NVIDIA’s largest conference of the year, GTC, begins Monday and is held in San Jose on Friday. TechCrunch will cover news happening on the ground – we expect healthy announcements.
CEO Jensen Huang will be giving a keynote at the SAP Center at 10 a.m. Tuesday with the focus – what else? – AI and accelerated computing technology, According to Nvidia. The company is also making fun of revelations related to robotics, Sovereign AI,,,,, Artificial Intelligence AgentAnd Automobile – Plus 1,000 meetings, there are 2,000 speakers and nearly 400 exhibitors.
Here’s how to watch NVIDIA GTC 2025 keynote speech onlineand many other meetings, talks and panels.
So, what do we want to see in GTC? Well, NVIDIA usually keeps most of the conferences with GPU-related debuts. The company’s Blackwell chip lineup seems likely to undergo a new, upgraded iteration.
In NVIDIA’s recent earnings call, Confirmed The upcoming Blackwell B300 series (codenamed Blackwell Ultra) will be released in the second half of this year. In addition to higher compute performance, the Blackwell Ultra card also includes more memory (288GB), an attractive feature for customers who want to run and train memory AI models.
NVIDIA’s next-generation GPU series, Rubin, will almost certainly be mentioned in GTC with Blackwell Ultra. Rubin expires in 2026 and Rubin promises to deliver what Huang calls “big, big, huge improvement” in computing power.
Huang said that during the aforementioned NVIDIA earnings call, he will also talk about the Robin post-product at GTC. That could be a Rubin Ultra GPU, or it could be a GPU building from the Rubin family. (Chip is Named after Vera Rubinan astronomer who discovered dark matter. )
In addition to GPUs, NVIDIA may shed light on its approach to recent advances in quantum computing. The company arranged a “Quantum DayFor GTC[map] The path to useful quantum applications. ”
To be sure: Nvidia can use victory.
Early Blackwellka It is reported that serious overheating problemscausing customers to cut orders. In recent months, U.S. export controls and concerns about tariffs have made NVIDIA’s stock price extremely high. Meanwhile, China’s AI Lab DeepSeek’s successful development of a model that competes with leading AI labs’ models has prompted investors to worry about the demand for powerful GPUs like Blackwell.
Huang asserts that the rise of DeepSeek will actually be Jingzheng For Nvidia, because it will accelerate a wider range of AI technologies. He also noted that “inference” models of desire, such as Openai’s O1, the next mountain in Nvidia, can be climbed.
It should be clear that NVIDIA is not completely harmed. The company reported a record quarter in February with revenue of $39.3 billion and expected revenue of $43 billion in the following quarter. Although competitors such as AMD have begun to encroach on the company’s territory, Nvidia Still command It is estimated to account for 82% of the GPU market. But Huang is It is reportedly impatient View AI applications beyond the tech industry.
