Goldman: AI Capex Hits $800B in 2026

What is Goldman Sachs forecasting for AI capital spending in 2026?

Goldman Sachs forecasts that AI capital expenditure will reach $800 billion in 2026. The bank's position, reported by Business Insider, is that the AI boom is bigger than investors currently think. Goldman made this call even as a tech selloff weighed on AI-linked stocks on Wall Street.

The bank's headline argument is that markets are underestimating the scale of AI infrastructure build-out. Whether profits can justify those soaring infrastructure costs remains a key concern for investors, Business Insider notes.

What is the $1.8 trillion off-balance-sheet risk?

ZeroHedge published a piece on June 15, 2026, titled "The $1.8 Trillion Off-Balance Sheet Time Bomb At The Heart Of The AI Supercycle." The article frames a $1.8 trillion exposure as sitting outside conventional balance sheets — meaning it may not be fully visible in standard financial reporting.

Off-balance-sheet financing is a method where companies hold assets or liabilities in ways that do not appear directly on their main financial statements. The ZeroHedge piece positions this structure as a systemic risk embedded in the current AI investment wave.

You might also like

OpenAI Simulates Deployments Before Release

Anthropic Ban Opens Door for Chinese Open-Source AI

Z.ai GLM-5.2 Beats GPT-5.5 on SWE-bench Pro

Altman, Amodei & Hassabis Head to G7 AI Summit

Here's what we know so far: the sources confirm the $1.8 trillion figure and the off-balance-sheet framing, but the detailed breakdown of which entities hold that exposure is behind a premium paywall.

Why are AI stocks under pressure right now?

A tech selloff on Wall Street has put AI-linked equities under strain. Business Insider reported on June 11, 2026, that the central investor question is whether AI profits can justify soaring infrastructure costs. That tension is driving the market pressure.

Goldman's counter-argument is that investors are wrong to be skeptical. The bank says the AI boom is larger than the market currently prices in — a direct pushback against the selloff narrative.

What are the key numbers in this story?

How does this connect to broader AI infrastructure spending?

The $800 billion Goldman forecast sits alongside massive physical build-outs happening across the industry. Nvidia's Texas factory is one example of the infrastructure race the Goldman number reflects. AI coding platforms are also straining existing infrastructure, as seen in GitHub's move to AWS.

On the revenue side, Perplexity's tripled revenue shows that some AI companies are generating returns. But the gap between capex commitments and proven revenue at scale is exactly what the Goldman vs. selloff debate is about.

The Adobe creator report found 87% of AI creators reporting business growth — another data point on the demand side of the equation.

What is driving the off-balance-sheet concern specifically?

The ZeroHedge framing centers on the structure of AI financing, not just its size. When large capital commitments are held off-balance-sheet, they can be harder for investors and regulators to track. The piece describes this as a "time bomb" at the heart of the AI supercycle — language that signals concern about visibility and systemic exposure, not just the dollar amount.

The Benzinga report on the Goldman $800 billion forecast, published in May 2026, also covered AI capex projections and Goldman's bullish stance heading into the year.

Key facts at a glance

• Goldman Sachs forecasts $800 billion in AI capital spending for 2026
• Goldman's position: investors are underestimating the AI boom
• ZeroHedge flags a $1.8 trillion off-balance-sheet exposure in AI financing
• A tech selloff has raised investor questions about AI return on investment
• Business Insider reported Goldman's view on June 11, 2026
• ZeroHedge published its off-balance-sheet analysis on June 15, 2026

The most confirmed fact from these reports: Goldman Sachs has put an $800 billion number on AI capex for 2026, while a separate $1.8 trillion off-balance-sheet figure has been flagged as a structural risk sitting beneath the AI supercycle.