Broadcom shares climbed 6% in extended trading on Wednesday after Google reported earnings and surging capital expenditures for artificial intelligence.

Google said on Wednesday that it expected to spend as much as $185 billion on capital expenditures this year, which is nearly double what it spent last year.

Ben Reitzes, Melius Research head of technology research, said the capex spend would be a boon for Broadcom and other names tied to Alphabet.

“That is an incredible number. We are laughing because that number is so good for the Google cohort,” Reitzes told CNBC’s “Closing Bell Overtime” following the release.

Google is among a handful of technology companies that are increasing spending on capex in order to build data centers focused on artificial intelligence.

Much of Google’s AI software doesn’t run on industry-standard Nvidia chips, but instead on its own tensor processing units. For example, Google’s state-of-the-art Gemini 3 model was brained on TPUs. Broadcom helps Google make its TPUs.

Broadcom has a burgeoning custom business focusing on chips called ASICs, which some experts believe may be more efficient for some artificial intelligence workloads. In December, Broadcom said that it would sell Google’s TPU Ironwood rack systems to Anthropic, another artificial intelligence lab.

Experts say that custom AI chips only make sense for the biggest and most sophisticated firms, which are often called hyperscalers. Broadcom refers to custom chips it is currently developing for five separate customers as “XPUs.”

In addition to Google’s TPU, Microsoft, Amazon, and Meta are working on their own custom chips, although Broadcom has not named customers beyond Google and Anthropic.

Hyperscalers typically need a partner in the semiconductor industry like Broadcom to add necessary intellectual property and to help the chips get manufactured.

Google also uses Nvidia chips. Nvidia shares rose 2% in extended trading.

“It’s probably good for Nvidia too because they are going to spread the love, not just their own TPU but also to Nvidia,” Reitzes said.