TLDR

Nvidia stock fell 3.2% in premarket trading after reports Meta may adopt Google’s TPU chips by 2027
Alphabet shares rose 2.1% as Meta explores using tensor processing units in data centers and cloud services
Google previously secured a deal with Anthropic for up to 1 million TPUs, validating the technology
Meta’s projected $70-72 billion AI infrastructure spending makes its chip choices highly influential
Tech companies are actively seeking alternatives to reduce reliance on Nvidia’s dominant GPU market position

Nvidia shares tumbled 3.2% in premarket trading Tuesday following reports that Meta is exploring Google’s AI chips. Alphabet gained 2.1% as investors assessed the potential challenge to Nvidia’s GPU dominance.

NVIDIA Corporation, NVDA

The Information reported Monday that Meta is in discussions to deploy Google’s tensor processing units in its data centers by 2027. Meta may also begin renting TPUs from Google Cloud next year, according to sources familiar with the talks.

The news represents a potential breakthrough for Google’s custom chip business. TPUs launched in 2018 for internal Google Cloud operations but have since evolved into a competitive alternative for AI workloads.

Google Gains Ground in Chip Market

Landing Meta as a customer would validate Google’s TPU technology and establish it as a serious Nvidia alternative. Meta ranks among the world’s largest AI infrastructure spenders, with projected capital expenditure of $70-72 billion this year.

Google recently closed a major deal with Anthropic for up to 1 million TPUs. Seaport analyst Jay Goldberg called it “really powerful validation” for the technology. He expects more companies to evaluate TPUs following high-profile partnerships.

TPUs differ from Nvidia’s GPUs in fundamental ways. Graphics processing units were originally built for video game rendering but proved effective for AI training. TPUs are application-specific integrated circuits designed specifically for machine learning tasks.

Diversification Push Reshapes Market

Tech companies have accelerated efforts to diversify chip suppliers beyond Nvidia. While Nvidia maintains market leadership, concerns about over-reliance have grown industry-wide.

Google’s advantage comes from its internal AI development. Teams working on models like Gemini provide direct feedback to chip designers. This creates a refinement cycle that competitors struggle to match.

Bloomberg Intelligence analysts estimate Meta will spend $40-50 billion on inferencing chip capacity next year alone. That assumes total 2026 capital expenditure of at least $100 billion.

The analysts suggest Google Cloud could see accelerated growth from enterprise customers wanting TPU and Gemini model access. Both require using Google’s cloud platform.

Asian suppliers connected to Alphabet surged in early trading. South Korea’s IsuPetasys jumped 18% to a new intraday record. Taiwan’s MediaTek rose nearly 5%.

Advanced Micro Devices remains a distant second to Nvidia in GPUs. Google’s TPUs introduce a third competitive option for companies building AI infrastructure.

The tensor chips were adapted as AI accelerators from Google’s internal applications. Because Google and DeepMind develop cutting-edge models, the company continuously improves TPU designs based on real-world performance data.

Meta and Google representatives declined to comment on the reported discussions. Deal terms and timing remain under negotiation, with final agreements yet to be reached.

The post Nvidia (NVDA) Stock Drops as Meta Considers Google Chips for 2027 Rollout appeared first on Blockonomi.

---

Filed under: Bitcoin - @ November 25, 2025 12:25 pm