AI Giants Take on Nvidia's Supremacy with New Chip Innovations

Prominent tech companies are initiating projects to create alternative AI chips in response to the disruptions caused by DeepSeek.

The entry of DeepSeek, a Chinese firm, into the AI chip sector has sparked leading tech companies to ramp up their initiatives to create their own processors and diminish their dependence on Nvidia, which presently controls around eighty percent of the global AI chip market.

Several prominent entities, including OpenAI, Meta, Microsoft, Amazon, Google, and Apple, are now implementing plans to manufacture chips deemed 'sufficient' for their forthcoming artificial intelligence models.

OpenAI, closely linked to the generative AI movement, is designing its own AI chip to reduce its reliance on Nvidia.

In January, OpenAI, in partnership with SoftBank and Oracle and under the guidance of President Donald Trump, initiated a project named StarGate.

This venture, valued at half a trillion dollars, aims to establish a substantial artificial intelligence computing framework in the United States.

Sources indicate that OpenAI is in advanced design phases and anticipates finishing the design within the next few months, after which the project will move into experimental production and testing by the Taiwanese chipmaker TSMC.

The experimental production phase is expected to incur costs in the tens of millions of dollars and last approximately six months, with plans for mass production aimed for 2026.

Meanwhile, Meta is negotiating to acquire the South Korean AI chip startup FuriosaAI, with reports suggesting that a deal could be reached soon.

Additionally, Meta has launched a new generation of AI chips, known as the Meta Training and Inference Accelerator, tailored for conventional artificial intelligence functions such as running recommendation models.

Microsoft is also creating a range of chip products designed to boost the performance of AI models and enhance its data centers.

In parallel, Amazon is developing an AI chip ecosystem for its large cloud computing division and revealed its latest AI training and execution chip, Trainium Three, in December.

Trainium Three utilizes technology from the Israeli firm Annapurna, which Amazon purchased in 2015 for $370 million, and is set to be released later this year.

Google is embedding its artificial intelligence models into the chip development process, a project spearheaded by its DeepMind division, aimed at automating the chip design process.

Apple is similarly ramping up its chip development activities.

In December, it was reported that Apple’s team, responsible for the proprietary M chip, is now at the forefront of developing its first server chip aimed at AI purposes.

This project includes collaboration with chipmaker Broadcom on network technology, with plans for mass production slated for 2026 following an agreement with TSMC.

These moves signify a strategic realignment by major technology firms to attain greater control over their computing assets and potentially challenge Nvidia's leading role in the AI chip landscape.

These efforts are perceived as a reaction to the evolving paradigm in AI model development, partly influenced by the cost-efficient method showcased by DeepSeek’s R1 model.