Google is testing whether its custom AI chips can become national infrastructure instead of remaining mostly a cloud advantage. A September 22 report from Seoul Economic Daily said Google met South Korean officials on September 9 to discuss supplying tensor processing units for the country’s sovereign-AI buildout. That puts Alphabet Inc. (NASDAQ:GOOGL) into a market where governments have leaned heavily on NVIDIA Corporation (NASDAQ:NVDA) GPUs, and it matters because Korea wants ess dependence on a single foreign supplier and more control over its compute strategy.
The important change is where the TPU can live
Alphabet Inc. has spent years using TPUs to lower the cost of running its own models and Google Cloud workloads. Supplying TPUs more directly to universities, research organizations or on-premise sovereign installations would widen the addressable market and weaken one of Nvidia’s strongest structural advantages: Nvidia has long dominated high-performance accelerators outside hyperscaler clouds. Google reportedly argues that TPUs can lower AI-compute costs materially, but those claims depend on workload, software support and utilization.
The bear case for Google is interoperability. Nvidia has spent years turning CUDA, networking and systems software into a platform that developers already know. A government can want diversification and still decide that migration costs or software maturity make GPUs the safer choice. Google also has to support external customers without diluting the economic advantage that helped differentiate its own cloud.
Nvidia can lose share without losing the market
For NVIDIA Corporation, sovereign AI remains a huge opportunity because countries are treating compute as strategic infrastructure. The threat is that the same strategic logic encourages governments to avoid a single vendor. A Korean TPU deployment would not mean Nvidia has been displaced, but it could establish a template for splitting national clusters across competing architectures.
Insider Monkey’s database showed 275 hedge funds holding Alphabet at the end of Q2 2026, up from 265 in Q1. Berkshire Hathaway increased its Class A Alphabet position about 45% during the quarter. Nvidia’s holder count rose to 285 from 275, while Fisher Asset Management increased its stake about 3%. Both filings predate the September Korea discussions. Alphabet short interest was 77.70 million shares as of August 31, around 0.72% of float with 3.48 days to cover.
The investable question is therefore broader than whose chip is faster. Google is trying to export a hardware advantage that used to reinforce Cloud, while Nvidia is defending the default architecture for almost everyone else. If sovereign buyers insist on second sources, both can grow while Nvidia’s share of incremental spending slips. That is a much more credible competitive threat than expecting governments to rip out existing GPU stacks overnight.
READ NEXT:
The Trillion-Dollar AI Capex Gap: Why Too Much Hardware Could Be Nvidia’s Trap and Microsoft’s Opportunity and Redditors Are Skeptical of SpaceX’s Orbital Data Centers. Are They Right?
Hedge Fund Resource Center