Nvidia's $300 and $350 Targets Rest on a Bet Beyond Chips: Owning the Rules for Autonomous AI

Morgan Stanley and Cantor Fitzgerald back Nvidia, citing AI data center demand, a $235 billion buyback and a shift into agent safety software.

Two of Wall Street's more vocal semiconductor analysts have planted their flags on Nvidia this week, and neither is particularly shy about the upside. Morgan Stanley's Joseph Moore, fresh from a sit-down with chief executive Jensen Huang, reiterated his preference for the stock as his top pick in the chip sector and pinned a $300 price target on it. Cantor Fitzgerald went further, reaffirming its "Overweight" rating on Thursday with a $350 objective.

The shares were trading at EUR 210.10 pre-market, up 1.1%, leaving them just 0.4% shy of their 52-week high of EUR 211.00. During European hours the stock added 0.9% to reach EUR 209.80.

The Bottleneck Has Moved

Morgan Stanley's case rests on a structural shift that has quietly redrawn the map of the AI buildout. The constraint is no longer wafer fabrication capacity. What sets the pace now is how fast new data centers can be constructed, whether suitable land can be secured, and above all whether the power grid can keep up.

That reordering plays to Nvidia's hand, in the bank's view. A portfolio spanning graphics processors, networking components and software architectures lets the company respond nimbly to whatever operators need, while its global customer base gives it a positional edge that rivals would struggle to replicate.

Amazon's Two Million Chips and the SpaceX Buildout

Underwriting all of this is demand from the largest technology buyers. Amazon Web Services intends to fold an additional two million Nvidia graphics processors into its infrastructure over the coming years, a partnership that also covers the company's Vera central processors and advanced networking technology. SpaceX, meanwhile, is pressing ahead with its own data center expansion.

The supply chain is sending matching signals. Reports from manufacturing partner Foxconn point to unabated appetite for specialized data center servers. And the workload itself is broadening: beyond conventional model training, the execution of autonomous AI agents is taking on greater weight, which translates into durable demand for compute.

From Chip Vendor to Control Layer

Cantor Fitzgerald's thesis looks past the hardware entirely. Selling accelerators alone, in its reading, will not sustain dominance indefinitely. The real test of the valuation is whether Nvidia remains indispensable across the software ecosystem.

About a week ago the company unveiled its Open Agent Safety Platform, which pairs OpenShell runtime software with the Sentry reference design to monitor and govern the activity of autonomous AI agents. More than 100 organizations were on board at launch, though TechCrunch noted that OpenAI was not yet among the listed participants.

That omission does little to blunt the strategic significance. Nvidia is repositioning itself not merely as a supplier of chips but as the safety and control authority for the very agents that Bloomberg identifies as a central driver of future infrastructure demand. Whoever defines the control architecture locks in long-term margins.

Buybacks, Margins and a Macro Tailwind

Capital allocation has turned notably aggressive alongside the operational push. Roughly a week ago the board authorized an additional $150 billion in share repurchases, lifting the remaining authorization to $235 billion, which Nvidia intends to deploy through fiscal 2028.

The scale of that program says two things at once: management considers its own cash generation sturdy enough to return astronomical sums to shareholders, and it is signaling confidence in the intrinsic value of the stock. Morgan Stanley, for its part, expects a gross margin of at least 72% over the medium term.

The macro backdrop has cooperated. Weaker-than-expected US employment data on Friday, as reported by Reuters, took the edge off interest rate worries. When monetary policy risk recedes while demand for enterprise software accelerates, technology heavyweights tend to benefit directly.

What to Watch

Market participants will treat the coming quarterly results from the major cloud operators — Microsoft, Alphabet, Amazon and Meta — as the key reference points. Further catalysts are expected from Nvidia's next GTC developer conference, where Huang plans to unveil new processor generations and software extensions.

The $350 target demands flawless execution of the company's expansion plans. Yet the combination of a software strategy built around safety and control, a $235 billion buyback authorization and persistent infrastructure hunger substantially limits the downside. For investors with a long horizon, the odds still tilt in their favor.