Google and Meta have reportedly joined hands to challenge Nvidia in the AI race through software and hardware development

Google is allegedly working with Meta to challenge Nvidia’s dominance in AI software and hardware development. Rumors suggest that the initiative is part of Google’s ambitious plan to advance its Tensor Processing ​Units as a suitable alternative to Nvidia’s industry-leading GPUs.

Alphabet Inc., a multinational conglomerate and parent company that owns Google, is reportedly working with the U.S.-based Social Media conglomerate Meta to advance artificial intelligence and challenge Nvidia’s monopoly in the sector.

NVIDIA’s stock drops by 6.21% as Google and Meta join efforts to weaken its AI dominance

According to people familiar with the matter, Google and Meta are reportedly working on the development of TorchTPU, which aims to make Google’s TPU chips fully compatible with PyTorch, the leading AI software framework. The initiative aims to streamline the adoption of TPU chips, favoring customers who already have PyTorch software infrastructure. TPU adoption was heavily inhibited by its poor compatibility with PyTorch. 

NVIDIA’s stock price has plummeted by 6.21% in the last 24 hours, and the stock is currently trading at $171.46. The stock is down 8.12% in the last month but has surged by 27.84% YTD, according to data from real-time stock price tracking platform Google Finance. 

The sources also revealed that Google has devoted more resources and organizational focus to the development of TorchTPUs due to the growing demand for TPU chips among tech companies. The company is also opting to open-source parts of the software to accelerate adoption among customers, according to the anonymous tips. AI developers utilize PyTorch to develop AI models, and Meta provides significant support for its development.

The partnership aims to challenge Nvidia’s dominance and monopoly over the AI sector. The chip maker currently dominates the AI processor industry through its CUDA software ecosystem. CUDA is Nvidia’s strongest shield against rising competition, according to some Wall Street analysts. The CUDA software is deeply integrated into PyTorch and has become the default method for training and running large AI models. 

The rumors also revealed that Meta has a strategic interest in developing software compatible with TPUs. Meta reportedly aims to diversify its AI infrastructure away from Nvidia’s monopoly and lower inference costs.

The partnership aligns with Google’s objective to make its TPUs a suitable alternative to Nvidia’s GPUs, which have gained a monopoly in the sector. TPU revenue has significantly boosted the company’s cloud revenue, as Google seeks to demonstrate to investors that its AI investments are profitable.

Google begins commissioning TPUs for commercial supply

For the longest time, Google had reserved a majority of its TPU supply for internal use only. However, the company switched policies in 2022 when Google’s cloud computing unit gained control over the group that sells TPUs. The reforms took place amid the ongoing AI race, which has increased global chip demand and AI infrastructure. 

Google’s chips are specifically designed to run on the Jax software frameworks. Therefore, companies that have already installed PyTorch frameworks would need to reengineer their infrastructure to accommodate the Jax frameworks, a costly process that would require significant time and resources to achieve.

Once Google’s initiative to make its chips PyTorch compatible is successful, the companies will incur significantly reduced costs to modify their infrastructure to house the chips. Google requires solutions to supply Google Cloud customers, such as Anthropic, who purchase TPUs and run its products, including Google AI-powered search and the Gemini chatbot.

Google’s ties to Meta in AI technology advancement were previously uncovered by a recent Cryptopolitan report, which revealed that the two tech giants were discussing a multi-billion-dollar deal to supply Meta with Google’s AI chips.

Google’s recent moves have revealed the company’s obsession with AI technology. According to another report by Cryptopolitan, the company backed a multi-year contract deal worth $7 billion that Hut 8 signed to lease a data center in Louisiana. A press release revealed that Google will provide financial support to cover the lease payments and related pass-through obligations.

Join Bybit now and claim a $50 bonus in minutes