Texas Retirement System Warned of AI Infrastructure Risks

Pension fund leaders have signaled caution over potential overspending and systematic risks tied to AI expansion.

Updated on Sept. 21, 2026 in Artificial Intelligence

Bold flat-color editorial illustration showing a massive geometric concrete structure and large power transformers, symbolizing institutional risk in infrastructure investment. — The Teacher Retirement System of Texas CIO Jase Auby has warned that aggressive spending on artificial intelligence infrastructure risks creating stranded assets and systematic overcapacity. AI Illustration. Upload story photo >

Live Poll

Is now a good time to be heavily invested in companies building out artificial intelligence infrastructure?

The Teacher Retirement System of Texas CIO Jase Auby has warned that current artificial intelligence infrastructure investment carries risks of overbuilding. This concern centers on whether capital spending can match future demand as AI adoption reaches 53%.

Why it matters

AI capital investment currently accounts for 3.5% of US annual GDP, creating systematic exposure across asset classes. Funds are now weighing the potential for corporate bankruptcies if revenue realization fails to keep pace with rapid infrastructure buildouts.

The Teacher Retirement System of Texas manages $225.3 billion, while the New York City Retirement Systems, which has moved to reject managers to reduce AI-related risk, holds $327 billion in assets.

The players

Jase Auby

Chief Investment Officer of the Teacher Retirement System of Texas who oversees the management of the state's $225.3 billion retirement fund.

Anthropic

An AI research lab and developer of large language models that recently raised $65 billion in a Series H round at a $965 billion valuation.

Nvidia

A semiconductor company that designs graphics processing units and remains a central holding in institutional portfolios due to its role in AI infrastructure.

The details

The Teacher Retirement System of Texas evaluates current AI buildouts by comparing them to historical capital-intensive infrastructure booms, such as railroads, highways, and the telecom fiber-optic network buildouts of the early 2000s. Pension managers monitor the gap between massive capital expenditure—the upfront cost of buying hardware like GPUs and building data centers—and the actual revenue generation required to sustain these investments. Without sufficient demand, this cycle risks creating stranded assets and bankruptcies.

The telecom and fiber infrastructure buildout occurred in the early 2000s.
The Teacher Retirement System of Texas reported its portfolio investments on Aug. 31, 2025.
Jagdeep Singh Bachher characterized AI as a revolution in March 2026.
Anthropic secured $65 billion in a Series H funding round in May 2026.
Jase Auby addressed the TRS investment committee in September 2026.

The Tech Race

Investment committees are reevaluating the sector following the early 2000s telecom and fiber infrastructure buildout. This comparison serves as a benchmark for determining whether current AI spending will lead to sustainable growth or a period of widespread corporate distress.

Residents should anticipate that institutional portfolio adjustments could influence long-term retirement fund performance and state investment stability. Investors should monitor ongoing committee disclosures for changes to asset allocation strategies as pension boards recalibrate their risk profiles.

The takeaway

The trajectory of AI development may mirror past infrastructure cycles, suggesting a period of volatility as markets reconcile high capital expenditure with actual utility. Watch upcoming quarterly pension reports for further changes in Nvidia equity exposure and diversification efforts by state funds.