Google's $185B AI Spending Spree Sends Broadcom Soaring 6%

One number changed everything: $185 billion. That's how much Google plans to spend on AI infrastructure this year—nearly double last year's expenditure—and it sent Broadcom shares rocketing 6% in after-hours trading Wednesday.

The Hidden Winner Emerges

While everyone expected Nvidia to be the biggest beneficiary of Google's AI spending bonanza, Broadcom stole the spotlight. The reason lies in Google's strategic choice: its cutting-edge Gemini 3 model runs on custom tensor processing units (TPUs), not industry-standard Nvidia chips.

Broadcom serves as Google's key partner in manufacturing these TPUs. The relationship runs deep—in December, Broadcom announced it would sell Google's TPU Ironwood rack systems to Anthropic, another major AI lab. "That is an incredible number," said Ben Reitzes from Melius Research. "We are laughing because that number is so good for the Google cohort."

The Custom Chip Revolution

The $185 billion investment signals a broader shift in AI infrastructure. Major tech companies are moving beyond off-the-shelf solutions toward application-specific integrated circuits (ASICs) tailored for their unique AI workloads. Microsoft, Amazon, and Meta are all developing custom chips, though only the largest "hyperscalers" can afford this approach.

Broadcom currently develops custom chips—dubbed "XPUs"—for five separate customers. While the company has only named Google and Anthropic, industry insiders suspect other tech giants are involved. These partnerships require sophisticated semiconductor expertise that only specialized firms like Broadcom can provide.

Nvidia Rides the Wave Too

Nvidia shares also climbed 2% in extended trading, proving this isn't a zero-sum game. Google uses both its custom TPUs and Nvidia chips, creating a hybrid approach that benefits multiple suppliers. "It's probably good for Nvidia too because they are going to spread the love," Reitzes noted.

This dual-sourcing strategy reflects the massive scale of AI infrastructure needs. When you're spending $185 billion, there's room for multiple winners.

The Broader Market Implications

Google's spending commitment represents more than infrastructure expansion—it's reshaping the semiconductor landscape. The shift from standardized chips to custom solutions creates new opportunities for specialized partners while potentially reducing dependence on any single supplier.

For investors, this trend raises important questions about market concentration. While Nvidia dominates general AI chips, the custom chip market remains fragmented, with Broadcom emerging as a key enabler for hyperscalers seeking differentiation.

Following the Money Trail

The $185 billion figure isn't just impressive—it's strategic. Google is essentially buying its way out of potential supply constraints while gaining performance advantages through custom silicon. This approach requires deep partnerships with companies like Broadcom that possess both the technical expertise and manufacturing relationships necessary for success.

The ripple effects extend throughout the semiconductor supply chain, from memory manufacturers to packaging specialists. When hyperscalers commit to this level of spending, entire industries realign to capture their share.