Timeline
Nvidia launches Cosmos 3 at CVPR
Launched physical AI agent skills, Alpamayo 2 Super, Cosmos 3, and other tools at CVPR
Rated Jensen Huang's ComputeX keynote 'F Tier' for no AI datacenter news
Released Nemotron 3 Ultra, a 550B open-weight model
Announced new Windows SoC for AI PCs targeting on-device inference
NVIDIA ARM laptop chip delayed 6-8 months due to video output issues
Blackwell confidential computing disables NVLink multicast, causing 61% regression on SGLang Qwen3.5 397B
Published analysis of 432k coding agent requests showing median 96k input tokens
SemiAnalysis conducted internal bake-off of AI tools, finding Perplexity Slack bot beats Claude
Flagged 8x SRAM understatement on Cerebras website, a transparency issue for IPO due diligence.
Ecosystem
SemiAnalysis
Nvidia
Evidence (7 articles)
SemiAnalysis: NVIDIA's Customer Data Drives Disaggregated Inference, LPU Surpasses GPU
Apr 22, 2026CPU Demand Flipping the AI Narrative as Datacenter Growth Shifts
Apr 28, 2026NVIDIA Vera Rubin VR NVL72: Value Extraction Engine Arrives
May 2, 2026SemiAnalysis Calls Jensen ComputeX Keynote 'F Tier' Over No AI DC News
Jun 1, 2026Blackwell NVLink Breaks Confidential Compute, 61% Regression Reported
May 30, 2026xAI Drops JAX, Builds Custom C Training Framework After <10% MFU
May 30, 2026Cerebras IPO Challenges GPU Scaling Orthodoxy
May 14, 2026