capital expenditures.
2027: Combined tech firm capital expenditure is expected to exceed $1 trillion.
2028: The global memory chip shortage is expected to persist until this time.

The Tech Race

This development follows the established trajectory of tech sector capital expenditure cycles, marking a potential shift from aggressive deployment to a phase of performance assessment. It contrasts with the hyper-growth narrative of previous years by highlighting the inflationary pressure of memory shortages and cooling token expenditure.

Investors may see continued volatility in semiconductor and tech-focused funds as firms weigh their $1 trillion annual expenditure goals against stalled growth. Users of AI-integrated workflows should expect costs to remain elevated until the global memory chip shortage begins to ease in 2028.

The takeaway

The massive capital commitment by hyperscalers indicates that infrastructure deployment remains the priority, even as market indices signal a shift in momentum. Watch for the 2027 aggregate capital expenditure reports to see if the sector maintains its projected $1 trillion investment pace.