AI Data Center Power Bottleneck Makes Bloom Energy This Analyst's Top 2026 Stock Pick
Beth_Kindig · x · 2026-09-19
Beth Kindig argues the real risk to the AI economy isn't compute or software but power availability — specifically time-to-power. Key data: Big Tech is spending $650B+ annually on AI data centers; NVIDIA's GB200/GB300 racks draw 120-140kW (2x H200's 70kW); rack-scale systems will need 300-600kW within 1-2 years, a 5x jump from H1 2025. Grid and nuclear expansion are too slow, so her top 2026 pick is Bloom Energy, whose solid oxide fuel cells provide behind-the-meter on-site generation to bring new data centers online faster ($BE, $NVDA).
More from Venture
- Those 'free' AI engineers in your enterprise are mostly sales engineers, says Linthicum — DavidLinthicum · 2026-09-19
- Indie Hacker's Twin SaaS Flywheel: SiteGPT Near $30k, Squad Near $10k MRR — tibo_maker · 2026-09-19
- Coefficient Giving has millions for AI safety nonprofits but can't find founders — AndyMasley · 2026-09-19
- AI is cutting startup costs, letting founders skip VC as Bending Spoons debate rages — prasanna_says · 2026-09-19
- Angle Health raises $200M Series C at $2.7B valuation for AI-native health benefits — ycombinator · 2026-09-19
- SEO tool Index My Site hits 100 daily visitors and 18.8k indexed pages in 11 days — iannuttall · 2026-09-19