Alphabet (GOOGL) stock edged up 1.17% as reports emerged that Google is aggressively expanding its custom AI chip business, taking direct aim at Nvidia’s (NVDA) dominance in the data center market.
Google’s TPUs were built for internal use back in 2013, when chief scientist Jeff Dean ran the numbers and realized Google would need to double its entire computing infrastructure just to handle an AI speech model at scale. The answer was custom silicon.
Now, those chips are the centerpiece of an external commercial push.
At Lake Mariner on the southern shore of Lake Ontario, Google has put $3.2 billion on the line through a financial guarantee for an AI data center project. The facility, developed by TeraWulf and FluidStack, will lease compute from thousands of Google TPUs to Anthropic.
It’s the same move Nvidia has run for years — use balance sheet muscle to help data centers raise cheaper debt, then watch the chip orders follow.
Google isn’t stopping there. It’s also backstopping a $7 billion Anthropic project called River Bend near Baton Rouge, Louisiana, and providing $1.4 billion in guarantees for an AI computing lease in Colorado City, Texas.
In May, Google escalated further. It announced plans to sell TPUs directly to customers for the first time and unveiled its first inference-specialized chip — a product that will go head-to-head with Nvidia’s new Groq 3 LPU.
Mark Lohmeyer, VP of AI and computing infrastructure for Google Cloud, said the inference chip has pulled in customers that wouldn’t have considered TPUs before.
Citadel Securities is one of them. The firm says it can run key workloads at 30% lower cost and up to four times faster using TPUs.
Google also struck a $5 billion deal with Blackstone to create a new cloud-services company to compete directly with CoreWeave and Nebius — both Nvidia-backed platforms that run exclusively on Nvidia hardware.
Nvidia CEO Jensen Huang hasn’t been quiet about it. In April, he said Anthropic is Google’s only real external TPU customer and challenged Google to demonstrate a cost advantage.
“Our market reach is far greater than any TPU or ASIC can possibly have,” Huang said.
Still, analysts are watching. Stacy Rasgon of Bernstein noted Google is being “more opportunistic and more aggressive” about monetizing its chip capabilities than it was a few years ago.
Internally, the push is being driven by Amin Vahdat, promoted in December to lead Google’s AI infrastructure build-out. He reports directly to Google Cloud head Thomas Kurian and CEO Sundar Pichai. People who’ve worked with him say he’s relentless about performance gains, pushing engineers to improve chip functions by 10% continuously.
Google this month also announced plans to raise $85 billion in equity, largely to fund AI infrastructure.
Vahdat framed it simply: “There’s so much demand out there.”
The post Alphabet (GOOGL) Stock: Google Is Using Nvidia’s Own Moves Against It appeared first on CoinCentral.
