At its annual developer conference, Nvidia CEO Jensen Huang addressed mounting concerns regarding the surging costs associated with artificial intelligence computing.
During the event, Huang introduced a series of advanced chips and technologies he claimed would deliver better value to customers. The highlights included a new version of Nvidia’s signature AI processor, the Blackwell Ultra, along with several additional generations projected to be released by 2027. He also presented Dynamo-branded software designed to optimize both current and future hardware, enhancing efficiency and profitability.
“It essentially acts as the operating system of an AI factory,” Huang explained during a two-hour presentation at the company’s GTC event in San Jose, California, which explored topics ranging from robotics to personal supercomputing.
The GTC conference has evolved from an obscure gathering to a key date on the calendars of technology enthusiasts and investors, with major implications for Wall Street. Throughout his speech, Huang showcased various hardware, software, and services, though no game-changing announcements were made for investors, leading to a more than three percent decline in the company’s stock on Tuesday.
Huang noted that Nvidia, originally concentrated on gaming chips, has broadened its influence across numerous sectors. The new Blackwell Ultra processor is expected to launch in late 2025, with a significant upgrade, named “Vera Rubin,” set for release in the second half of 2026.
The event featured several noteworthy announcements, including:
- A platform dubbed Isaac GR00T N1 aimed at accelerating humanoid robot development. Nvidia is collaborating with Walt Disney Co. and Google’s DeepMind on this initiative, which will be accessible to external developers.
- A strategic partnership with General Motors Co. to integrate AI within future vehicles, manufacturing processes, and robotics.
- A wireless initiative involving T-Mobile US Inc. and Cisco Systems Inc. to develop “AI-native” hardware for upcoming 6G networks, which will succeed today’s 5G technology.
- New Nvidia-powered personal supercomputers from manufacturers such as Dell Technologies Inc. and HP Inc., enabling developers and researchers to work on AI models directly from their desks.
Nvidia is at a crucial juncture, as it faces skepticism from investors regarding the sustainability of its recent two-year streak of unprecedented growth in revenue and market valuation. This strain became evident earlier in the year when Chinese startup DeepSeek claimed it could develop a competitive AI model using significantly fewer resources.
DeepSeek’s assertions raised questions about the necessity and pace of investments in AI computing infrastructure. However, Nvidia’s major clients—including Microsoft and Amazon’s AWS—have pledged to maintain their spending in 2025.
According to a report from Bloomberg Intelligence, leading data center operators—commonly referred to as hyperscalers—are expected to allocate $371 billion (approximately Rs. 32,10,561 crore) towards AI facilities and computing resources in 2025, marking a 44 percent increase from the previous year. This figure is anticipated to rise to $525 billion (around Rs. 45,43,247 crore) by 2032, accelerating faster than analysts originally predicted following DeepSeek’s viral success.
Concerns regarding trade tensions and a potential recession have adversely impacted Nvidia’s stock, which is down 14 percent this year. Shares slid 3.3 percent to $115.53 (approximately Rs. 9,997) by the end of trading on Tuesday.
Furthermore, the partnership announcement with General Motors negatively affected the stock of Mobileye Global Inc., a developer of autonomous vehicle technology, leading to a 3.5 percent drop to $14.44 (approximately Rs. 1,250). Mobileye went public in 2022 and is majority-owned by Intel Corp.
The ongoing GTC event serves as a platform for Nvidia to reinforce its position as a fundamental player in the AI sector, which Huang believes is on the brink of catalyzing a new industrial revolution. He referred to the conference as the “Super Bowl of AI.”
One of the critical issues for Nvidia is whether investment in AI will continue to rise in 2026, as pointed out by Wolfe Research analyst Chris Caso. He highlighted that while AI spending is likely not to be the first area cloud customers will cut in their budgets, any financial distress affecting funding could pressure capital expenditures.
Huang’s presentation did not fully assuage investors’ fears but laid out a strategic plan featuring future chip developments and introduced a cutting-edge system combining silicon and photonics—utilizing light waves.
Nvidia is also establishing a quantum computing research facility in Boston to explore this burgeoning field further.
Headquartered in Santa Clara, California, Nvidia has faced production challenges in its rush to enhance chip offerings. Some initial versions of the Blackwell processor required modifications, delaying their release. Although Nvidia claims that these issues are resolved, supply still lags behind demand, prompting the company to increase spending to boost output, which may affect margins this year.
Huang revealed that the four largest public cloud providers—Amazon, Microsoft, Alphabet Inc.’s Google, and Oracle Corp.—acquired 1.3 million of Nvidia’s older-generation Hopper AI chips last year, while they have already purchased 3.6 million Blackwell AI chips in 2025.
Following the introduction of Vera Rubin in 2026, Nvidia plans to roll out a subsequent model called Rubin Ultra the following year. Vera Rubin, after whom the processor is named, was an influential American astronomer who contributed to the discovery of dark matter.
According to Huang, the next generation of chips will be called Feynman, likely paying homage to Richard Feynman, an esteemed American physicist notable for his work in quantum mechanics.
© 2025 Bloomberg LP
(This story has not been edited by NDTV staff and is auto-generated from a syndicated feed.)
