Cloud Backlog Hits $1.5T: Why GOOG Is Eating AWS's Lunch

Three of the Magnificent Seven are walking into earnings this week, and the Bloomberg Podcasts team has a spicy take: this isn't really an earnings story. It's an infrastructure arms race, and one company is pulling ahead.

$GOOGL, $MSFT, and $AMZN collectively sit on a cloud backlog exceeding $1.5 trillion. That's not a typo. The demand engine is largely fueled by AI frontier labs, with OpenAI and Anthropic alone committing eye-watering sums to these platforms. Anthropic's $750 billion spend commitment is the number everyone should be screenshotting right now, and Google Cloud is reportedly capturing the largest share of it.

Here's why that matters: Google Cloud's edge isn't just relationships. It's structural. Bloomberg's analysts point to vertical integration with TPUs (Google's custom AI chips), lower-cost infrastructure, and the fact that unlike its rivals, Alphabet doesn't have to rent its own brain. While AWS built its empire on breadth and MSFT Azure rides the enterprise Office 365 flywheel, Google Cloud is increasingly the choice for AI-native workloads where cost per compute matters most.

The bull case for $GOOGL heading into earnings is basically this: capex increases are not a red flag here the way they might be at a pure software company. Bloomberg's analysts note that the hyperscalers can point directly to backlog as justification, a luxury Meta didn't have when it announced its own spending surge. Bears will argue that Google Cloud still trails AWS and Azure in total market share, and that TPU advantages could narrow as competitors build out their own silicon. Fair point. But the trajectory is hard to argue with.

Meanwhile, the Patrick Boyle deep dive on YouTube lays out the uncomfortable flip side of all this ambition: the US already has 65 gigawatts of data center capacity with another 413 gigawatts planned, equivalent to powering 330 million new homes. Three out of four Americans oppose local data center construction. The AI buildout is being financed by debt secured against rapidly depreciating chips. That's a lot of things that could go sideways.

So yes, the cloud backlog is enormous, the AI demand is real, and Google Cloud has a genuine structural advantage. But the entire trade is one rate shock, one grid failure, or one regulatory overreach away from looking very different. The scoreboard this week will tell us a lot about who's actually winning, and who's just spending like they are.