Amid Market Fluctuations, Here’s the AI Risk We’re Not Talking About

It’s clear that the stock market is reacting to AI, with a sharp selloff hitting companies powering the AI boom last week, including chip companies that collectively lost over $1 trillion. But is there something about different companies’ approach to AI that’s triggering the different ways the market is reacting to those companies?

As July came to a close, Microsoft was up 9 percent and Meta was down almost as much, with business headlines reporting that AI trade was splitting Big Tech and that markets are riding a roller coaster of Fed confusion and AI hope.

Of course, there are many variables at play, but there’s one major trend that our five-year analysis of 275 global publicly trade companies picked out – and that looks like it could be playing a role in the Microsoft/Meta split as well.

A CNBC article Thursday contained what seemed to be a throwaway line that actually ties closely to what our analysis found. The article – about the significant directional differences between Microsoft, whose Q4 results beat analyst expectations, and Meta, which issued a revenue forecast that was weaker than expected, cited Ben Barringer, head of technology research at Quilter Cheviot, as saying: “Right now, the narrative from Mark Zuckerberg is a little light on detail.”

That may sound minor, but our analysis has found that actually, those details are correlated with market success over time – not only for enterprises providing AI tools but also for those using them.

The AI Hype Tax: Why Vague AI Talk Is Costing Companies

The stock prices of the largest global listed enterprises show a strong correlation between going big on AI and enjoying a superior stock price evolution – as long as those enterprises disclose concrete actions and precise targets.

That’s what we found in the 2026 AI Barometer, our analysis of the share price and annual reports of 275 companies traded on the S&P 100 (U.S.), FTSE 100 (UK), DAX (Germany), and CAC 40 (France) between 2020 and 2025.

The biggest lesson for companies looking to use AI is that general declarations about how AI will transform the business aren’t cutting it anymore with investors, based on the share price analysis. Our report found a large stock outperformance gap tied to transparent AI communication, with companies that actively discussed AI in their financial reports outperforming companies that weren’t talking about it by 18 percentage points on average.

But talking the talk isn’t enough. We found that companies that either described their concrete actions to implement AI or committed to precise business targets tied to increased productivity or revenue were outperforming companies that mentioned AI in a vague way, without tying it to their actions or goals, by 14 percentage points the following year.

The bottom line: Be specific on how the company will be deploying AI and which improvements in efficiency or revenue will result. For instance, instead of making vague jargon-filled statements about AI plans, tell investors how much return you expect from concrete tasks like using AI to improve inventory allocation, optimize cash collection reconciliation or increase forecasting accuracy.

Tying Your Company’s Fate to a Single AI Player Is One of the Biggest AI Risks

Anthropic has unfrozen access to its Fable and Mythos AI models, but the uncertainty introduced by the U.S. government decision to ban them has made many companies question their dependence on advanced AI models that carry risks due to government control, cyber threats and high usage costs. The companies we analyzed communicated frequently about the risks related to AI’s impact on the business landscape, especially the fragmented AI regulation and compliance burden, the risk of AI-enabled cyber threats and attacks, and competitive pressure resulting from AI adoption.

But one of the biggest AI risks for a company may be relying too heavily on any single AI provider.

Our findings showed that from 2020-2025, Microsoft had the most AI/cloud platform mentions in financial reports and OpenAI had the most model mentions, but there’s been a sharp decline in ChatGPT mentions, a strong uptick of Google, and the emergence of Chinese players such as DeepSeek and Qwen. Amazon and Anthropic are strong contenders in the mix and could well move up the rankings.

Of course, the deck is already starting to be reshuffled. In the last few weeks alone, Alphabet raised guidance for capital expenditures and reported a quarter of negative free cash flow, meaning that it spent more than the business generated, which seems to have triggered the market selloff.

That constant change is teaching enterprises something important. The market fluctuations, the jockeying for position, the government ban – all are signals that none of these AI bets is certain. While implementing AI is critical, selecting any one of these providers as the company’s sole gateway to AI poses an additional risk that companies may not be accounting for.

Rather than focusing on a single AI model or seeking out the most advanced AI use cases – no matter how high the usage costs – companies would do well to mitigate the risk of putting all their AI eggs in one basket and focus more on what specific tasks they can do to increase revenue and which specific actions they can take with AI that will get them there.