CNBC reports that Google's online dominance is showing signs of cracking in the AI era, with generative AI assistants steadily pulling queries away from the search box that has anchored the company's economics for two decades. Ads still account for roughly three-quarters of Alphabet's revenue, and those advertising margins are what let Google fund long-horizon bets like Waymo and pour close to $200 billion into AI infrastructure. The threat the article describes is twofold: Google could lose share to AI rivals, and in racing to compete it could cannibalize search in favor of a discovery model that has no proven ad engine behind it.
The cracks are still hairline, not structural. Google continues to account for roughly 80 percent of global query volume and, by Cloudflare Radar's measurement, sent close to 88 percent of all search referral traffic in May 2026. AI engines combined remain a rounding error in referral terms. What has changed is the trend line and the narrative around it. ChatGPT crossed 1 billion monthly active users, Perplexity has grown sharply as a citation-first research tool, and Google felt compelled to redesign its search box for the first time in 25 years, placing an "AI Mode" button inside the box and demoting the classic search button beneath it.
For the cooling industry, the relevant detail is that the advertising cash flow now under question is the same cash flow underwriting the largest data-center construction program in Alphabet's history. The company raised its 2026 capital expenditure guidance to a range of $180 billion to $190 billion, up from an earlier $175 billion to $185 billion, and Chief Financial Officer Anat Ashkenazi has said 2027 spending will "significantly increase" again. About 40 percent of that capex flows to data centers and networking equipment, with the balance going to servers. Those buildings need cooling, and increasingly they need liquid cooling.
Investors are already pricing the tension. On June 22, Alphabet had its worst trading day in over a year, shedding roughly $269 billion in market value after two prominent AI researchers departed for OpenAI and Anthropic and after the capex guidance moved higher. Markets are weighing whether the returns on a $190 billion spend justify the outlay while the core business that pays for it faces a credible long-term challenge. That is the exact equation that sits upstream of every cold plate, coolant distribution unit, and chiller order tied to Google's pipeline.
The cooling demand curve rests on the competitive scramble among hyperscalers to build AI capacity faster than their rivals, and that scramble is precisely what funds the data-center and liquid-cooling buildout. Google's response to the AI threat has been to spend more, which keeps procurement flowing to thermal vendors even as the search franchise wobbles. The same dynamic shows up in the broader hyperscaler capex wave that is driving data-center construction and cooling demand, where competition rather than proven AI revenue is the engine of the order book.
Google's own footprint makes the stakes concrete. The company has committed $15 billion to data-center development in Missouri, a campus-scale commitment that translates directly into thermal procurement, and it is building out a liquid-cooling supply chain that reaches into China through partners like Envicool. Each of those moves assumes the capital keeps coming. The capital keeps coming because the ad business keeps paying.
For cooling vendors, the lesson is to watch the funding source as closely as the construction announcements. A squeeze on Google's advertising economics would not stop the buildout tomorrow, but it would raise the cost of capital for every gigawatt Alphabet plans and put real pressure on the spending guidance that vendors treat as a forecast. The cooling demand curve is a derivative of hyperscaler cash flow, and right now the most profitable cash flow in technology is being asked, for the first time in 25 years, to prove it can survive the thing it is paying to build.