Alphabet is making billions of dollars selling homegrown AI chips it once reserved for internal use, but building inventory to meet surging demand is draining cash as the tech giant spends heavily on data centres. The trade-off has emerged as one of the more consequential developments in the company’s recent financial disclosures, and it captures a tension that runs through the entire AI economy: the businesses generating the most growth are also the ones consuming the most capital, and the timing of when that capital returns is anything but certain.
Unlike its traditional search and cloud businesses, selling AI chips that compete with Nvidia‘s requires Alphabet to spend on hardware well before customers pay, putting pressure on cash flow. That distinction is easy to overlook in headline revenue figures, but it matters enormously to how the business behaves quarter to quarter. The company is effectively manufacturing inventory on speculation, building tensor processing units that have no confirmed buyer at the moment of production, and carrying that cost on its books until a customer signs. Traditional capital expenditure buys infrastructure the company itself will use. This is something closer to a hardware business, with all the working capital demands that entails.
The trade-off, outlined by finance chief Anat Ashkenazi on Alphabet’s post-earnings call on Wednesday, offers an early look at the costs of a business that has emerged as one of the company’s biggest growth areas as investors fret over AI’s threat to its lucrative search advertising business. “Where previously inventory balances were primarily some of our hardware products, now those TPU (custom chip) systems that are not yet sold will be showing up. So you see cash going out … it’s not the traditional kind of capex but rather a different cycle of cash flow generation,” Ashkenazi said. The remark was unusually candid for an earnings call, and it acknowledged what analysts had begun to suspect: the company’s cash flow profile is changing in ways that are not fully captured by conventional metrics.
That shift could pressure Alphabet just as investors grow more concerned about the mounting cost of AI. Its shares fell nearly 11% on Thursday after the company reported another quarter of cash burn and reiterated that spending on AI infrastructure would rise sharply next year. The market’s reaction is instructive. Alphabet is not being punished for weak demand; its cloud business is expanding briskly. The concern is about the durability of the returns and the length of the runway before spending translates into free cash flow. When a company tells investors that its most promising growth engine requires it to front-load enormous costs, the market tends to ask how long the front-loading will continue before it stops being an investment and starts being a liability.
Still, the processors are helping drive growth. Alphabet does not break out revenue or margins from the chips, called tensor processing units, but said their sales made a “meaningful” contribution as Google Cloud revenue more than doubled in the September quarter. That disclosure, limited as it is, tells us something important. The TPU business is no longer a side project or a cost centre buried inside the cloud division. It is a genuine revenue line, and its growth is intertwined with the broader expansion of Google Cloud, which has become the company’s most credible answer to the question of where future profits will come from if search advertising matures.
The competitive dimension deserves attention as well. Nvidia dominates the market for AI accelerators, and its margins reflect that dominance. Alphabet’s TPUs offer an alternative for customers who want to avoid single-supplier dependence or who value integration with Google’s software stack. But competing with Nvidia means competing on performance, availability, and price simultaneously, and it means holding inventory that Nvidia’s customers have historically been willing to commit to in advance. Alphabet is betting that demand for its chips will remain strong enough to justify the inventory risk. If it is right, the cash flow pressure will prove temporary and the business will mature into a durable profit centre. If demand softens or a competitor undercuts it, the unsold systems sitting on its balance sheet will become a visible problem.
There is a broader lesson here about how the AI boom is reshaping even the largest and most diversified technology companies. Alphabet spent years using TPUs internally, treating them as a competitive advantage in cost and efficiency rather than a product. Opening them to external customers was a strategic decision, and like most strategic decisions in the current environment, it came with an uncomfortable financial trade-off. The company has not indicated how long it expects the inventory build to continue, nor has it offered guidance on when cash flow might normalise. Those unanswered questions will keep analysts focused on the balance sheet rather than the income statement for the next several quarters. Whether Alphabet’s chip business becomes a model for others or a cautionary tale about the hidden costs of vertical integration in AI is a question the market has not yet resolved, and the company’s own commentary suggests it is still working out the answer.
