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

Nvidia took San Jose via typhoon this yr, with a record-breaking 25,000 attendees flocking to the San Jose Conference Heart and surrounding downtown structures. Many workshops, talks, and panels had been so packed that individuals needed to lean towards partitions or sit down at the flooring — and endure the wrath of organizers shouting instructions to get them to line up correctly.

Nvidia lately sits on the most sensible of the AI international, with record-breaking financials, sky-high benefit margins, and no severe competition but. However the coming months additionally hang extraordinary possibility for the corporate because it faces U.S. price lists, DeepSeek, and transferring priorities from most sensible AI consumers. 

At GTC 2025, Nvidia CEO Jensen Huang tried to mission self assurance, unveiling robust new chips, private “supercomputers,” and, in fact, in reality lovely robots. It used to be an exhaustive gross sales pitch – one aimed toward buyers reeling from Nvidia’s nosediving inventory.

“The extra you purchase, the extra you save,” Huang stated at one level right through a keynote on Tuesday. “It’s even higher than that. Now, the extra you purchase, the extra you’re making.”

Inference increase

Greater than the rest, Nvidia at this yr’s GTC sought to guarantee attendees – and the remainder of the sector staring at – that call for for its chips gained’t decelerate anytime quickly. 

Right through his keynote, Huang claimed that almost the “complete international were given it unsuitable” on conventional AI scaling falling out of fashion. Chinese language AI lab DeepSeek, which previous this yr launched a extremely environment friendly “reasoning” fashion known as R1, brought on fears amongst buyers that Nvidia’s monster chips would possibly now not be essential for coaching aggressive AI. 

However Huang has again and again insisted that power-hungry reasoning fashions will, if truth be told, force extra call for for the corporate’s chips, now not much less. That’s why at GTC, Huang confirmed off Nvidia’s subsequent line of Vera Rubin GPUs, claiming they’ll carry out inference (this is, run AI fashions) at kind of double the velocity of Nvidia’s present absolute best Blackwell chip.

The risk to Nvidia’s trade Huang spent much less time addressing used to be upstarts like Cerebras, Groq, and different low cost inference {hardware} and cloud suppliers. Just about each hyperscaler is growing a customized chip for inference, if now not coaching, as smartly. AWS has Graviton and Inferentia (which it’s reportedly aggressively discounting), Google has TPUs, and Microsoft has Cobalt 100.

Alongside the similar vein, tech giants lately extraordinarily reliant on Nvidia chips, together with OpenAI and Meta, wish to cut back the ones ties by way of in-house {hardware} efforts. In the event that they – and the aforementioned different opponents – are a hit, it’ll virtually usually weaken Nvidia’s stranglehold at the AI chips marketplace.

That’s in all probability why Nvidia’s percentage worth dipped round 4% following Huang’s keynote. Traders may’ve been maintaining out hope for “one very last thing” — or in all probability an speeded up release window. In spite of everything, they were given neither.

Tariff tensions

Nvidia additionally sought to allay worries about price lists at GTC 2025.

The U.S. hasn’t imposed any price lists on Taiwan (the place Nvidia will get maximum of its chips), and Huang claimed price lists wouldn’t do “important harm” within the brief run. He stopped in need of promising that Nvidia could be protected from the long-term financial affects, on the other hand — no matter shape they in the end take.

Nvidia has obviously gained the Trump Management’s “The usa First” message, with Huang generation/2025/mar/20/nvidia-us-manufacturing-ceo-jensen-huang-donald-trump”>pledging at GTC to spend masses of billions of greenbacks on production within the U.S. Whilst that will lend a hand the corporate diversify its provide chains, it’s additionally an enormous price for Nvidia, whose multitrillion-dollar valuation is dependent upon wholesome benefit margins.

New trade

Because it seems to seed and develop companies instead of its core chips line, Nvidia at GTC drew consideration to its new investments in quantum, an trade that the corporate has traditionally not noted. At GTC’s first Quantum Day, Huang apologized to the CEOs of main quantum firms for inflicting a minor inventory crash in January 2025 after he steered that the tech wouldn’t be very helpful for the following 15 to 30 years.

On Tuesday, Nvidia introduced that it could open a brand new heart in Boston, NVAQC, to advance quantum computing in collaboration with “main” {hardware} and instrument markers. The middle will, in fact, be supplied with Nvidia chips, which the corporate says will allow researchers to simulate quantum methods and the fashions essential for quantum error correction.

Within the extra fast long term, Nvidia sees what it’s calling “private AI supercomputers” as a possible new revenue-maker. 

At GTC, the corporate introduced DGX Spark (up to now known as Undertaking Digits) and DGX Station, either one of that are designed to permit customers to prototype, fine-tune, and run AI fashions in a variety of sizes on the edge. Nor is precisely reasonably priced – they retail for hundreds of greenbacks – however Huang boldly proclaimed that they constitute the way forward for the private PC.  

“That is the pc of the age of AI,” Huang stated right through his keynote. “That is what computer systems must seem like, and that is what computer systems will run someday.”

We’ll quickly see if consumers agree.