TL;DR

CES 2026 wrapped with NVIDIA’s next-gen Vera Rubin platform stealing the show. OpenAI launched a healthcare app built with 260 physicians. Anthropic closed a massive $10B round at $350B valuation. And the federal government is gearing up to fight state AI laws in court.

Top Story: NVIDIA Unveils Vera Rubin Platform

NVIDIA officially launched Vera Rubin, the successor to Blackwell, featuring six new chips including the Vera CPU and Rubin GPU. The numbers are striking:

  • 10x reduction in inference token costs
  • 4x fewer GPUs needed for training Mixture-of-Experts models
  • Major cloud providers committed to deploying Rubin-based “superfactories” in H2 2026

The company also unveiled Alpamayo, an open AI model for autonomous vehicle reasoning, and Nemotron Speech ASR for real-time voice recognition. Physical AI—robots and embodied systems—dominated the CES floor, officially dethroning “agentic AI” as the buzzword of choice.

OpenAI Goes Healthcare

OpenAI debuted a dedicated ChatGPT Health experience, developed in collaboration with over 260 physicians. The app integrates with medical records and wellness apps to help users interpret lab results and prepare for appointments.

Key details:

  • Purpose-built encryption for health data
  • Conversations explicitly not used for model training
  • Designed as a “digital advice nurse” for health guidance

Separately, OpenAI created a new Head of Preparedness role focused on frontier AI risk—covering threat modeling, cyber risks, biological misuse, and guardrails for self-improving systems. Sam Altman acknowledged “serious risks from rapid progress.”

Funding & Deals

Anthropic secured a $10 billion funding round at a $350 billion valuation, cementing its position as the leading OpenAI competitor. The company also shipped Claude Code 2.1.0 on January 7.

Mobileye agreed to acquire humanoid robotics startup Mentee Robotics for $900 million, marking a bold expansion from automotive autonomy into physical AI. Expected to close Q1 2026.

OpenAI is exploring a Pinterest acquisition to access the platform’s ~200 billion images and intent-labeled visual data—a move aimed at strengthening ChatGPT’s understanding of visual planning and competing with Google’s search advantages.

Consumer AI Updates

Product What’s New
Gmail Gemini 3 integration with “Help Me Write” and conversational inbox search
Samsung Galaxy AI Expanding from 400M to 800M AI-enabled devices in 2026
Lenovo Qira New AI assistant spanning PCs and Motorola phones
Ford AI voice assistant coming 2026 (app) and 2027 (in-vehicle)

Policy Watch: Federal vs. State Showdown

The Trump administration is moving to preempt state AI laws. Key developments:

  • December 11: Executive order establishing a “national policy framework” for AI
  • This week: DOJ created an AI Litigation Taskforce to challenge state regulations as unconstitutional
  • January 1: California, Texas, and Illinois AI laws went into effect despite federal pushback
  • Colorado postponed its AI Act to June 30, 2026

This sets up a major federalism battle over who regulates AI in 2026.

Robotics Roundup

CES 2026 was the coming-out party for humanoid robots:

  • Boston Dynamics + Google DeepMind announced a partnership to equip humanoids with advanced AI capabilities
  • EngineAI showcased the PM01 lightweight agent and T800 full-scale humanoid
  • Hexagon + Microsoft partnered on robots for data-driven manufacturing
  • LG CLOiD smart home robot launched, powered by NVIDIA Jetson Thor

Research Corner

Falcon-H1R 7B from UAE’s Technology Innovation Institute is turning heads—a Transformer-Mamba hybrid that matches models 7x its size on reasoning benchmarks while running efficiently on edge hardware.

MIT and Microsoft introduced an AI system that designs peptide-based molecular sensors for early cancer detection, demonstrating end-to-end inverse design from functional requirements to candidate molecules.

The broader trend: 2026 is shifting from “bigger models” to smaller, specialized systems optimized for specific domains.

Quick Hits

  • Meta signed nuclear power deals with TerraPower, Oklo, and Vistra for AI data centers
  • Zhipu AI became the first LLM company to IPO (~$6.8B market cap)
  • Chinese open-source models gaining traction in Western apps as the frontier gap shrinks
  • Gartner: 80% of organizations expected to formalize AI governance policies this year
  • Industry analysts predict 40% of agentic AI projects will be canceled by 2027 due to unclear ROI

Controversy Corner

  • Meta leaked documents show executives approved “sensual” AI conversations with minors
  • Psychiatrists warning of rising “AI psychosis” cases
  • OpenAI lawsuit pending over teen suicide allegedly linked to chatbot interaction
  • Baltimore AI security system mistook a Doritos bag for a gun

What to Watch

This week: More CES fallout and product deep-dives

This month: State vs. federal AI regulation legal challenges begin

This quarter: Anthropic’s plans post-funding, OpenAI-Pinterest acquisition talks

This newsletter was researched and compiled using multi-source parallel queries across Perplexity, Claude, and Gemini.