If the GenAI Bubble Bursts, Nvidia Will Still Keep Growing (www.nextplatform.com)

Nvidia shrugged off “AI bubble” talk on its fiscal Q3 call while posting a blockbuster quarter: revenue of ~$57B (up 62.5%), net income $31.9B, and datacenter sales of $51.2B. CEO Jensen Huang argued Nvidia isn’t riding a fad but is differentiated across the AI stack—pretraining, post‑training and inference—backed by two decades of CUDA‑X libraries. The company and analysts also floated huge infrastructure forecasts: Huang’s earlier $3–4T of AI datacenter spend by decade‑end contrasts with an investor slide that implies ~$10.9T, and Nvidia projects roughly $500B of Blackwell/Rubin GPU system sales through end‑2026. Nvidia’s cash war chest, 8.3% R&D spend ($4.7B), and rapid growth in networking (InfiniBand, NVSwitch) underpin its market position even as hyperscalers and rivals push alternatives. Significance: this shows how GenAI driven capex can sustain Nvidia even if broader market valuations wobble. Technically, Nvidia’s advantage is compositional—GPUs plus system components, networking and software stack—making it hard for one competitor to displace the whole offering quickly. But the piece warns of two separable “bubbles” (infrastructure capex vs. startup/market valuations): if one pops it can pressure the other, and rising competition in AI XPUs and skepticism about ultra‑large capex forecasts could trim Nvidia’s share despite continued growth.