Apollo Chief Economist: Slowing AI Returns Could Trigger Market Repricing
Miles Bennett
Apollo chief economist Torsten Sløk warns that AI investment returns remain confined to tech, with almost no profit improvement visible elsewhere — stocks priced for instant earnings growth face a painful repricing.
What exactly is the risk he's flagging?
Sløk's core call: for the foreseeable future, returns on massive AI spending accrue almost entirely to the tech sector. Other industries see little benefit in the near term.
This means → the market has priced in "AI lifts everyone's margins" as a done deal, but cash-flow reality is far behind the valuations.
His own words are blunt: "If the hockey-stick productivity surge takes five years rather than five months, equities priced for instant earnings growth face a painful repricing."
Where does the AI bull case break down?
The bull thesis runs: companies adopt AI → productivity rises → margins expand. Outside of tech, this chain has barely materialised.
Sløk lists a long roster of slow-to-change sectors: healthcare, banking & insurance, energy, defence, pharma, manufacturing, transport, construction, education, law, and the public sector.
In plain terms = the money so far has gone to the people selling the shovels (tech companies). The people digging for gold (traditional industries) haven't struck anything yet.
What does the "token optimisation" signal tell us?
Sløk flags a shift in corporate focus toward "token optimisation" — controlling the cost of calling AI models — as an early warning sign.
This means → companies are finding AI far more expensive to run than expected. Internal conversations have pivoted from "how to make money with AI" to "how to keep AI costs down."
His conclusion is direct: "If companies don't see fast returns on investment, they will slow AI spending."
Is he turning fully bearish on AI?
No. Sløk has previously been optimistic about AI's economic impact, citing the Jevons paradox — the idea that when technology raises efficiency, total demand expands rather than shrinks — to argue AI will fuel a jobs boom.
His target this time is narrow: the market has fully priced in rapid, AI-driven margin expansion, and reality is not delivering at that speed.
This reflects a critical juncture: the Nasdaq 100 is on track for its strongest quarterly performance in years. Whether upcoming earnings can validate AI-linked valuations is the key test for this gap between price and reality.
Content is for reference only, not financial advice.