AINewsnow

Nvidia Shakes Up Leadership as Google Floods the Zone With Gemini

2026-11-13

Nvidia Shakes Up Leadership as Google Floods the Zone With Gemini

Nvidia's CEO exit dominated the day, but Google's multi-front Gemini push — new models, smart glasses, and genomics AI — kept the pressure on rivals. Meanwhile, Anthropic turned Claude into a workspace and researchers showed how easily AI agents can be sabotaged.

  1. Jensen Huang steps down as Nvidia CEO amid scandal Nvidia announced on November 12, 2026 that Jensen Huang is leaving as CEO following a scandal, with CFO Colette Kress appointed interim CEO. Industry figures are weighing the impact on Nvidia's chips, investments, and rivals.
    Why it matters: The leader of the world's most valuable AI chipmaker stepping down abruptly creates uncertainty across the entire AI supply chain.
  2. Google launches Gemini 4 Hydrogen, a lower-cost model for coding and agents Google released Gemini 4 Hydrogen on Nov 12, priced at $1.50/$7.50 per million tokens with a 1M-token context, four thinking levels, and 2.5x faster task completion than Gemini 3.8 Flash. It is rolling out to Pro and Ultra users and via API.
    Why it matters: A cheaper, faster Gemini tier intensifies price competition in the coding and agentic AI market where margins are being fought over.
  3. Google opens pre-orders for Gemini smart glasses with Warby Parker and Samsung Google opened pre-orders on November 12 for Gemini-powered Intelligent Eyewear starting at $395, designed with Gentle Monster and Warby Parker and engineered with Samsung, with Samsung unveiling 30 models.
    Why it matters: A broad, fashion-forward hardware lineup signals Google's serious bid to make AI wearables mainstream rather than a niche experiment.
  4. DeepMind unveils AlphaGenome 2 for cell-type-specific DNA predictions Google DeepMind announced AlphaGenome 2 on November 11, a deep-learning model predicting how DNA variants affect gene regulation in specific cell types and states, doubling causal variant recall over its predecessor.
    Why it matters: Better variant prediction could accelerate genetic research and personalized medicine by pinpointing which mutations actually matter.
  5. Gemini 4 Carbon opens to Pro users and API after cybersecurity debut Google's Gemini 4 Carbon, first released to cybersecurity defenders two weeks earlier, became available to Pro plan users and via API in AI Studio on November 10, topping Artificial Analysis' Intelligence Index v5.
    Why it matters: Google's staged rollout strategy — defenders first, then everyone — is becoming a template for how frontier models reach the market.
  6. Gemini formally verifies Signal's encryption protocol Google DeepMind used Gemini to formally verify Signal's protocol, an effort covered by Techmeme, The Register, and Unite.AI.
    Why it matters: Formal verification of a widely trusted messaging protocol shows AI moving from code generation to proving critical software correct.
  7. Anthropic expands Claude into a full workspace Anthropic launched Claude Sheets and Whiteboards in beta, made Claude Data generally available, and began rolling out Claude Meetings in beta for Team and Enterprise plans. It also released Claude for Manufacturing with 17 workflow skills.
    Why it matters: Claude is now competing directly with Google Workspace and Microsoft 365, pushing Anthropic from chatbot vendor into productivity platform territory.
  8. Microsoft shows AI agents can be backdoored by editing fewer than 10 neurons Microsoft Research's 'Lazarus' technique plants a working backdoor in an AI agent by reprogramming fewer than 10 dormant neurons with no retraining; a trigger string activates the hidden behavior, reportedly succeeding 98.8% of the time.
    Why it matters: The ease and reliability of the attack exposes a serious security gap in agent deployments that current safeguards may not catch.
  9. ElevenLabs ships Scribe v3 as Fish Audio launches full-duplex Live 3 On November 10, ElevenLabs released Scribe v3 and Scribe v3 Live, claiming 27% fewer word errors than Scribe v2 across 102 languages and 43% fewer semantic errors in the realtime variant. Fish Audio also shipped Live 3, a full-duplex voice model.
    Why it matters: Rapid gains in speech accuracy and real-time voice interaction are bringing natural, low-latency voice AI closer to production quality.
  10. AWS launches emulator of 200+ services written entirely by AI agents AWS released AWS Emulator, a free CC0-licensed tool simulating 200+ services from 438 API models, saying autonomous AI agents wrote and tested all code without human review. It passes 56.6% of 42,857 published fidelity tests.
    Why it matters: Shipping unreviewed AI-generated code at this scale is a landmark test of whether autonomous agents can produce production-grade software.