Goldman Sachs Trading Desk: Market Flashing FOMO Signals as Inference Economy Accelerates
nashnova research
Goldman's single-delta trading desk head Rich Privorotsky says the market is showing classic FOMO — spot and vol rising together; he frames the AI economy as a split between "rails" and "interfaces," favoring bottleneck assets and infrastructure over business models that profit from human friction.
What signal made Goldman's trading desk call FOMO?
Privorotsky flagged the textbook fear-of-missing-out pattern: spot prices rising while volatility rises in tandem. This means → buyers are chasing, not positioning calmly.
He personally trimmed some exposure during the Nasdaq 100 rally but kept his overall constructive stance. In plain terms = he lightened up, but didn't flip bearish.
He is bullish on the acceleration of the inference economy — the phase where AI models are called at scale to generate real commercial value — and expects every major player to ship Muse-like products.
"Rails" vs. "interfaces" — how does the AI economy split in two?
Privorotsky divides the AI economy into two layers: rails (back-end infrastructure) and interfaces (the front end users touch directly).
His core call: AI agents will destroy value at the interface layer while creating more activity on the rails. In plain terms = a user may never visit a travel site again, but the AI agent still has to call inventory systems, booking channels, and payment infrastructure — the pipes get busier, not quieter.
This reflects a deeper signal: systems of record (the back ends that actually hold data and settle transactions) gain importance, while systems of engagement (front ends that compete for user attention) get marginalized.
Which business models face disruption from AI agents?
Any model that earns excess returns from friction, complexity, or user inertia faces structural challenge. This means → comparison sites, affiliate marketing, lead-gen platforms, and businesses whose moat is essentially "users can't be bothered to switch" are all in the risk zone.
The logic of advertising value is shifting too: from capturing human attention to being selected by AI agents. In plain terms = ads used to need humans to see them; soon they'll need AI to recommend them.
His trading framework follows directly: go long bottleneck assets, foundational rails, proprietary data, and genuinely differentiated names; avoid models that profit from human friction.
What about the labor market and macro outlook?
Privorotsky cites former IBM CEO Lou Gerstner's framing: the likelier outcome is that corporate revenue keeps growing while demand for incremental headcount falls. In plain terms = not mass layoffs, but "less hiring, less firing" — productivity finally delivering.
On the macro side, he flags two signals to watch: oil prices and the front end and belly of the real-rate curve.
If cyclical sectors can rally in sync with tech, he sees that as the strongest confirmation of the current bullish thesis. This reflects his view: a tech-only rally isn't stable enough on its own.
市场有风险，内容仅供研究参考，不构成投资建议。