In spite of its recent release of the latest generation of AI chips at the annual GPU Technology Conference (GTC), Nvidia’s share price dipped, reflecting the market’s conservative disposition.
The firm introduced its generation newer AI processors, intended to advance industries such as robotics and autonomous vehicles, during the San Jose conference in California. However, Nvidia stock declined over 3% following the announcement instead of sparking investor enthusiasm. All of the Magnificent Seven closed the session in negative territory following a broader sell-off in tech shares precipitated by market worries regarding economic uncertainty fueled by Trump’s tariffs.
Nvidia shares are down 15% this year so far, dragged by the general risk-off mood across world markets and ramped-up Chinese competition from DeepSeek, whose new cost-effective AI model recently hit the streets.
The firm’s recent quarterly earnings report also didn’t sit well with investors, as decelerating sales growth raised eyebrows about its staying power. Nevertheless, some analysts consider the pullback a buying opportunity. “Investors might view that as an opportunity, especially with its valuation remaining appealing on the backdrop of sustained growth,” eToro Australia market analyst Josh Gilbert said.
The GTC is an important event for Nvidia, drawing AI developers and investors globally. Persuading large cloud service providers, or hyperscalers, to keep investing in its next-generation AI chips is still important for the company’s expansion, particularly with DeepSeek’s new competition on the horizon. The Chinese company’s lower-cost generative AI model is regarded as a direct threat to the US tech giants’ dominance.
Perhaps the most significant reveal at the show was Nvidia’s future successor to its Blackwell supercomputing chip, named Blackwell Ultra. Planned to arrive in the second half of 2025, the next-generation upgrade holds a lot of promise in the way of enhanced performance. Current Blackwell chips, which have been shipping in late 2024, are a key money maker for the company.
Blackwell Ultra is engineered to run more tasks in the same amount of time as the previous generation, making it a more productive option for cloud providers. As with Nvidia, this will allow businesses to create 50 times the revenue generated by the last generation of Hopper GPUs. We architected Blackwell Ultra for this moment — it’s one general-purpose platform that can do pretraining, post-training, and reasoning AI inference easily and efficiently,” said CEO Jensen Huang.
The second big announcement was the Vera Rubin system, a next-generation computing platform that combines CPU and GPU functionality. Coming in the latter half of 2026, Vera Rubin takes its name from the astronomer who contributed so much to research on dark matter. It’s being designed as a bespoke supercomputing system that can execute 50 petaflops of work performing inference tasks—over twice what Blackwell chips can currently accomplish.
Nvidia also made a number of high-profile partnerships during the event. Included among them is a collaboration with Walt Disney and Google DeepMind known as Isaac GR00T N1, which is geared towards speeding up robotics progress. The company is also partnering with General Motors to create AI-based automotive technology and collaborating with T-Mobile US and Cisco Systems to build AI-based 6G network infrastructure.
AI investments are set to dump hundreds of billions of dollars into the sector within the next decade, and Nvidia is still at the forefront of this revolution. The firm’s main clients are technology giants such as Amazon, Microsoft, Alphabet, and Oracle, all of which have cumulatively bought 3.6 million Blackwell AI chips in 2025 alone, as per Huang.
A Bloomberg report points out that hyperscalers like Microsoft, Amazon, and Meta Platforms are likely to spend a whopping $371 billion (€340 billion) on AI infrastructure in 2025. That amount is likely to grow to $525 billion (€480 billion) by 2032, reflecting a continued focus on AI-led growth.
Spending in the future on AI will be less on developing entirely new AI models and more on augmenting processing capacity and developing inference capabilities, as per a company. It is moving toward increasing computational efficiency with inspiration drawn from innovations like DeepSeek’s groundbreaking method.
“In the last 2 to 3 years, a major breakthrough happened, a fundamental advance in artificial intelligence happened. We call it agentic AI,” Huang said. “It can reason about how to answer or how to solve a problem.”
Nvidia’s ability to drive AI innovation and maintain its leadership in the industry will be closely watched as it navigates changing market trends and increased competition. The long-term AI investment trend is that Nvidia’s innovative technology will continue to dominate the shaping of the future, although near-term market responses remain uncertain.