Apollo says there is no AI payoff outside tech. The number to check is your own
Apollo's chief economist Torsten Slok published a note on 24 October 2026 titled "No Signs of AI Payoff Outside the Tech Sector". The title is not a joke and it is not a hot take. It is the boring version of a number everyone in AI keeps avoiding. His argument is narrow. Revenue per worker and marg
Posted on Oct 26 • Originally published at voxahq.in
Apollo's chief economist Torsten Slok published a note on 24 October 2026 titled "No Signs of AI Payoff Outside the Tech Sector". The title is not a joke and it is not a hot take. It is the boring version of a number everyone in AI keeps avoiding.
His argument is narrow. Revenue per worker and margins are rising for the largest technology companies, and they are flat for the rest of the S&P 500. The money is showing up in the firms selling the infrastructure, not in the firms buying it.
Every general purpose technology looks like this at first. Electricity took decades to show up in factory productivity, because the gain required rebuilding the factory around the motor, not installing a motor. Computers took longer than the spreadsheet.
Slok's read is that AI is in the same phase: the tool is deployed, the work is not reorganised yet, so the number does not move.
If the payoff is not showing up, it usually means the tool was added on top of an unchanged process. A team buys a chatbot, keeps the same approval chain, and gets a faster draft attached to the same bottleneck.
So the useful question is not adoption. It is whether one number moved:
If no number moved, you bought a subscription, not an improvement.
The obvious reply is that measurement lags. Firms are in year one of deployment, and the gain appears when the process is rebuilt around the tool, which takes two to three years. That is also what the historical record says, which makes this a warning about expectations rather than a verdict on the technology.
The second reply is selection. The companies seeing gains have the data, the engineering staff and the willingness to change how work is done. Everyone else bought the same model and got the same demo.
Pick one process this quarter. Measure it before. Change the process rather than adding another subscription. Measure it again. If the number does not move, stop paying for it.
That is cheaper than waiting for the macro data to agree with the pitch deck.
Originally published at voxahq.in. VOXA is an independent AI tools and automation site. No affiliate links, no sponsored verdicts.