Big Tech doubles down on even bigger AI spend

Quarterly results are in for Microsoft, Alphabet and Meta — and while the numbers are mixed, they all agree on big capital expenditures — needed for building data centers — for fiscal year 2025.

The coming AI wars will be fought with data centers and gigawatts, and nobody wants to lose out.

Spending big
Meta reported an 83% drop in income this quarter, and saw its stock drop 8% in after hours trading. That did not stop its AI expansion, offering an updated guidance of $70-72 billion in yearly capex, up from $66-72.

This is needed for building out the huge data centers CEO Zuckerberg has committed to, along with huge wage bills for his Superintelligence lab.

Alphabet reported net income up 33% and saw a jump in its after hours stock of over 5%.

And bigger
In their quarterly report they suggest growth and cloud demand increases will now lead to capital expenditures in the range of $91 to $93 billion.

Google has recently entered deals with OpenAI and Anthropic, and their own Gemini models have over 650 million monthly users.

Microsoft are eerily quiet on expenditures and AI in their quarterly report, citing only a $3B «impact from investment in OpenAI.»

$35 billion per quarter
Their reported income was up a mere 12% and the stock fell about 4% in late trading.

Reuters snagged the details from their earnings call, however, where they say capex for the third quarter was at $35 billion and said spending would rise this year, rather than moderate.

$35B times three is about $100 billion for the year total, which makes Microsoft the biggest spender on capex of the currently reporting Big Tech companies.

Are we in a bubble?
This all drowns in comparison to OpenAI’s spend on data centers, expected to total over a trillion dollars in the coming years.

New reporting suggests OpenAI might go public as soon as 2027.

With all this spending, talk of bubbles are getting ever louder, but as Nvidia reaches the world’s highest stock market evaluation at $5 trillion, CEO Jensen Huang dismisses the idea, saying,

— I don’t believe we’re in an AI bubble. All of these different AI models we’re using—we’re using plenty of services and paying happily to do it.

Read more: Wired, BBC and Business Insider.