Nvidia’s recent GTC 2025 conference in San Jose marked a record-breaking event with over 25,000 attendees flocking to the convention center. The company showcased its powerful new chips, “supercomputers,” and robots, aiming to reassure investors about demand for its products.
CEO Jensen Huang projected confidence, claiming that traditional AI scaling was not falling out of vogue. He highlighted Nvidia’s next-gen Vera Rubin GPUs, which will perform inference at double the rate of current best Blackwell chip. However, Huang acknowledged the threat from low-cost inference hardware and cloud providers like Cerebras, Groq, and AWS.
Tech giants OpenAI and Meta are also reducing their reliance on Nvidia chips by developing in-house hardware. If successful, this could weaken Nvidia’s stranglehold on the AI chips market. This concern led to a 4% dip in Nvidia’s share price following Huang’s keynote.
Tariff tensions were also addressed, with Huang stating that tariffs wouldn’t cause significant damage in the short run. However, he didn’t promise protection from long-term economic impacts.
Nvidia has committed to spending hundreds of billions of dollars on US manufacturing to diversify its supply chains. The company also announced new investments in quantum computing, including a Boston center for collaboration with leading hardware and software markers.
The launch of “personal AI supercomputers” like DGX Spark and DGX Station aims to generate new revenue. These devices allow users to prototype, fine-tune, and run AI models at the edge, but they are not inexpensive. The future of these products remains uncertain, and customers will need to decide if they align with Nvidia’s vision for the future of computing.
Source: https://techcrunch.com/2025/03/20/gtc-felt-more-bullish-than-ever-but-nvidias-challenges-are-piling-up