Nvidia took the San Jose from the storm this year, with a record of 25,000 participants flocking to the San Jose conference center and around the city’s buildings. Many workshops, conversations and panels were so full that people had to lean against the walls or sit on the floor – and suffer from organizers’ rage to shout orders to make them align properly.
Nvidia is today at the top of the world AI, with a record of economics, high quality profit margins and even serious competitors. However, the coming months also have an unprecedented risk for the company, as it faces tariffs, Deepseek and changing priorities from AI’s top customers.
At GTC 2025, NVIDIA CEO Jensen Huang tried to display confidence, revealing strong new brands, personal “supercomputers” and, of course, truly cute robots. It was an exhaustive sales step – one aimed at investors unfolding from Nvidia’s stock.
“The more you buy, the more you save,” Huang said at a point during a keynote on Tuesday. “It’s even better than that. Now, the more you buy, the more you do.”
Outburst
More than anything else, Nvidia at this year’s GTC tried to secure participants – and the rest of the world she was watching – that demand for her chips would not slow down anytime soon.
During his keynote, Huang claimed that almost the “whole world took it wrong” in the traditional escalation of AI falling out of fashion. The Chinese AI Lab Deepseek, which earlier published an extremely effective “reasoning” model called R1, has caused fears between investors that Nvidia chips may no longer be necessary for the preparation of competitive AI.
But Huang repeatedly insisted that the models of hungry power would actually lead more demand for the company’s chips, no less. That is why in the GTC, Huang showed Vera Rubin’s next GPU line, claiming that he will draw conclusions (ie, runs AI models) at about twice the current Best Blackwell Chip.
The threat to Nvidia’s business that Huang spent less time to deal with was like strokes, Groq and other low -cost material conclusions and cloud providers. Almost every supernatural develops a custom chip for conclusions, if not training, too. AWS has Graviton and Inferentia (alleged to be aggressive discount), Google has a TPU and Microsoft has a Cobalt 100.
Along the same spirit, technological giants are currently dependent on Nvidia chips, including Openai and Meta, trying to reduce these bonds through home hardware efforts. If they – and the aforementioned other opponents – are successful, it will almost certainly weaken Nvidia’s strangulation on the AI chips market.
This is perhaps the reason why the price of Nvidia’s stock sank around 4% after Huang’s central point. Investors could have kept the hope of “one last thing” – or perhaps a window of accelerated launch. In the end, they didn’t even get.
Tariff tensions
Nvidia also tried to mitigate concerns about invoices at GTC 2025.
The US did not impose any invoices on Taiwan (where Nvidia gets most of the chips) and Huang claimed that invoices would not do shortly “significant damage”. He stopped promising that Nvidia would be protected from the long-term financial impact, but just regardless of the form it takes.
Nvidia clearly received the “America First” message of Trump administration, with Huang binding to the GTC To spend hundreds of billions of dollars on US construction, while this would help the company differentiate its supply chains, it is also a huge cost for Nvidia, whose estimation of hundreds of dollars depends on healthy profit margins.
New business
Since it seems to be seeds and developing businesses in addition to the Core Chips line, NVIDIA in the GTC has drew attention to its new investment in Quantum, an industry that the company has historically neglected. On the first Quantum Day of the GTC, Huang apologized to the CEOs of the big quantum companies for causing a small stock crash in January 2025, after proposing that technology would not be very useful for the next 15 to 30 years.
On Tuesday, Nvidia announced That will open a new center in Boston, NVAQC, to promote quantum computing in collaboration with “top” hardware and software indicators. The center, of course, will be equipped with Nvidia chips, which the company says will allow researchers to simulate quantum systems and models necessary to correct quantum errors.
In the most near future, Nvidia sees what it calls “personal AI supercomputers” as a possible new revenue manufacturer.
In the GTC, the company started the DGX Spark (previously called project digit) and DGX, both of which are designed to allow users to original, refine AI models in a series of sizes aside. Nor is it exactly cheap – retail for thousands of dollars – but Huang has boldly declared that they represent the future of the personal computer.
“This is the computer of the AI era,” Huang said during his keynote. “This is what computers should look like, and that’s what computers will run in the future.”
Soon we will see if customers agree.
