Former Intel CEO Abandons OpenAI for DeepSeek's Open Source Model
Pat Gelsinger's startup Gloo is switching to DeepSeek's R1 model, trained at a fraction of typical costs using 2,000 Nvidia H800 GPUs. The move signals a shift in AI landscape as open-source alternatives challenge dominant closed-source models.
Nvidia's H20 GPUs Surge in Chinese Market Despite Export Controls
Nvidia's export-compliant H20 GPUs are seeing explosive 50% quarterly growth in China, outpacing their flagship H100 processors. The success demonstrates Nvidia's ability to maintain market dominance while navigating complex international trade restrictions.
NVIDIA Blackwell GPUs Face Extended Testing Times Due to Complex Architecture
NVIDIA's latest Blackwell GPUs require 3-4 times longer testing periods compared to previous Hopper generation, due to sophisticated dual compute chiplets and HBM3E memory design. The extended validation process reflects increasing complexity in modern AI accelerators and data center processors.
Microsoft's Massive AI Push: 485,000 Nvidia Hopper Chips Purchased in 2024
Microsoft intensifies its AI ambitions by acquiring nearly half a million Nvidia Hopper chips, outpacing competitors threefold. This unprecedented investment, coupled with custom chip development and strategic power agreements, demonstrates Microsoft's commitment to dominating AI infrastructure.
Nvidia Shatters Records with 94% Revenue Surge Fueled by AI Chip Demand
Nvidia reports extraordinary Q4 2023 results with $22.1 billion revenue, driven by soaring demand for AI processors. The company's transformation from gaming hardware maker to AI powerhouse is highlighted by a 409% increase in data center revenue, reaching $18.4 billion.
Intel's Gaudi 3: A Cost-Effective Challenger to Nvidia's AI Dominance
Intel unveils the Gaudi 3 AI accelerator, offering competitive performance at a fraction of the cost of Nvidia's H100. With impressive specs and aggressive pricing, Intel aims to disrupt the AI hardware market.