Nvidia (NVDA) stock has gained about 53% over the past year, at the time of writing, Tuesday afternoon, March 17, according to Yahoo Finance. Meanwhile, the SPDR S&P 500 index (SPY) is up about 18% in the same period.
While the stock has performed well in the long term, today’s performance is a little bit disappointing. The stock is slightly down (0.04%) at the moment. You may think it is not a big deal, but Nvidia’s GPU Technology Conference (GTC) is ongoing.
The company announced many new things, including the forecast that the revenue opportunity for its AI chips may reach $1 trillion or more through 2027, as reported by Reuters.
The likeliest explanation for this unimpressed market reaction is that revenue opportunity heavily depends on hyperscalers. We have to wait for hyperscalers to announce their capital expenditure plans for 2027 before we can conclude that the revenue opportunity is real.
The problem here is that most hyperscalers have already had to raise debt to fund their 2026 spending plans, which is why the sentiment is that the same spending pace will not continue in 2027.
Shutterstock
Key Nvidia GTC announcements
The most important and somewhat shocking announcement from GTC is the Nvidia Groq 3 LPX accelerator. The shocking part is that Nvidia licensed Groq technology in December 2025 and is already launching new chips.
I’ve analyzed this deal in my article “What’s next for Nvidia stock in 2026.” I correctly predicted that these chips, specialized for inference, would come in their own separate racks.
What I didn’t expect was to see them in 2026. No one can design brand-new chips in just three months, so the only conclusion I can make from this is that the Groq team was already nearing completion of the chip design, and Nvidia contributed to the rack design and networking.
I also wrote that Nvidia’s strategy is shifting more towards open-source models, as they fit much better with the AI factories business model.
Nvidia unveiled its Open Agent Development Platform — Agent Toolkit, which provides open source models and software for enterprises and developers. The toolkit includes Nvidia OpenShell, an open source runtime that increases safety by enforcing policy-based security, network, and privacy guardrails.
“Claude Code and OpenClaw have sparked the agent inflection point — extending AI beyond generation and reasoning into action,” stated Jensen Huang, founder and CEO of Nvidia. “Employees will be supercharged by teams of frontier, specialized, and custom-built agents they deploy and manage. The enterprise software industry will evolve into specialized agentic platforms, and the IT industry is on the brink of its next great expansion.”
The company also announced the NVIDIA Nemotron Coalition, a global collaboration between open model builders and AI developers. The coalition members include Black Forest Labs, Cursor, LangChain, Mistral AI, Perplexity, Reflection AI, Sarvam, and Thinking Machines Lab.
More Nvidia:
- Nvidia stock gets major reality check on ‘$100B’ number
- Nvidia CEO delivers blunt 7-word rebuttal on software stocks
- Bank of America resets Nvidia price target after earnings
The first project resulting from the team up will be a base model codeveloped by Mistral AI and Nvidia. Other coalition members will contribute data and expertise to support the model’s post-training and continued development.
The goal for this model is to enable developers and organizations to post-train and specialize AI systems for their industries, regions, and unique needs. The company stated that the model will be trained on Nvidia DGX Cloud, and will be shared with the open ecosystem and support the upcoming Nvidia Nemotron 4 family of models.
Bank of America reiterates Nvidia as its top AI pick
Bank of America analyst Vivek Arya and his team updated their opinion on Nvidia stock following the GTC and post-keynote meeting with the CFO.
Analysts said that Nvidia updated its data center sales outlook from $0.5 trillion or more between 2025 and 2026 to $1.0 trillion or more between 2025 and 2027. This number spans GPU system sales and thus includes relevant CPU and networking sales.
The sales outlook does not include any potential standalone Vera CPUs and LPX solutions.
Arya believes this new $1.0 trillion or more outlook validates the prior Wall Street expectations for Data Center spending of approximately $970 billion for the three years. He noted that Nvidia expects 60% of spending to come from the top five hyperscalers, and 40% from enterprise, industrial, sovereign, etc.
Related: Morgan Stanley resets Broadcom price target after earnings
Analysts said that the cost of building 1 GW of data center power now amounts to approximately $40 billion in capital expenditures. Nvidia’s revenue share of that $40 billion is in the range of $20 billion to $30 billion, depending on the networking content.
In a research note shared with me, Arya reiterated a buy rating for Nvidia stock and the target price of $300, based on 28 multiple of his estimate for price to earnings ratio excluding cash for calendar year 2027, which is within Nvidia’s historical forward year price to earnings range of 25 to 56.
Analysts noted downside risk factors for Nvidia:
- Weakness in consumer driven gaming market,
- Competition with major public firms,
- Larger than expected impact from restrictions on compute shipments to China
- Lumpy and unpredictable sales in new enterprise, data center, and autos
markets, - Potential for decelerating capital returns,
- Enhanced government scrutiny of Nvidia’s dominant market position in AI
chips.
