This week, Fortune, HackerNews’ front page, and Yale School of Management all said the same thing: the AI bubble is bursting. Salesforce has lost 30% of its market cap since January. ServiceNow likewise. OpenAI is showing ads in ChatGPT – something Sam Altman once called a “last resort”. Anthropic is raising prices and cutting usage limits.
Many business leaders will react with concern. “Should we slow down our AI efforts?”
The answer is the opposite.
Two different AI markets
There are two entirely separate AI markets right now, and most people conflate them.
One is the investment market. That’s where Nvidia trades at fantasy multiples, OpenAI seeks a $730 billion valuation, and tech giants burn through $539 billion in capex they don’t yet have customers for. That market has a problem. A big one.
The other is the usage market. That’s where a company with 50 employees uses AI to automate data processing, draft reports, or handle customer inquiries. That market has never been better.
The bubble is in the first market. The value is in the second.
Why the bubble is good news for you
When investors lose their appetite, something predictable happens: AI companies start competing on price and quality instead of competing for the next fundraise. That’s exactly what we’re seeing now.
Google’s new Gemini models are dramatically cheaper than a year ago. Anthropic and OpenAI are cutting prices quarter by quarter. Open source models from Meta and Alibaba are pushing the entire market. For companies buying AI as a service – which is most of them – prices are falling while quality is rising.
That’s the opposite of a bubble. That’s a market maturing.
What’s changing – and what isn’t
Capital Economics’ chief markets economist John Higgins points to something interesting: the bubble may not be in stock prices but in earnings themselves. Tech giants’ profits have grown so explosively that the earnings themselves look like a bubble. What happens when customers won’t pay what the pricing demands?
That question doesn’t hit you. Most companies pay a few hundred dollars a month for AI tools – not billions in infrastructure. The risk is concentrated with those building the foundation, not those using it.
What’s changing is the market for AI vendors. There will be fewer of them, and the survivors will be those that actually make money. That means more stable partners, better support, and products built to last – not to impress investors.
What isn’t changing is the technology’s capability. A language model that can analyse contracts, write code, or summarise meetings doesn’t get worse because Nvidia’s stock price drops. The models are trained. The infrastructure is built. It doesn’t disappear.
The right response
The worst response to AI bubble news is to slow down. “Let’s wait and see.” That’s exactly what your competitors hope you’ll do.
The right response is to distinguish between investment and usage. You don’t need to bet on whether Anthropic wins the war against Google. You need to know, concretely, which processes in your company can be improved with AI – and then do it. With the tools available now, at the prices that apply now.
McKinsey’s latest figures show that 88% of companies use AI regularly. But Harvard Business Review reports that adoption is stalling – not because the technology is failing, but because employees are nervous about being replaced.
That’s a leadership problem, not a technology problem. And it won’t be solved by waiting for the bubble to land.
What you can do right now
Start by answering one question: Which three processes in your company cost the most time and create the least value?
That’s where AI has the greatest impact. Not in grand visions of transformation, but in the boring, repetitive work that nobody wants to do but everyone depends on. Data entry. Reporting. Sorting enquiries. Quality checks.
The companies winning with AI aren’t the ones investing the most. They’re the ones starting with the right problems.
We help companies find exactly the processes where AI makes sense – and implement it without breaking the organisation. Read more about our approach to AI strategy and advisory, or see how it works in practice in our cases.
Read also:
