“I hope you realize this is not a concert,” Nvidia (NVDA) CEO Jensen Huang said as he took to the stage at GTC exactly a year ago, the first in-person version since 2019. But after a troubled start to 2025 for the stock, new products slated to be revealed at this year’s event in California could well be music to investors’ ears.
Light Up your Portfolio with Spark:
Fast forward a year from that event – and some major changes in the AI space in the intervening period – and Huang is expected to unveil the company’s latest chip at its GTC developers conference in San Jose, California, dubbed the AI Woodstock. It’s thought that Huang will use his keynote address today at 1 p.m. ET to reveal the Blackwell Ultra, a souped-up version of the firm’s blockbuster AI chips.
The company is also expected to release details about Rubin, the chipmaker’s next AI graphics processor, which could deliver an even bigger jump in computing power. Rubin is slated for launch in 2026 and could deliver the kind of performance improvement that Blackwell delivered when it was the talk of GTC last year.
We’re Gonna Need a Bigger GPU
GTC, which kicked off on Monday and runs all week, is the chipmaker’s major showcase event and has in the past been used as the launchpad for major product updates. Last year Blackwell took center stage – successor to the hugely successful H100, which was in short supply after the AI boom took off in 2023.
“We need bigger GPUs. So ladies and gentlemen, I’d like to introduce you to a very big GPU,” said Huang as he introduced the Blackwell chip.
Now Nvidia may need even bigger GPUs again. Blackwell solved one problem: as AI took off it led to a marked shift in what was required: operating the underlying large language models (LLMs) became a lot more important than providing the tools to train them, a process known as “inference.”
Blackwell delivered 30 times faster performance than the previous generation on AI inferencing – five-star Citgroup analyst Atif Malik reckons the Rubin generation could do the same.
The pace of change in the AI requires Nvidia to keep pushing the boundaries with its chips. The stock has had a hard time lately and investors will want to see progress. Competition is intense and stiffening. Tech companies like Amazon (AMZN) and Alphabet (GOOGL) are developing their own AI inference chips to compete with NVDA’s as they try to wean themselves off relying too much on the chipmaker. And breakthroughs in China suggest AI can be done well without the need for the kind of expensive chips Nvidia makes.
Huang is also expected to appear on a panel about the future of quantum computing on Thursday. While NVDA has kept things pretty tight on quantum computing, it has set March 20th as its first Quantum Day.
Is NVDA a Buy?
On Wall Street, NVDA has a Strong Buy consensus rating, based on 39 Buys and three Holds. The average NVDA price target of $177.23 implies over 48% upside to current levels.