NEW YORK, March 16, 2026, 09:14 (EDT)
- Nvidia gained 1.1% in premarket hours as of 7:12 a.m. ET, trading ahead of GTC. The stock ended Friday at $180.25.
- Nvidia shares slipped after blockbuster results in late February, putting the focus squarely on what’s coming next for the chip giant’s roadmap and inference strategy. Investors want more clarity.
- Investors are watching the usual suspects: ramped-up custom-chip rivalry, crude flirting with $100, and a Fed widely expected to hold rates steady this week.
Nvidia climbed 1.1% in early U.S. premarket action Monday, trading higher at 7:12 a.m. ET as investors looked ahead to CEO Jensen Huang’s keynote at the GTC conference. Shares had closed out Friday at $180.25, still searching for momentum after the late-February earnings rally faded.
The timing isn’t random. Nvidia delivered a blowout fourth-quarter revenue figure in February, yet shares slid 4% the following day. Investors weren’t convinced that the chip giant’s aggressive outlays across the AI sector would ultimately pay off for shareholders.
This week, investors aren’t settling for rosy talk. They’re pressing for specifics on inference—the moment an AI model actually responds to a prompt—and on agentic AI, those autonomous software agents that handle tasks for users. Both are areas where competitors and big clients are picking up speed.
Huang’s set to speak at 2 p.m. ET in San Jose. Nvidia previewed a keynote focused on chips, “AI factories”—the massive systems behind model training and deployment—plus open models, agentic systems, and physical AI. An investor and analyst Q&A is scheduled for Tuesday. NVIDIA Newsroom
Matt Britzman at Hargreaves Lansdown thinks if Huang can prove Nvidia’s hardware is ready not only to build AI but to run it in daily life, it could shore up belief in what he calls the “next leg of the AI race.” BNP Paribas’s David O’Connor points to possible reveals like the Feynman architecture, a new inference chip, and a stronger play in optical networking, but he doesn’t see GTC sparking a major move in the stock right now. Reuters
That’s no coincidence. Huang is set to unveil the next-generation Feynman chip, plus updates for data centers, CUDA, Nvidia’s chip-programming software, digital assistants, and robotics. Groq-related inference tech is also likely on the agenda, as the industry’s AI appetite moves away from training massive models and toward real-time user services.
Nvidia’s blowout quarter brought in $68.1 billion in revenue, $62.3 billion of that from data-center sales alone. CEO Jensen Huang called it: “the agentic AI inflection point has arrived.” Still, investors weren’t satisfied—shares dropped right after the results, a sign that expectations are sky-high. NVIDIA Investor Relations
The competitive noise is intensifying. OpenAI has been looking into inference options from AMD, Cerebras, and Groq; Meta, for its part, is speeding up internal AI chip development. Still, Nvidia claims over 90% of training and inference right now, according to Summit Insights’ KinNgai Chan. He thinks that grip starts to loosen in 2027, as more customers ramp up their own chip projects.
Evidence of new investment flowing into the Nvidia orbit keeps surfacing. On Monday, Nebius said Meta has committed to purchase $12 billion of AI computing capacity by 2027, and potentially another $15 billion after that. Just last week, Nvidia revealed it had taken a $2 billion stake in Nebius, which relies on Nvidia chips to power its data centers.
The picture remains messy. Broadcom is pushing its custom AI chips further into a space where Nvidia has held sway for years, while oil prices are sticking close to $100 a barrel amid Middle East shipping disruptions. Traders are also betting the Fed keeps rates elevated for an extended stretch. Given all that, spotty product timelines or scant details on inference returns might limit any GTC rally.
Nvidia still tops the charts as the world’s most valuable publicly traded company, sporting a roughly $4.53 trillion market cap. Monday brings the keynote, then analysts get their Q&A with Huang on Tuesday—a brief shot for him to convince Wall Street that AI’s momentum hasn’t run out of steam for the stock.
