Goldman strategist sees cleaner AI positioning after Democratic sweep

By Grace Kim

Data-center capital spending could move into 2027 after the midterm election, according to Rich Privorotsky, Goldman Sachs’ head of European equities trading, whose client note argues that a Democratic outcome may accelerate hyperscaler investment rather than deter it.

What changed after the election?

Privorotsky described investor sentiment toward AI stocks as unambiguously negative in recent weeks. He said the midterm result appeared to realize the sector’s worst fears for some investors.

He said he had watched AI shares be “sold to death,” while AI plays had derated and short sellers now had to cover their positions. His advice to fund managers was: “the micro remains strong, electoral uncertainty has diminished and positioning looks considerably cleaner.”

Before Tuesday’s vote, investor attention centered on whether a Democrat victory would curb data-center proliferation. The Texas governor race drew particular focus because about one-quarter of all planned data centers were in the state.

In the note dispatched Wednesday, Privorotsky said Republican candidate and incumbent Greg Abbott’s defeat by Gina Hinojosa pointed to a potentially more restrictive regulatory environment. He also said the election had reduced uncertainty without removing tighter-rule risk.

Takeaway: The political overhang has eased, but the threat of tighter data-center regulation remains.

Can capital spending arrive in 2027?

Privorotsky said corporate decision-makers might pull capital deployment forward to 2027 to use current permitting and strategic transactions before the planning regime adjusts. He said a new president in 2029 might oversee tax changes, increased administrative oversight and greater scrutiny of mergers. In that setting, hyperscalers may decide to simply get a move on.

The pull-forward gives operators a narrower window to act before the next presidential cycle changes the policy backdrop.

Takeaway: A tougher 2029 setup could pull hyperscaler commitments into 2027.

Who benefits if compute expansion slows?

Privorotsky argued that constraints on future compute expansion would raise the value of existing, energized compute. That favors operators with capacity already deployed.

He also acknowledged that delays or abandonment of some data centers could slow the investor-defined “picks-and-shovels” trade. The note separated three parts of the AI infrastructure chain:

• Amazon, Microsoft and Alphabet: Their deployed capacity could become more valuable if future expansion is constrained.
• Semiconductor-chip, cooling and electrical-equipment manufacturers, plus energy providers: Delayed or abandoned data centers could slow the picks-and-shovels trade.
• Nebius and CoreWeave: Privorotsky called the investment case more nuanced and said it depends on whether compute is contracted or existing.

Takeaway: Installed compute may gain value while suppliers face delays, leaving neoclouds dependent on contract status.

What do AMD’s numbers and falling valuations show?

The correction was already substantial: the Philadelphia Semiconductor Index exchange-traded fund had fallen by one-fifth, or 20%, in the past month, according to Privorotsky’s note.

The note cited AMD’s strong third-quarter results as evidence that the AI capital-expenditure cycle remained strong at present. Privorotsky described AMD’s fourth-quarter guidance as extremely strong and said it implied another 27% sequential acceleration.

He said throttling supply may not destroy demand; it may push pricing higher for longer. He also said hyperscalers might have over-ordered relative to their needs if data centers slipped, but he argued lower valuations already reflect that possibility.

Takeaway: Demand still appears strong enough to support pricing, while lower valuations absorb project-delay and over-ordering risks.

Can lower volatility lift major indexes?

Privorotsky doubted that current levels of the VIX, the so-called fear index, could persist now that the election was over and uncertainty had receded. If volatility falls as quickly as he expects, its inverse relationship with stocks could push major indexes higher, he said.

He noted that AI and AI-adjacent equities make up almost half of those benchmarks, leaving the group with substantial influence over the broader index move.

Takeaway: A fast VIX decline could lift AI-heavy benchmarks even as tighter rules delay some new projects.

For investors, the near-term test is whether hyperscalers pull spending into 2027 and whether the VIX contracts quickly enough to revive an AI-heavy market.