Technology

Google negative cash flow quarter follows AI infrastructure surge

Alphabet reported $119.8 billion in Q2 revenue, but AI data center spending pushed Google’s free cash flow to negative $5.8 billion.

Hana Yoshida

By Hana Yoshida · Markets Reporter

3 min read

Google negative cash flow quarter follows AI infrastructure surge
Photo: Ars Technica

Google negative cash flow became a first-time milestone in Alphabet’s second-quarter 2026 results, even as the company reported $119.8 billion in revenue. Alphabet’s earnings release showed that heavy spending on AI infrastructure outweighed operating cash flow for the quarter, drawing investor attention despite strong sales.

According to Alphabet, search remained the company’s largest business, producing $63.3 billion in revenue. Google Cloud brought in $24.8 billion, up 23.8% from the first quarter, while subscriptions, platforms and devices generated $12.9 billion and YouTube ads added $11.1 billion.

Ars Technica reported that Alphabet’s revenue beat analyst expectations, but the company’s stock still fell about 4.5% overnight after the results and continued to move lower afterward. The pressure came from a higher capital spending plan tied to AI data centers and related infrastructure.

Why did Google have negative cash flow?

Alphabet reported about $39.1 billion in operating cash flow for the quarter, a figure Ars Technica said was up roughly 40% from the same quarter in 2025. The company also reported $44.9 billion in capital spending connected to expanding its AI footprint.

That gap left Alphabet with negative $5.8 billion in free cash flow for the quarter, according to the figures reported by Ars Technica. Free cash flow is the cash left after a company pays for operations and capital investments; investors watch it because it shows how much money a business can generate without selling investments or borrowing.

Alphabet remains profitable and holds more than $100 billion in cash, Ars Technica reported. The negative free cash flow mark is notable because it is the first such quarter since Google became a public company, according to the report.

How much is Alphabet spending on AI?

Before the latest results, Alphabet had told investors it expected capital expenditures of $180 billion to $190 billion in 2026, according to Ars Technica. The company now says infrastructure spending could reach $205 billion this year.

That would be far above the $91 billion Alphabet spent in 2025 and about six times the $22 billion it spent in 2022, before the current AI buildout accelerated. Ars Technica reported that Alphabet also expects spending to rise again next year.

The spending reflects the cost of building and running data centers used for AI models. Ars Technica said AI-related capital expenditures across the industry are expected to exceed $700 billion this year, a scale that has led some investors to question how quickly the spending will pay off.

Alphabet has advantages that many AI competitors lack, including its advertising business, cloud revenue and in-house AI chips. Ars Technica reported that Google’s Tensor 8i and 8t chips are intended to improve efficiency in AI data centers.

What happens next for Google’s AI push?

Alphabet’s results suggest management is preparing investors for a period of elevated AI spending. The company’s cloud growth shows demand for AI services, but the latest quarter showed that demand did not fully offset infrastructure costs.

Ars Technica reported that Google recently delayed its flagship Gemini 3.5 Pro model and said it remains in testing with a small group of partners. The report also said other reports have raised questions about whether Google is getting the performance gains it needs to compete with models such as GPT 5.6 and Claude Mythos.

Google has also faced resignations among senior AI researchers, according to Ars Technica. With capital spending climbing and investors watching cash flow more closely, the next few months will test whether Google can turn its AI investment into stronger competitive results.

This story draws on original reporting from Ars Technica.