Nvidia's bumpy November (www.businessinsider.com)

November knocked Nvidia off its recent pedestal: after hitting all‑time highs and delivering another blockbuster quarter, the AI chipmaker saw its stock slide roughly 11% amid renewed “AI bubble” worries, big investors cashing out (SoftBank sold $5.8B), vocal short‑sellers and Michael Burry’s public criticisms. The rout accelerated when reports surfaced that Google is negotiating to supply Meta with its own advanced chips — a sign that Nvidia’s dominance in accelerators used to train and run large language models (LLMs) and chatbots could face real competition. Nvidia nevertheless beat Q3 expectations, doubled down on claims that its Blackwell generation remains “a generation ahead,” and projected massive demand (CFO: roughly $500B of AI chip orders across 2025–26). For the AI/ML community this matters because Nvidia has been the de‑facto supplier for high‑performance model training and inference; any erosion of its share would reshape procurement, costs and performance baselines across cloud providers and labs. If hyperscalers successfully scale alternative accelerators, ecosystems (software stacks, optimizers, compiler toolchains) and ML workflows will need to adapt. The episode also underscores market fragility: stellar technical execution (Blackwell adoption) can coexist with valuation risks and strategic pressure from rivals, making Nvidia both central to and vulnerable within the evolving AI infrastructure landscape.