AI News Roundup — March 16, 2026
Nvidia GTC 2026 dominated the day as Jensen Huang delivered a blockbuster keynote announcing $1 trillion in chip orders through 2027, the Groq 3 LPU, the NemoClaw agentic AI platform, and Uber robotaxis in 28 cities by 2028.
Nvidia GTC 2026 Keynote: Jensen Huang Announces $1 Trillion in Orders Through 2027
In his two-hour keynote at the SAP Center in San Jose, Nvidia CEO Jensen Huang revealed that the company expects purchase orders between Blackwell and Vera Rubin chips to reach $1 trillion through 2027 — doubling last year's $500 billion projection. Huang also previewed the long-term GPU roadmap, highlighting Vera Rubin systems arriving this year and Feynman-class chips later in the decade. The event drew 30,000 attendees from 190 countries.
Nvidia Unveils Groq 3 Language Processing Unit
Huang introduced the Groq 3 LPU, Nvidia's first chip from the startup it acquired for $20 billion in December. The Groq 3 LPX rack holds 256 LPUs and is designed to work alongside Vera Rubin rack-scale systems for ultra-fast AI inference. The chip is expected to ship in the third quarter of 2026.
NemoClaw Open Source Agentic AI Platform Debuts at GTC
Nvidia launched NemoClaw, an open-source stack for developing and deploying autonomous, long-running AI agents that can reason, plan, write code, call tools, and analyze data. Paired with DGX Spark and DGX Station hardware, the platform provides secure enterprise agentic AI with built-in privacy and governance. Nvidia also announced Nemotron 3 Ultra and a Nemotron 4 coalition including Perplexity, Mistral, Cursor, and Black Forest Labs.
Tom's Hardware · Analytics Insight
Uber to Launch Nvidia-Powered Robotaxis in 28 Cities by 2028
Huang announced that Uber will deploy a ride-hail fleet powered by Nvidia's Drive AV software across 28 cities on four continents by 2028, starting with Los Angeles and San Francisco next year. Nissan, BYD, Geely, Isuzu, and Hyundai are also building Level 4 autonomous vehicles on Nvidia's Drive Hyperion platform. The GTC showcased 110 robots demonstrating Nvidia's physical AI vision.
Disney's Walking Talking Olaf Robot Steals the Show at GTC
One of the keynote's standout moments was a walking, talking Olaf robot from Frozen that appeared on stage alongside Jensen Huang. The robot was trained in an Nvidia simulation co-developed with Disney and was part of a showcase of 110 robots at GTC demonstrating advances in physical AI and embodied intelligence.
Claude Gets 1 Million Token Context Window at Standard Pricing
Anthropic expanded Claude Opus 4.6 and Sonnet 4.6 to include the full 1 million token context window at standard pricing on the Claude Platform, eliminating the previous pricing multiplier. The upgrade means fewer conversation compactions and is also available in Claude Code for Max, Team, and Enterprise users.
Cerebras Coming to AWS with Fastest AI Inference via Bedrock
AWS announced it is deploying Cerebras CS-3 wafer-scale systems to offer the industry's fastest AI inference through AWS Bedrock. The partnership supports open-source LLMs and Amazon's Nova models, bringing Cerebras's unique computing architecture to the world's largest cloud platform.