GTC felt more bullish than ever, but Nvidia’s challenges are piling up

NVIDIA acquired San Jose from the storm this year, with 25,000 of those present at the record that flows to the San Jose Conference Center and the surrounding city center buildings. Many workshops, conversations and paintings were so packed that people had to rely on the walls or sit on the ground – and suffer from the anger of the organizers who scream on orders to obtain them properly.

NVIDIA is currently on the top of the world of artificial intelligence, with super financial data, a high -sky margin, and there are no serious competitors yet. But the coming months also holds unprecedented risks to the company because they face customs tariffs, Deepseek, and transform priorities from the best artificial intelligence customers.

In GTC 2025, Nvidia Jensen Huang CEO tried to display confidence, detecting strong new chips, and personal “super computers”, of course, really nice robots. It was a comprehensive sales stadium – one aimed at investors who reeling from the shares of Nvidia Nossivive.

“The more I bought, the more you save,” Huang said at one point during one of the main words on Tuesday. “It is better than that. Now, the more you buy, the more you are made.”

The conclusion boom

More than anything else, NVIDIA in GTC this year has sought to confirm the attendees – and the rest of the world is monitoring – that the demand for its chips will not slow down any time soon.

During the keynote, Huang claimed that “the whole world has sinned” almost the limitation of traditional artificial intelligence. AI LAB Deepseek, which earlier this year released a high -efficiency “thinking” model called R1, prompted fears among investors that the monster chips in NVIDIA may no longer be necessary to train competitive artificial intelligence.

But Huang has repeatedly insisted that thirsty thinking models for power will actually lead more demand for company chips, not less. For this reason for GTC, HUANG has offered the next line of GPUS Vera Rubin, claiming that it will lead to reasoning (i.e., operating artificial intelligence models) at a rate of almost the rate of the best current Blackwell chips in NVIDIA.

Huang’s NVIDIA’s threat was less time to treat, such as tanks such as the brain, GROQ, and other low -cost inferences and cloud services providers. Almost all Easter develops a slide intended for reasoning, if not the training too. AWS has Graviton and Indeedia (which is said to be strongly), Google has TPUS, and Microsoft has cobalt 100.

With the same path, technology giants are currently looking for NVIDIA chips, including Openai and Meta, to reduce these relationships through the efforts of internal devices. If they succeed – and the other competitors mentioned above – they will definitely weaken the stifling of NVIDIA in the artificial intelligence chips market.

Perhaps this is the reason why the NVIDIA share price decreased by 4 % after the keyword for Huang. Investors may keep hope for a “last thing” – or perhaps an accelerated launch window. In the end, they did not get.

Introductory tensions

NVIDIA has also sought to calm concerns about the tariffs in GTC 2025.

The United States has not imposed any tariff on Taiwan (where NVIDIA gets most of its chips), and Huang claimed that the definitions will not lead to “great damage” in the short term. The promise has stopped NVIDIA from long-term economic impacts, however-whatever the form they eventually take.

NVIDIA clearly received the “America First” message to the Trump administration, where Huang CTC pledged to spend hundreds of billions of dollars to manufacture in the United States while this would help the company diversify its supply chains, which is also a huge cost for NVIDIA, whose evaluation of billions of dollars depends on healthy profit.

New work

Since it looks forward to a seed and developing business other than the basic chips line, NVIDIA in GTC drew attention to its new investment in Quantum, an industry that has been historically neglected. On the first day of the quantity in GTC, Huang apologized to the executives of the major quantum companies for causing slight destruction in January 2025 after suggested that technology would not be very useful in the next 15 to 30 years.

On Tuesday, NVIDIA announced that it will open a new center in Boston, NVAQC, to enhance quantum computing in cooperation with the signs of “leading” devices and programs. Of course, the center will be provided with NVIDIA chips, which the company says will enable researchers to simulate quantitative systems and models needed to correct quantum error.

In the most immediate future, NVIDIA sees what it is called “personal giant computers” as a possible new revenue maker.

In GTC, the company DGX Spark (which was previously called project numbers) and DGX, both are designed to allow users to preliminary models, adjust them, and operate artificial intelligence models in a range of sizes on the edge. It is not completely inexpensive – they sell thousands of dollars – but Huang has boldly declared that they represent the future of the personal computer.

“This is the computer in the era of artificial intelligence,” Huang said through the keyword. “This is what computers should look at, and this is what computers will work in the future.”

We will soon see whether customers agree.