Is Google investing too much in data centers for AI? Its CEO doesn’t think so

  • Alphabet revenue was up 14% year on year and profit jumped nearly 29%
  • Capex, primarily focused on servers and data centers, nearly doubled as the company works to build out AI infrastructure
  • Executives said they're not worried about the possibility that they could end up building more compute than is needed

Google parent Alphabet posted another monster quarter, once again reporting double-digit increases in revenue and profit in Q2. But analysts were quick to notice one other standout figure: capital expenditures. The company nearly doubled its capex year on year, increasing from $6.9 billion in Q2 2023 to $13.2 billion in the recent quarter. CFO Ruth Porat attributed the increase to spending on servers and data centers, but the dramatic jump left analysts wondering how much is too much.

Asked explicitly on the company’s earnings call whether executives were worried they might end up building a little too much compute capacity, CEO Sundar Pichai said no.

“When we go through a curve like this, the risk of under-investing is dramatically greater than the risk of over-investing for us here, even in scenarios where if it turns out that we are over-investing…these are infrastructure which are widely useful for us,” he explained.

He continued: “They have long useful lives, and we can apply it across, and we can work through that. But I think not investing to be at the front here, I think, definitely has much more significant downside.”

Porat added that the company expects capex to be “at or above” $12 billion throughout 2024.

New Street Research analysts said in a research note that Alphabet’s decision to spend big “makes a lot of sense given: 1) the company’s lead in custom silicon through TPUs (which are the most GPU-competitive custom silicon available today) and 2) Google’s ability to directly monetize the investments through public-facing cloud services while simultaneously benefiting from the direct impact on its core consumer business, including AIOs and Gemini Advanced/Google One AI Premium.”

Milestone metrics

Consolidated revenue increased 14% year on year to $84.7 billion, while net income jumped nearly 29% to $23.6 billion. Google Cloud hit milestones in both revenue and operating income, crossing the $10 billion and $1 billion marks, respectively. Cloud revenue rose from $8 billion to $10.3 billion, with operating income hitting a total of nearly $1.2 billion.

Though Alphabet didn’t break out artificial intelligence revenue, Pichai said that year to date the company’s AI infrastructure and generative AI products have “already generated billions in revenues and are being used by more than 2 million developers.”

Pichai added Google Cloud is “innovating at every layer of the AI stack,” including everything from chips to agents.

“While it lacks the scale of AWS and Azure still, we think proprietary infrastructure like TPUs are a key differentiator for Google Cloud,” New Street’s team wrote. “We have long believed that Google Cloud has a considerable advantage in their proprietary TPU and Generative AI tooling infrastructure (Vertex, Gemma, etc.) and that this advantage is disproportionately favorable to Google Cloud relative to their traditional offering versus Cloud peers.”