Latest AI (Artificial Intelligence) News
Stanford Releases 2026 AI Index Report
The Stanford AI Index Report for 2026 highlights generative AI adoption at 53% population-level, spreading faster than the internet or PCs. US and China have closed the performance gap, with US leading private investment at $285.9B in 2025 while China tops publications and patents. The report underscores a fractured global AI landscape.
NVIDIA Blackwell GPUs Sold Out Through Mid-2026
NVIDIA's Blackwell systems, essential for large AI models with 4-bit floating point inference, are sold out through mid-year at ~$40,000 per GPU. Microsoft and Meta deploy massive clusters, solidifying NVIDIA's compute dominance. Upcoming Vera Rubin promises 10x performance per watt.
Taiwan's TSMC Critical to Global AI Hardware
Stanford's AI Index emphasizes TSMC's monopoly on cutting-edge AI chips, making Taiwan vital yet vulnerable in the supply chain. US leads with 5,427 data centers, but all advanced chips rely on Taiwan. Geopolitical risks prompt US-Japan investments in overseas fabs.
Meta Plans 10% Workforce Cut for AI Focus
Meta to lay off ~8,000 employees (10% of workforce) starting May 2026, amid Zuckerberg's hundreds-of-billions AI investment. Cuts aim for leaner structure with AI-assisted work; further reductions possible based on AI evolution. Similar moves at Amazon and Block.
Anthropic Launches Claude Opus 4.7
Anthropic released Claude Opus 4.7, enhancing capabilities in collaboration with Google's TPU access up to one million chips for gigawatt-scale compute by 2026. This follows tight US-China model performance leads per Stanford Index. Google's quantum Willow chip adds to hardware pushes.
Qwen Releases Open-Source 3.6-Plus and 35B Models
Alibaba's Qwen unveiled Qwen3.6-Plus and efficient Qwen3.6-35B-A3B mixture-of-experts model, activating only 3B of 35B parameters yet rivaling larger models in coding. This boosts open-source AI accessibility amid global competition. Stanford notes China's publication lead.