Alphabet’s AI Chip Business Boosts Growth but Pressures Cash Flow

Alphabet is generating billions of dollars from sales of its in-house artificial intelligence (AI) chips, known as tensor processing units (TPUs), which were initially developed for internal use. However, meeting surging demand requires substantial upfront investment in hardware and data centres, putting pressure on the company’s cash flow. Unlike its traditional search and cloud businesses, chip sales require Alphabet to build inventory before receiving payments from customers.

The growing costs of AI infrastructure have raised investor concerns, particularly as Alphabet faces questions about the long-term impact of AI on its lucrative search advertising business. Chief Financial Officer Anat Ashkenazi highlighted the changing cash flow dynamics during the company’s post-earnings call. Alphabet shares fell nearly 11% on Thursday after the company reported continued cash outflows and reiterated plans to significantly increase AI infrastructure spending next year. Despite these concerns, TPU sales made a meaningful contribution to Google Cloud’s revenue, which more than doubled in the September quarter.

Alphabet expects sales of its TPU systems, which combine chips with networking, storage and software, to increase further in the current quarter and accelerate significantly in 2027, according to Google Cloud chief Thomas Kurian. The company previously announced a deal to supply Anthropic with TPUs worth tens of billions of dollars and has reportedly held discussions with OpenAI over a similar arrangement. While Alphabet does not separately disclose chip revenue or margins, some investors believe the business could deliver attractive returns over time, with Check Capital’s Jeremy Saltzberg expressing confidence that margins could remain comparable to those of Nvidia’s graphics processing units.

Pic Courtesy: google/ images are subject to copyright