Air Cargo Is the New Leading Indicator for AI Capex
AI goods are 7% of air cargo volume but 53.5% of its value. Air freight is now a real-time gauge of the AI buildout.
AI News
Latest AI news, breakthroughs, and industry trends.
China's Indium Grip: The Hidden Bottleneck Behind CPO
China holds 70% of proven indium reserves and tightened export caps, choking the InP wafer supply behind Nvidia CPO optics.
AI Daily Briefing – 2026-08-20: Anthropic ARR Tops $65B
Welcome to the AI Daily Briefing for Thursday, August 20. Today the money story dominates: Anthropic is closing in on an IPO with annualized revenue past $65 billion, AI chip startup Etched has doubled its valuation to $21 billion in a month, and Cerebras is claiming 30x the inference throughput of GPU servers with a new rack-scale system. Yet the bill for the AI buildout is starting to come due —
WAM to WTM: Why Task Completion Is Robotics AI's Next KPI
Task completion is replacing per-step accuracy as embodied AI's core KPI. Why WAM hits a wall and how WTM changes the game.
GPT-5.6 Sol: 3x Vision Jump, Coding SOTA at Lower Cost
GPT-5.6 flagship Sol nearly triples detection accuracy and tops coding benchmarks; Terra and Luna match rivals at a fraction of the cost.
Seedance 2.5: Stronger AI Video, Higher Production Bar
Seedance 2.5 improves AI video, but studios face higher costs and a consistency wall. The real shift is to unit economics.
Research Preference Models: Meta Prioritizes AI Experiments
Meta's Research Preference Models rank which experiments deserve compute before execution, cutting the GPU budget by a third.
Banks Lend Against AI Tokens: Compute Becomes Collateral
Chinese banks now lend against AI token usage, treating usage data as collateral for light-asset AI startups.
Claude's Autonomous Protein Design Hits 14 of 15 Targets
Claude designed protein binders for 14 of 15 targets, beating the 10-15% industry baseline with 22-35% hit rates.
Why Arm CPUs Are Now Core to AI Data Centers
Arm CPUs are quietly becoming the backbone of AI data centers. Agent workloads make CPU as important as GPU, and hyperscalers are betting on it.
BoltzMol-1: Virtual Screening That Finds Real Hits
BoltzMol-1 turns Boltz-2 into a prospective screening pipeline: real hits on 6 of 10 targets, 28-96 compounds each.
OpenAI Pauses RL Training: Cyber-Safety Gates Move Earlier
OpenAI paused frontier RL training as Astra neared critical cyber capability, moving safety gates into the training phase.
