Goldman Sachs Says 2026 AI Spending Could Top $1 Trillion, Prediction Markets Push Anthropic's Odds of a $2 Trillion Valuation to 77.5%

Wall Street's AI spending numbers keep climbing, and so does the price tag on betting that Anthropic becomes a $2 trillion company by New Year's Eve.

According to a Motley Fool analysis, consensus estimates already call for roughly $800 billion in AI capital expenditure in 2026, driven by hyperscalers like Amazon, Microsoft and Google building out data centers. Goldman Sachs thinks that number is too low. The bank's analysts told Motley Fool the true 2026 figure could land closer to $1 trillion, because most public estimates "do not include investment in AI by private companies or by companies outside the U.S., including firms in Asia." Goldman's math cuts both ways: it expects reported U.S. spending to come in a bit lighter than headline forecasts suggest, but global and private spending to run more than $200 billion over expectations.

McKinsey & Co. is even more aggressive on the long horizon, projecting AI capex could hit $7 trillion globally by 2030, a figure the firm calls "one of the largest infrastructure build-outs in modern history," per Motley Fool's reporting.

The Skeptics Get Their Say

Not everyone is buying the growth story at face value. Research from AllianceBernstein, cited by Motley Fool, argues "the use cases of AI are slowly emerging, albeit it's frankly too early to really lay out a high confidence path." The firm says clients keep bringing up memories of the dot-com bubble and the 19th-century railway-building frenzy, warning that in past infrastructure booms, adoption and payoff took far longer than investors hoped, and the eventual winners weren't the companies everyone expected.

Hundreds of billions in capital spending built on projections, not signed contracts, is exactly the kind of thing that inflates and then corrects hard when reality catches up to the forecast. AllianceBernstein isn't predicting collapse, but flagged that nobody actually knows yet who profits long-term from this build-out.

Motley Fool's counterargument is that this cycle looks different because the companies doing the spending, think Nvidia, Amazon, Google and Microsoft, are largely profitable already and can self-fund the bet rather than borrowing on hype alone. Nvidia gets singled out as the biggest beneficiary of rising global capex, described by Motley Fool as not just the largest AI company but the most valuable company in the world, period, with its CUDA software platform making its hardware stickier than a typical chipmaker's.

Anthropic's Valuation Becomes a Betting Market

Separately from the capex debate, prediction-market activity tracked by Crypto Briefing and KuCoin, using data from the analytics platform Vera, shows the odds of Anthropic reaching a $2.0 trillion valuation by December 31 climbed from 67% to 77.5% over the past week.

That's a real move in a real market, but it's not an audited valuation, a funding round term sheet, or an SEC filing. It's traders wagering on an outcome, and the price shifted because of sentiment, not new financial statements.

What's actually feeding that sentiment is a string of recent developments in the AI sector. Anthropic reported turning profitable in the second quarter of 2026, according to Crypto Briefing's related coverage, while OpenAI is reportedly eyeing profitability by the third quarter. OpenAI separately secured $122 billion in funding, with CEO Sam Altman calling AI compute the most expensive project humanity has undertaken, per the same coverage. Anthropic has also reportedly agreed to pay AI infrastructure firm Nscale $45 billion for computing capacity, and separate reporting referenced by Crypto Briefing said Anthropic was pursuing a new funding round exceeding $86 billion. None of these figures come with independent verification in the sources reviewed here, and they should be read as reported deal activity, not confirmed closings.

What's Actually Unresolved

One piece of related reporting flagged by Crypto Briefing found that cost, not technical limitations, is now the primary barrier to enterprise AI adoption. That's the tension sitting underneath every one of these numbers. Hundreds of billions are being committed on the bet that businesses will eventually pay enough to justify the infrastructure, but the near-term obstacle these companies are running into is that their own customers think the product is too expensive.

The next real data points will come from hyperscaler earnings later this year and whatever happens with Anthropic's reported funding round. Until money actually changes hands and gets disclosed in a filing, the $1 trillion capex number and the 77.5% valuation odds are both forecasts, priced by people betting on the future, not statements of what has already occurred.