Goldman Sachs Warns AI Stock Investors of New Risk
Yahoo Finance ·
Goldman Sachs Warns AI Stock Investors of New Risk Moz Farooque ACCA Fri, September 4, 2026 at 8:13 AM EDT 2 min read GS MSFT This article first appeared on GuruFocus . Goldman Sachs is flagging a new risk hiding inside the artificial intelligence boom: AI is getting cheaper so quickly that the industry could eventually build more computing capacity than customers can profitably consume . The bank's Delta One desk warned that collapsing prices for AI tokens could threaten technology-stock valuations unless usage grows fast enough to absorb the enormous amount of infrastructure coming online. Warning! GuruFocus has detected 3 Warning Sign with MSFT. Is MSFT fairly valued? Test your thesis with our free DCF calculator. Tokens are the basic units AI models process when generating text, code and other outputs. Lower token prices make AI cheaper for customers and can accelerate adoption, but they also reduce the revenue generated from each unit of AI activity. That tradeoff is becoming increasingly important. Goldman highlighted an industry benchmark from Silicon Data that recently fell to $0.97 per million tokens , including a 29% decline during August alone . Competition between frontier-model developers, cheaper open models and more inference moving toward lower-cost or local hardware are all pushing prices down. Add more inference moving toward cheaper/local hardware and I struggle with per token pricing as a durable end state model, Goldman said. The danger is straightforward: if AI prices fall faster than overall usage grows, demand for expensive data-center capacity may eventually disappoint. Goldman warned that periods of compute oversupply become plausible unless future models produce major improvements that stimulate substantially more consumption. That matters because the industry is still spending extraordinary amounts on infrastructure. Microsoft said capital expenditures reached $41 billion in its latest quarter and expects roughly $175 billion of calendar-2026 investment , while Alphabet expects $175 billion to $185 billion of 2026 capital expenditures as it expands AI and cloud capacity. Goldman's warning does not necessarily mean AI demand is weakening. In fact, lower prices could dramatically expand usage. The critical question is whether token consumption grows faster than token prices decline . That makes AI utilization, cloud revenue growth and data-center returns increasingly important metrics for investors in Microsoft ( NASDAQ:MSFT ), Alphabet ( NASDAQ:GOOGL ), Meta ( NASDAQ:META ) and the broader AI infrastructure trade. If cheaper AI triggers an explosion in applications and usage, today's massive compute buildout could still prove justified. But if prices continue collapsing while consumption grows more slowly, investors may eventually discover that the AI boom created plenty of demand but even more supply .
DYAX Investor Sentiment
Bullish (Long) 45% · Bearish (Short) 55%
252 participants
Related News
- Analyst Says Nvidia's Huge AI Deal Goes Beyond Chips
- Morgan Stanley Hands Elon Musk a New X Headache
- Anthropic Makes Stunning $15 Billion Move Before IPO
- Meet the Low-Cost Vanguard ETF With 26.2% Invested in Nvidia and Alphabet, While VOO Has Just 13.4%.
- Grid Dynamics and Turning Point Brands have been highlighted as Zacks Bull and Bear of the Day
- Cathie Wood Cuts AMD and Puts $53 Million Into Nvidia
