Short answer: Jim Covello, identified in a September 2024 Futurism report as a senior Goldman Sachs stock researcher, warned that artificial-intelligence investment could fail if the technology stayed too expensive or delivered too little practical productivity. That warning was an argument about economics, not a prediction that a crash was imminent. Goldman Sachs discussions in 2025 and 2026 show the question remains unresolved: spending is enormous, some suppliers are already earning substantial profits, and investors still have to prove that those profits and customer benefits will last.
What Covello actually warned about
The headline dates from September 25, 2024. It framed Covello’s concern as an AI bubble that could “explode,” but neither the report nor the later evidence establishes a timetable for a market collapse.
As an Amazon Associate I earn from qualifying purchases.
As reported by Futurism, Covello argued that AI’s price and infrastructure build-out had moved ahead of demonstrated usefulness. The report attributed these statements to him:
Crashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteWindows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstall“Despite its expensive price tag, the technology is nowhere near where it needs to be in order to be useful.”
#1 Best Overall
“Overbuilding things the world doesn’t have use for, or is not ready for, typically ends badly.”
Those quotations should be treated as the 2024 report’s rendering of Covello’s research, rather than as an independently checked transcription of the underlying Goldman Sachs document. His central test was straightforward: will the revenue, cost savings and productivity created by AI justify the capital being spent on it?
Why high spending alone does not prove a bubble
A bubble claim requires more than a large investment number. The relevant questions are whether customers receive durable value, whether that value becomes recurring revenue and earnings, and whether stock prices assume profits that may not persist.
Recommended Free Tools
- Spending versus earnings: Data centers, chips, networking and electricity can generate sales for suppliers before end users have earned a return. The key issue is whether the chain ultimately produces sustainable profits rather than simply passing capital from one company to another.
- Projected productivity versus delivered value: AI may improve output, but investors need evidence that those gains exceed model, software, labor, energy and integration costs.
- Supplier profits versus future demand: Infrastructure companies can benefit during a construction boom. Their earnings become vulnerable if customers slow capital expenditure before AI applications produce enough cash to finance the next spending cycle.
- Valuation versus durability: A high share price can be rational if future earnings are durable. It becomes fragile when forecasts assume rapid growth without evidence that margins and demand will hold.
What Goldman Sachs said after the 2024 warning
The debate did not end with Covello’s 2024 comments.
Renewed concerns in 2025
Goldman Sachs revisited AI-bubble concerns in an October 2025 discussion, indicating that questions about spending, valuations and eventual returns were still active rather than settled.
Covello’s view in June 2026
In a June 2026 Goldman Sachs interview, Covello said AI economics appeared more questionable than they had two years earlier and discussed whether the investment boom had yet demonstrated returns. That is a continuing evaluation of the economics, not a dated forecast that a crash would occur.
The 2026 valuation perspective
A separate Goldman Sachs valuation analysis in 2026 presented both sides. It noted that investment in AI is already generating profits that support some stock prices, while warning that the market may overestimate how durable those investment-related earnings will be. The distinction matters: real current profits can coexist with risk that future growth slows or margins normalize.
Free tools Windows power users keep installed
One-click scans. No signup required.
How large is the investment boom?
Goldman Sachs Research forecast global AI investment to exceed $1 trillion in 2026. This is a forecast, not a finalized tally of money already spent. The estimates depend on modeling assumptions, and Goldman Sachs cautioned that some companies’ capital expenditure may be counted more than once as spending flows through the technology supply chain.
| Measure | 2026 | 2027 | 2028 |
|---|---|---|---|
| U.S. AI investment as a share of GDP | 1.8% | 2.5% | 2.8% |
| Global AI investment as a share of global GDP | 0.9% | 1.3% | 1.4% |
These are Goldman Sachs modeled estimates, not measured outcomes. They should not be read as proof that the forecast will be reached or that every dollar represents new, distinct investment.
The bearish case: why the boom could disappoint
Capital may outrun customer economics
Companies can keep ordering computing capacity while the applications using it remain experimental or price-sensitive. If customers cannot capture enough savings or new revenue, infrastructure demand may slow before investors recover the capital committed.
Returns may arrive later than valuations assume
AI projects often require data preparation, workflow redesign, security controls and specialized staff. Delayed adoption does not make the technology worthless, but it can undermine share prices built on near-term growth.
Profits may be cyclical
Chip and data-center suppliers can report exceptional earnings during expansion. Those earnings could weaken if cloud companies and other buyers reduce capital expenditure, even if long-term AI use continues to grow.
Best Value
The countercase: why “bubble” is not established
Some investment is already profitable
Goldman Sachs’s 2026 analysis says profits generated by the investment boom are supporting some stock prices. That is evidence against the simplest version of the bubble argument—that the spending has no economic output at all.
Productivity benefits may compound
Early returns can be difficult to measure when companies are integrating AI into existing products and processes. If reliability improves and costs fall, current experiments could become recurring, high-margin services.
Valuations differ across companies
“AI” is not one asset or one business model. A profitable infrastructure supplier, a software company with paying customers and a firm spending heavily without revenue face different risks. A broad bubble label can hide those differences.
Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Repair Windows errors before they cause bigger problems3Fix the driver behind crashes, sound loss and screen glitchesWhat would show that the investment is paying off?
- Customer-level evidence: Businesses should report measurable savings, higher output or new revenue attributable to AI, not just user counts or pilot projects.
- Recurring revenue: Providers need contracts and usage that renew at prices covering computing, energy and support costs.
- Cash returns on infrastructure: Data-center and cloud investment should produce cash flow over a reasonable period rather than relying on continually rising demand.
- Durable margins: Earnings should remain healthy as competitors add capacity and model prices decline.
- More disciplined valuations: Share prices should increasingly reflect demonstrated earnings instead of distant scenarios alone.
So, is AI in a bubble?
The evidence supports a more precise answer than a yes-or-no headline. AI investment is large enough, and expectations high enough, to create bubble-like risks. Covello’s warning is therefore relevant as a framework for testing returns. But high spending, current supplier profits and continuing adoption do not establish that the entire market is a bubble, nor do they show that a crash is imminent.
The decisive issue is durability. If customers turn AI capabilities into dependable productivity and revenue, today’s spending can support long-lived earnings. If spending keeps rising while customer returns remain uncertain, valuations and supplier profits become increasingly exposed to a slowdown.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
