Nvidia's Telecom Push and Buyback Muscle Keep the Stock Within Striking Distance of Its Peak

Nvidia unveiled its Nemotron 3 Large Telco Model and a NeMo recipe for carriers, as shares ended at EUR 212.50, 1.7% below the 52-week high.

Nvidia is widening its reach beyond the hyperscale data centers that built its fortune, rolling out industry-specific artificial intelligence tools aimed at locking telecommunications operators into its software stack. The chipmaker unveiled its Nemotron 3 Large Telco Model on Tuesday, fine-tuned by AdaptKey, alongside a NeMo recipe that lets carriers tailor open models to their own operational data. Company research cited in the announcement found that 89 percent of industry respondents consider open models and associated software a key pillar of their AI strategy.

The move marks a deliberate shift toward monetization that extends past raw accelerator sales. By embedding network operators and IT service providers inside its NeMo toolset and specialized model ecosystem, Nvidia raises the cost of switching to rival platforms—a recurring-revenue stream that could prove stickier than one-off hardware orders.

A Stock Trading Just Below Its High-Water Mark

Shares ended Tuesday's session at EUR 212.50, leaving the stock just 1.7 percent shy of the 52-week high of EUR 216.10 touched on Monday. The advance came as part of a broad recovery across US equity markets, with the S&P 500 and Nasdaq climbing to record levels. According to Reuters, falling US Treasury yields and softening oil prices underpinned the buying mood, with market-wide momentum—rather than any single company announcement—doing the heavy lifting. Nvidia itself added a modest 0.2 percent on the day, changing hands at EUR 213.30 at one point.

That tide has carried the stock up 33 percent since the start of the year on European trading venues.

Buybacks and Safety Software Round Out the Story

Supporting the equity is an aggressive capital-return posture. On September 28, the board authorized an additional USD 150 billion share repurchase program. Combined with prior approvals, Nvidia now has USD 235 billion available for buybacks, which it intends to deploy through fiscal 2028. Repurchases of this scale shrink the share count and bolster earnings per share.

The company has also been building out its autonomous-software franchise. Late last month it introduced Open Agent Safety, a platform pairing OpenShell software with the Sentry reference design to deliver governance and oversight for AI agents—standardized security checks that stretch from early testing through production deployment.

Foxconn Data Underscores Hardware Demand

On the manufacturing side, key partners are signaling no letup in AI infrastructure activity. Server assembly partner Hon Hai Precision Industry, better known as Foxconn, posted September-quarter revenue of TWD 3.03 trillion, a 47 percent jump from the year-earlier period. Media coverage framed the result as evidence that appetite for AI infrastructure remains robust.

Further validation came from CoreWeave, which late last month made Nvidia Vera Rubin NVL72 systems available on its platform. Startup Cognition simultaneously went live as the first customer running production workloads on the new architecture. Whether such deployments scale smoothly will shape margins in the quarters ahead.

Analyst Upgrades and a Grid-Side Caveat

Sentiment in the analyst community remains constructive. Karl Ackerman of Exane BNP Paribas raised his price target on the stock Tuesday to USD 345 from USD 285, keeping a "Buy" rating.

Morgan Stanley weighed in Monday with a more nuanced view, arguing that Nvidia and Broadcom are relatively well insulated from power-supply bottlenecks afflicting US data centers. While delays in data center construction could hit secondary chip suppliers, demand for core processors stays protected for now. Still, the analysts cautioned that slower delivery of data center capacity could trigger order deferrals or cancellations among memory, optics, and secondary chipmakers. If physical buildouts stall, the pull for primary compute units could eventually lose steam as well.

A second risk lies with customers themselves: justifying enormous compute investments. Should agentic AI systems and telecom-specific models underdeliver in practice, network operators may plan specialized compute budgets more cautiously.

Chart Levels and the Road Ahead

Technically, the uptrend holds as long as the stock stays above recent breakout zones and defends the EUR 210 mark. A sustained push past the 52-week high of EUR 216.10 would open the door to further gains. A slide below key support, by contrast, would refocus attention on how quickly new hardware platforms are actually shipping.

The next concrete catalyst on the technology side is organizational: applications for Nvidia's university support program close on October 30, 2026, offering participants a window into the company's research collaborations. On the operational front, the pace of Vera Rubin rollouts at partners like CoreWeave will reveal how fast new customers can bring productive workloads online.