22V Research: AI Adds 180 BPS to Margins at 25 S&P Firms

The number worth staring at from this week's earnings tally is 25. That is how many companies in the S&P 500 have now put a real figure on what AI is doing to their margins, according to research firm 22V, reported by Bloomberg. The average lift those firms cite is 180 basis points. Strip out the ones bundling AI with other efficiency programs and the typical bump is still 150.

That is small in one sense (25 of roughly 500 names) and striking in another. A quarter ago, per the same tally reproduced by Briefs.co, only 17 constituents had bothered to quantify anything and the average boost was 20 basis points. In one quarter, both the number of companies willing to attach a specific figure and the size of that figure have jumped an order of magnitude. 22V president Dennis DeBusschere frames it plainly in the coverage: direction matters more than precision in these early estimates, and the direction is toward more AI users reporting better margin improvement.

The examples are what make it land, because these are not names anyone bought for the AI trade. Waste Management says its SmartTruck routing system is running at more than $300 million of annual EBITDA. Logistics broker CH Robinson credits AI for a 60% productivity improvement since 2022. Fortinet reported operating margin up 490 basis points in Q2. Willis Towers Watson expects $400 million in cost cuts, thanks mostly to process automation. 22V's punchline is that if that margin picture generalizes across the index, the S&P 500 is worth at least 10% more than it trades for today.

The honest caveat is that 25 out of roughly 500 is a very small pilot to draw an index-level conclusion from, the number falls to 150 basis points once mixed-attribution firms are excluded, and none of the reporting nets those gains against what any of these companies are spending on AI in the first place. What the coverage does not give you is a revenue-side view, or a durability test on whether the lift survives past the first round of one-time headcount cuts.

What is worth watching from here is less the specific 180 figure and more the disclosure trend. If the count of S&P firms attaching a hard basis-point number to AI keeps compounding, the next few earnings seasons will do more to reprice non-tech names than any hyperscaler capex line item.