Goldman Sachs Says Hyperscaler AI Investment Will Peak, US Investment-Grade Bond Issuance Hits Record

Goldman Sachs Chief Economist Jan Hatzius recently said that the growth in investment by hyperscale data centers in the field of artificial intelligence (AI) is expected to peak in the coming weeks.

Specifically, he pointed out that the year-over-year change in hyperscale data center investment as a share of U.S. GDP is expected to peak in the fourth quarter of 2026 at about 1.5%.

The top hyperscale data center operators—Amazon, Microsoft, Alphabet, Meta Platforms, Oracle and SpaceX—have invested record amounts of capital in AI infrastructure, and capital expenditures are expected to reach about $725 billion to $800 billion this year alone, up 64% from 2025.

Alphabet and Microsoft will report earnings on Wednesday, while Amazon will report on Thursday, giving investors a glimpse into future investment plans.

Companies accelerate borrowing

At the same time, the surge in AI investment is pushing companies to accelerate borrowing. According to reports, U.S. investment-grade bond issuance this year is expected to set a historical record on Monday, surpassing the $1.75 trillion record set during the pandemic-induced borrowing boom in 2020. According to data, after Monday's trades are priced, issuance so far this year will reach $1.76 trillion.

Many analysts had already expected annual issuance to set a record this year, driven by expectations of a sharp increase in spending in the AI field. Since 2026, AI-related financing in the global high-grade bond market has exceeded $300 billion, nearly three times the full-year scale in 2025. There have been nine U.S. deals this year with individual issuance of at least $25 billion, and apart from the financing by Skydance Corp. for an acquisition, all the rest came from technology companies.

"We are witnessing one of the most capital-intensive corporate cycles in decades," said Lindsay Rosner, head of multi-sector fixed income investing at Goldman Sachs Asset Management.

Goldman Sachs predicts that under an optimistic scenario, the wealth effect brought by AI investment will drive AI-related stocks up 55%, other stocks up 20%, and the broader market up 30%, with the wealth effect boosting consumption growth by 0.7 percentage points. But under a pessimistic scenario (AI stocks down 40%, other stocks down 15%, and the broader market down 20%), the wealth effect will drag on consumption growth by 0.3 percentage points.

AI investment is transmitted through the path of "stock market gains → increased household wealth → higher consumer spending." In a previous report, the bank wrote that over the past two years, the wealth effect has boosted consumption growth by an average of about 0.3 percentage points per year, of which about 0.2 percentage points can be attributed to AI-related stock returns. Over the next year, it is expected to boost consumption growth by 0.4 percentage points, driven mainly by the strong performance of AI stocks.

Shi Nande, president of Goldman Sachs Group for Asia-Pacific (excluding Japan), previously warned that the wealth effect brought by the rise in AI-related stocks is boosting consumption, but this also means that once the market corrects, it will immediately be transmitted to U.S. consumption through the wealth effect, which is also one of the biggest financial risks currently concerning global financiers.

Hatzius previously pointed out that AI investment's contribution to U.S. GDP growth in 2025 was "basically zero," mainly because a large amount of hardware relies on imports, which are offset by imports in GDP accounting. Goldman Sachs predicts that if the data center investment boom ends, it could reduce the U.S. GDP level by about 2 percentage points.

This article is from the WeChat public account "Cailian Press," author: Huang Junzhi, published by 36Kr with authorization.