Goldman expects debt to fund 35% of hyperscaler capex in 2027, about $440 billion
Goldman Sachs credit strategist Amanda Lynam says 2026 borrowing by hyperscalers already topped the bank's $250 billion forecast after a recent surge in debt issuance.
What happened
Hyperscalers are the giant tech firms that run huge data centers for AI. They are borrowing more money to pay for those buildings and chips. Goldman Sachs says that borrowing is running higher than it predicted for 2026.
Why it matters to you
The AI buildout is increasingly paid for with debt, not only with the cash these companies earn. That ties AI growth to credit markets.
The takeaway
Big tech is leaning more on borrowed money to build AI infrastructure, and Goldman expects that to continue in 2027.
What it means for you
Companies that sell to or compete with hyperscalers should note that AI capex is increasingly financed by debt, per Goldman.
Goldman says 2026 hyperscaler borrowing exceeded its $250 billion forecast and projects debt covering 35% of 2027 capex, about $440 billion.