Perplexity Chooses CoreWeave to Scale AI Inference, Stock Rises
CoreWeave announces multi-year deal to power Perplexity's AI inference on NVIDIA GB200 clusters.
CoreWeave announces multi-year deal to power Perplexity's AI inference on NVIDIA GB200 clusters.
The US is considering capping AI chip exports like Nvidia's H200 and AMD's MI325 to just 75,000 units per Chinese company. This news is pressuring stocks amid ongoing trade tensions and global market worries.
NVIDIA commits $4B ($2B each) to Lumentum & Coherent to advance silicon photonics and energy-efficient AI infrastructure. Stocks jump.
Meta abandons its most advanced in-house AI training chip (Olympus) due to design struggles and shifts focus to simpler MTIA versions while securing multi-billion deals to rent Google TPUs and buy chips from AMD & Nvidia.
Nvidia gets US license for limited H200 AI chip exports to China with 25% duty, but zero revenue so far amid Beijing import uncertainty and rising local rivals.
NVIDIA reports record $68.1B Q4 revenue (+73% YoY) and guides $78B for Q1 2027, beating estimates as AI demand surges. Data center hits $62.3B.