In This Article:
The world of artificial intelligence (AI) is poised for even more growth in the coming year as innovation takes shape. This week, the industry is focused on the Nvidia GTC (GPU Technology Conference) 2025, an event that will bring new updates from Nvidia (NVDA) CEO Jensen Huang.
Both investors and consumers are likely watching the conference for updates on the upcoming Nvidia Blackwell Ultra, the company’s newest AI chip. Although this highly anticipated product will likely spark some short-term momentum for Nvidia, the company is also facing new challenges.
For years, Nvidia has dominated the AI chip market, benefitting from the fact that leading tech companies have become increasingly dependent on its graphics processing units (GPUs) that they require to train their AI models.
But one of its competitors is doubling down on creating its own in-house chips, and if that chip demonstrates progress this year, Nvidia may be in trouble.
Nvidia must prepare for new challenges and rising competition
For years, other tech companies have been working quietly on developing their own AI chips that can ease their dependence on Nvidia. Given the high prices the leading chipmaker charges for its GPUs, this makes sense.
Now, Amazon (AMZN) is advancing toward that goal, but the company isn’t content to stop at building its own chips. The tech conglomerate’s subsidiaries include Annapurna Labs, an Israeli microelectronics startup acquired in 2015, which is building Amazon’s five nanometer Trainium 2 microprocessors.
Related: Nvidia CEO Jensen Huang could reboot AI trade with GTC address
It’s true that these custom silicon chips still aren’t as powerful as the Nvidia GPUs that are used by companies like OpenAI and Elon Musk’s xAI. However, Amazon is focused on using them to power something significant - the supercomputer being built by Amazon Web Services (AWS) titled “Project Rainier.”
An ultracluster made up of hundreds of thousands of Amazon’s custom Trainium chips, Project Rainier is expected to be ready before the end of 2025. This puts it in competition with xAI, which is currently working on its own AI supercomputer.
However, Amazon may have an early edge over its competitors as it works toward reaching this new milestone.
“Amazon’s success or failure is not riding on the raw power of each individual chip, but on a meticulously planned vertical integration in which entire data centers, down to each screw, copper wire and cooling fan, are engineered to squeeze every ounce of compute power from hundreds of thousands of Trainium 2 chips,” reports Semafor.
One of Amazon’s partners in this endeavor is Anthropic, a prominent AI startup in which the larger company has a sizable investment. Anthropic has agreed to train the next version of its popular Claude AI model, a well-liked chatbot in the tech community.
-
OpenAI rival powers toward milestone with massive implications
-
Anthropic CEO issues frightening warning on Chinese AI rival
Tom Brown, Chief Compute Officer at Anthropic, is optimistic that the partnership will prove beneficial for both parties.
“Project Rainier will help power both our research and our next generation of scaling. For our customers, this means more intelligence, lower prices, and faster speeds”, he states. “We're not just building faster AI, we're building trustworthy AI that scales.”
Amazon’s AI bet could be bad news for Nvidia
Nvidia entered 2025 at the top of the AI field, and most Wall Street analysts are still bullish on its growth prospects. However, the threat of rising competition, particularly from a company with Amazon’s resources, could cause some experts to reconsider their positions as Nvidia’s market share is increasingly threatened.
Related: Amazon eyes the future of AI with a massive new project
Amazon has been on a major AI push lately, making it clear that it wants to continue conquering the fast-growing AI market. It recently announced plans to establish a new group within AWS that will focus exclusively on agentic AI and use this technology to further enhance its already popular products.
Now, it seems to be on the verge of losing a customer making significant waves in the AI space. Semafor also notes that “Anthropic, boosted by an $8 billion investment from Amazon on its way to a $60 billion valuation, used Google Tensor Processors and Nvidia GPUs to train the previous versions of Claude models.”
As Anthropic trains the next version of Claude on Project Rainier, it could compel other Nvidia customers to take similar action, lowering their own dependencies on Nvidia’s expensive chips.
Related: Veteran fund manager unveils eye-popping S&P 500 forecast