Alphabet, the parent company of Google, is preparing to raise $80 billion through a stock sale to fund its AI infrastructure buildout. It's the largest equity raise in tech industry history—and a sign of just how expensive the AI arms race has become.

The Guardian reports that the funds will go toward data centers, specialized AI chips, and the massive computing infrastructure needed to train and deploy frontier AI models. For context, training a single large language model can cost hundreds of millions of dollars. Alphabet is planning for a future where it's training dozens simultaneously.

This follows similar moves by competitors. Microsoft has committed over $100 billion to AI infrastructure over the next three years. Amazon is spending tens of billions on its own data center expansion. The message is clear: whoever controls the most compute wins.

But here's what makes this particularly interesting—and risky. Unlike previous tech infrastructure buildouts, there's no guarantee this investment will pay off. The entire AI industry is betting that frontier models will generate enough revenue to justify the astronomical costs. If that bet is wrong, someone's going to be holding a very expensive bag.

Warren Buffett's Berkshire Hathaway, historically one of Apple's largest shareholders, has been notably reducing its tech exposure lately. That should give investors pause. When the world's most successful value investor is pulling back from tech, maybe pouring $80 billion into AI infrastructure isn't the sure thing everyone assumes.

Then again, maybe Alphabet has no choice. In the AI race, standing still means falling behind. The technology is impressive. The question is whether the economics make sense.