Nvidia's New Playbook: Bond-Fueled Investments, Ethernet Dominance, and Grid-Ready AI Factories

Nvidia raises $25B in oversubscribed bond sale, invests $40B in top AI labs, overtakes Arista in Ethernet switching, and builds AI factories. Stock holds steady near €182.

Nvidia is executing a multi-front strategy that extends far beyond chip design. The company has just raised $25 billion through a bond issuance that drew $85 billion in orders — a signal that institutional investors are eager to back its broader ambition. With cash in hand, Nvidia is pouring $30 billion into OpenAI and $10 billion into Anthropic, effectively locking in chip demand from the world’s most influential AI labs for the next decade. This is not a defensive capital raise; it is a calculated play to entrench dependency on its hardware.

The bond sale came shortly after S&P upgraded Nvidia’s credit rating to AA, reflecting confidence in its financial discipline. Yet the company is also returning capital to shareholders, running an $80 billion buyback program and paying a quarterly dividend of $0.25 per share — the ex-dividend date was June 4, 2026. The message is clear: Nvidia can afford to invest aggressively and still reward investors.

A key milestone in that expansion came in the first quarter of fiscal 2026, when Nvidia’s Spectrum-X platform overtook Arista to claim the top spot in the data-center Ethernet switching market. Segment revenue hit $2.1 billion, surging nearly 193% year over year. By controlling both the GPUs that power AI training and the networking infrastructure that connects them, Nvidia has locked in a powerful two-sided position in the data-center value chain.

Alongside that networking push, Nvidia is tackling a physical bottleneck: energy. The U.S. Federal Energy Regulatory Commission (FERC) recently ordered six major grid operators to fast-track the connection of new AI data centers, with the facilities bearing their own interconnection costs. Nvidia is already capitalizing on this regulatory shift. In partnership with Emerald AI, it is developing so-called AI factories that generate their own power and respond in real time to grid fluctuations. The first commercial project, using Nvidia’s Vera Rubin DSX design, is set to launch in Virginia later this year, delivering 96 megawatts of capacity.

That system is built on Nvidia’s new Rubin platform, which has already entered mass production. The first server deployments from partners are expected in the second half of 2026, with Amazon, Google, and Microsoft among the early customers. Nvidia sees total cloud-capital spending reaching $1 trillion by 2027, and its own revenue growth of 85% in the latest quarter confirms that demand shows no sign of slowing.

Shareholders will have their next chance to weigh in at the company’s virtual annual meeting next Wednesday, where seven items are on the agenda — including board elections and activist proposals related to greenhouse-gas reporting and diversity policies.

On the charts, the stock is holding steady at around €182.00, having defended the 50-day moving average. That is roughly 10% below the all-time high of €202.50 set in mid-May. The RSI sits at 50.5, a neutral reading, while the 200-day line at €163.00 provides a comfortable floor far below. Since the start of the year, shares have gained nearly 13%. Analysts see substantial room to run, with a consensus target of €260.70 — implying upside of more than 40%.

The broader energy backdrop supports this expansion. Data centers currently consume about 5% of U.S. electricity, a share that could triple by 2035. Nvidia’s evolution from a pure chip supplier into an architect of flexible, on-site power networks is not incidental — it is becoming a fundamental requirement for continued growth. The company’s next big bet, “Physical AI” — bringing intelligence into humanoid robots and industrial automation — demands even more embedded infrastructure, and Nvidia is building the tools to deliver it. The question is not whether the technology works, but how fast the world can adapt to the infrastructure Nvidia is now constructing.