AI News Roundup: NVIDIA GTC Unveils Rubin, Anthropic Dominates Enterprise, Tesla’s Terafab Takes Shape

NVIDIA GTC 2026 Kicks Off With Vera Rubin and a Vision for AI’s Next Era

NVIDIA CEO Jensen Huang takes the stage today in San Jose for his two-hour GTC 2026 keynote, officially launching the Vera Rubin platform — the successor to Blackwell and the company’s most ambitious compute architecture yet. The platform encompasses six new chips: the Rubin GPU with up to 288GB of HBM4 memory, the Vera CPU, a next-generation DPU, advanced NICs, NVLink 6 scale-up networking, and new Ethernet switching infrastructure.

The numbers are staggering: Rubin promises up to a 10x reduction in inference token cost and a 4x reduction in the number of GPUs needed to train mixture-of-experts models compared to Blackwell. AWS, Google Cloud, Microsoft Azure, and Oracle Cloud are among the first providers set to deploy Vera Rubin–based instances in the second half of 2026. The four-day conference, running through March 19, will feature sessions on physical AI, robotics, agentic AI, and AI factories.

Anthropic Surges Past OpenAI in Enterprise Adoption

The Ramp AI Index for March 2026 reveals a dramatic shift in the enterprise AI market: Anthropic now wins approximately 70% of head-to-head matchups against OpenAI among businesses purchasing AI services for the first time — a complete reversal from 2025 trends. Overall business AI adoption hit a record 47.6%, with nearly one in four businesses on Ramp now paying for Anthropic, up from one in 25 a year ago.

OpenAI posted its largest single-month decline in adoption since tracking began, while Anthropic’s 4.9% month-over-month growth was its biggest gain ever. The shift comes despite Anthropic charging more for roughly equivalent performance — and despite the company actively turning away revenue because it lacks the compute to serve demand. The data suggests enterprise buyers are increasingly prioritizing reliability and safety guarantees over price.

Tesla’s $20B Terafab AI Chip Factory Nears Launch

Elon Musk announced on March 14 that Tesla’s Terafab semiconductor project launches on March 21. The facility aims to combine logic processing, memory storage, and advanced packaging under one roof — vertically integrated chip manufacturing at a scale no private company outside Taiwan and South Korea currently operates.

Terafab is targeting 2-nanometer process technology, the most advanced node in commercial production. Tesla’s fifth-generation AI chip, AI5, is among the first products the facility will produce, with small-batch production expected in 2026 and volume production projected for 2027. The chips are destined for Full Self-Driving, the Cybercab robotaxi program, the Optimus humanoid robot line, and xAI’s Grok training infrastructure.

Meta and Nebius Ink $27B AI Infrastructure Deal

Nebius signed a five-year, $27 billion AI infrastructure agreement with Meta on March 16, marking one of the first large-scale deployments of NVIDIA’s Vera Rubin platform. Nebius will deliver $12 billion of dedicated GPU capacity across multiple locations starting in early 2027, deepening its position as a major AI cloud provider.

The deal underscores the insatiable demand for AI compute capacity. Meta has been on a spending spree, having already committed to a multiyear partnership with NVIDIA for hyperscale data centers deploying millions of Blackwell and Rubin GPUs. Combined with Microsoft and Amazon’s plans, Big Tech is on track to invest more than $650 billion in AI infrastructure in 2026 alone.

Eli Lilly Fires Up Pharma’s Most Powerful AI Supercomputer

Eli Lilly inaugurated LillyPod, the world’s first NVIDIA DGX SuperPOD built with DGX B300 systems, delivering more than 9,000 petaflops of AI performance. Assembled in just four months and powered by 1,016 Blackwell Ultra GPUs, the Indianapolis-based system is now in production use across genomics, molecule design, single-cell biology, imaging, and manufacturing optimization.

Lilly aims to use the system to cut the typical 10-year drug development timeline in half by accelerating every stage from target discovery to clinical trials. The company has committed to running its AI supercomputing infrastructure on 100% renewable electricity by 2030, using efficient liquid cooling to minimize energy impact.

Anthropic Launches Research Institute to Tackle AI’s Societal Impact

Anthropic introduced The Anthropic Institute on March 11, a new research arm led by co-founder Jack Clark in a new role as Head of Public Benefit. The institute brings together machine learning engineers, economists, and social scientists from three existing teams: Frontier Red Team, Societal Impacts, and Economic Research.

The roughly 30-member institute includes notable hires like Matt Botvinick (formerly Google DeepMind), Anton Korinek (University of Virginia economics), and Zoe Hitzig (formerly OpenAI). The launch coincides with Anthropic’s ongoing legal battle with the Pentagon, positioning the company as a voice for responsible AI development at a time when the industry faces mounting scrutiny over military and surveillance applications.