Nvidia today commands visibility into nearly $500 billion of combined Blackwell and Rubin platform revenue from the start of 2025 through the end of calendar 2026, of which roughly $150 billion has already been shipped to hyperscale and enterprise customers. Its networking business—anchored by proprietary NVLink interconnects, InfiniBand fabrics and the AI-optimized Ethernet platform Spectrum-X—now generates multibillion-dollar annual sales, reinforcing Nvidia’s position as an end-to-end AI infrastructure provider rather than solely a GPU designer.

Nvidia remains on track to commence volume production of its Vera Rubin architecture in the second half of 2026. By combining the newly introduced Vera CPU with the advanced Rubin GPU, the platform targets a range of AI workloads spanning cloud services, enterprise inference, robotics and emerging ‘physical AI’ markets. Early third-party benchmarks suggest double-digit performance gains and improved power efficiency compared with current data-center configurations, setting the stage for broader enterprise adoption as demand for turnkey AI solutions accelerates.

In a bid to deepen its AI software and algorithmic prowess, Nvidia is in advanced discussions to acquire Israeli LLM specialist AI21 Labs at a valuation between $2 billion and $3 billion—up from its last private valuation of $1.4 billion—and gain access to a 200-strong research team. Concurrently, the company is planning a new R&D campus in Israel capable of housing up to 10,000 employees, and has secured regulatory approval for a $5 billion strategic share purchase in Intel. These moves reflect Nvidia’s broader strategy to build a vertically integrated AI ecosystem that spans chip design, advanced packaging, software frameworks and global infrastructure.