NewsMacroGoldman Sachs Lifts 2027 AI Project Finance Forecast to $545 Billion

Goldman Sachs Lifts 2027 AI Project Finance Forecast to $545 Billion

Author: CryptoBriefing·

Key Takeaways

  • •Goldman Sachs raised its 2027 AI project finance forecast to $545 billion, an increase of roughly 80 percent from its earlier $300 billion estimate.
  • •The forecast divides into about $225 billion for data center shells and $320 billion for chips, with chip financing projected to grow almost sevenfold from $48 billion this year.
  • •Goldman lifted its 2027 hyperscaler capital expenditure estimate to between $1.2 trillion and $1.4 trillion, where the top of the range represents nearly 4 percent of US GDP.
  • •Chip financing deals typically involve special-purpose vehicles owning the hardware and customer contracts of five years or less, and a large share may come without traditional corporate guarantees.
  • •Reported transactions include a $40 billion GPU financing package involving SpaceX backed by Goldman Sachs and Apollo, plus a $22 billion chip financing linked to Crux AI, a venture connected to Blackstone and Alphabet.
Goldman Sachs Lifts 2027 AI Project Finance Forecast to $545 Billion

Goldman Sachs' credit team has raised its forecast for artificial intelligence project finance in 2027 to $545 billion, up sharply from its previous estimate of $300 billion—an increase of roughly 80 percent—according to new research from the bank.

The projection is split into two components. About $225 billion would fund data center shells—the buildings, power hookups, and physical structures that house the hardware—while the remaining $320 billion would finance the chips inside those facilities. Put simply, Wall Street now expects to lend more against what sits in the server racks than against the walls around them.

Compute is where the growth is most pronounced. Chip financing has reached $48 billion so far this year, according to Goldman, and the bank expects that figure to grow almost sevenfold to $320 billion by 2027.

The revision also reflects a broader shift in Goldman's thinking. The bank has lifted its estimate for hyperscaler capital expenditure in 2027 to between $1.2 trillion and $1.4 trillion. At the top of that range, annualized spending on AI hardware and intellectual property would amount to nearly 4% of US GDP.

How a GPU financing deal works

Under the structure Goldman describes, a special-purpose vehicle, or SPV, owns the chips. An SPV is a standalone legal entity created for a single job, which keeps the assets and their debt separate from the parent company's balance sheet. This is the classic project finance playbook—a structure banks have long used to fund big-ticket infrastructure such as power plants, pipelines, and toll roads—now being applied to computing hardware.

A customer then signs a lease or service contract for the computing capacity, with those contracts running five years or less. Lenders are repaid from the cash the contract produces. If the customer keeps paying, the debt gets serviced. If the customer stops paying, lenders are mostly left holding the hardware and the paperwork.

That structure leads to the key caveat in Goldman's research: a large share of the $320 billion in chip financing may come without traditional corporate guarantees. In a typical corporate loan, a parent company with deep pockets promises to cover the debt if things go sideways. Without that backstop, lenders are betting on two things—the contract itself and the creditworthiness of the customer who signed it.

The deals already on the table

A $40 billion GPU financing package involving SpaceX was recently reported, with Goldman Sachs and Apollo backing the arrangement. Separately, a $22 billion chip financing has been tied to Crux AI, a venture connected to Blackstone and Alphabet.

These transactions sit alongside a separate wave of borrowing. Earlier reports pointed to hundreds of billions of dollars in direct investment-grade bond issuance from the major hyperscalers in 2026—a funding route that leans on the issuers' own balance sheets and credit ratings, in contrast to the contract-based chip deals above.

What it means for lenders and borrowers

Goldman's research flags several pressure points that could test lenders' assumptions. They include power supply constraints, regulatory opposition to data center projects, and whether end users can earn enough revenue to meet their contract obligations.

Growth from $48 billion to $320 billion means a market that barely existed at scale is expected to become one of the largest corners of project finance within roughly two years. Banks like Goldman Sachs and alternative asset managers like Apollo and Blackstone are clearly positioned to arrange and fund these deals, which gives them a direct stake in the market's growth.

The figures worth tracking from here are concrete: how much of the $320 billion chip forecast actually closes, how many of those deals carry a corporate guarantee, and whether power and permitting constraints start delaying the facilities those chips are meant to fill.

This article is based on reporting by CryptoBriefing, edited by Diego Almada Lopez.