Goldman Sachs executives are drawing critical distinctions between today’s AI market frenzy and the dot-com bubble of 25 years ago, while simultaneously warning that artificial intelligence is contributing almost nothing to current US GDP growth. Ben Snider, who is set to become Goldman Sachs’ US equity chief, explained in a Bloomberg Odd Lots podcast episode that modern investors have learned from past mistakes and are now focusing on tangible, near-term earnings rather than speculative long-term AI potential.
During the late 1990s dot-com bubble, investors attempted to estimate the long-term productivity gains and economic benefits of the internet, causing valuations to expand dramatically and unsustainably. Today’s market approach is fundamentally different, with investors concentrating on immediate returns from AI infrastructure companies including semiconductors, hyperscalers, and power companies that are directly benefiting from AI investment.
Snider pointed to Goldman Sachs’ newly developed Speculative Trading Indicator, which measures trading activity in unprofitable stocks, penny stocks, and high-valuation companies. This gauge reveals that speculative activity remains “well below levels” seen 25 years ago during the dot-com era and even below the 2021 market euphoria. “I think this has been one of the least enthusiastic markets that is often described as a bubble in recent history,” Snider noted.
However, Jan Hatzius, Goldman’s chief economist and head of research, delivered a sobering reality check for AI enthusiasts. Despite the boom in mega tech stocks, Hatzius stated that “close to 0%” of US GDP growth can be attributed to AI in 2025. His analysis reveals that goods invested in the AI sector are largely imported, and semiconductors are treated as intermediate inputs rather than investments in GDP calculations.
Over the last three to four years, AI investment has contributed only about 20 basis points to GDP growth, and that contribution has dropped to nearly zero over the past year. This disconnect between market enthusiasm for AI stocks and actual economic impact raises important questions about the sustainability of current AI valuations and whether the technology’s transformative potential is being realized as quickly as investors hope.
Key Quotes
Today, investors are saying we saw what happened that time. It’s too hard. And so what we’re really going to focus on is the earnings today
Ben Snider, Goldman Sachs executive becoming US equity chief, explained how modern investors have learned from the dot-com bubble and are now prioritizing immediate, tangible earnings from AI companies rather than speculative long-term projections.
I think this has been one of the least enthusiastic markets that is often described as a bubble in recent history
Snider characterized the current AI market boom as surprisingly restrained compared to historical bubbles, based on Goldman’s Speculative Trading Indicator showing activity well below 2021 and dot-com era levels.
When we look at the impact of AI investment on measured GDP growth, on the numbers that are actually being printed, we’re getting only about 20 basis points of contribution over the last three or four years, and pretty close to zero over the last year
Jan Hatzius, Goldman’s chief economist, delivered a stark assessment of AI’s actual economic impact, revealing that despite massive investment and market enthusiasm, AI has contributed almost nothing to measurable US GDP growth in 2025.
Our Take
Goldman Sachs is essentially telling two stories simultaneously: investors are being more rational about AI than they were about the internet, but AI isn’t delivering economic results yet. This creates a fascinating paradox—if the market is truly focused on near-term earnings as Snider claims, why are AI stocks booming when GDP contribution is zero? The answer likely lies in the concentration of benefits among infrastructure providers like NVIDIA and cloud hyperscalers who are capturing immediate revenue from AI buildout, even if broader economic productivity gains remain elusive. This suggests we’re in an investment phase rather than a productivity phase of AI adoption. The real test will come when companies demand returns on their AI spending. If productivity gains don’t materialize within the next few years, even this more disciplined market could face a reckoning, albeit perhaps a more orderly one than the dot-com crash.
Why This Matters
This analysis from Goldman Sachs executives matters because it challenges the narrative that AI is already transforming the economy while simultaneously suggesting the current market rally may be more rational than previous tech bubbles. The distinction between today’s earnings-focused approach and the dot-com era’s speculative frenzy indicates investors have matured, potentially reducing the risk of a catastrophic market collapse. However, Hatzius’s warning about AI’s negligible GDP contribution reveals a critical gap between hype and reality. If AI investments continue without measurable economic returns, even a more disciplined market could face corrections. This matters for businesses making massive AI infrastructure investments, policymakers evaluating AI’s economic impact, and workers wondering when AI productivity gains will materialize. The focus on semiconductors and hyperscalers as immediate beneficiaries also signals where real value is being created today versus speculative future gains. Understanding this disconnect is crucial for anyone navigating AI investment decisions or assessing the technology’s true economic trajectory.
Related Stories
- Goldman Sachs Hires Google’s Melissa Goldman as Tech Head for AI Push
- Meta and Nvidia Billionaires’ Wealth Soars $152B in AI Boom
- Google Founders’ $511B Fortune Soars on AI Breakthrough Success
- CEOs Express Insecurity About AI Strategy and Implementation
- Global Funds Turn to Indian Stocks as Hedge Against AI Risks