Bank of England smells hint of dotcom bubble 2.0 in AI froth (www.theregister.com)

The Bank of England’s Financial Policy Committee warned that sky-high valuations in tech and AI stocks—driven by a recent flurry of deals and massive infrastructure commitments—look strikingly like a replay of the dotcom era and leave markets vulnerable to a “sudden correction.” The committee flagged record concentration in US indices (the top five S&P 500 firms now account for roughly 30% of the market) and said valuation metrics such as the CAPE-implied earnings yield are comparable to dotcom peaks, even if not yet exceeding them. High-profile tie-ups and spending commitments (e.g., Nvidia/AMD/OpenAI-linked deals and Oracle’s cited $455B cloud pipeline, with OpenAI reportedly committing ~$300B) have helped push prices to stretched levels. For the AI/ML community this is a cautionary signal: downside risks include slower-than-expected capability or adoption, intensified competition, and material bottlenecks in power, data and commodity supply chains. Analysts warn a shakeout among model builders is likely (Gartner calls it an “extinction event” for some providers), while consultancies estimate infrastructure spending needs of up to $500B/year—so a sharp pullback could ripple through growth figures and funding. Real-world failures (for example, Deloitte’s AI-generated government report with fabricated citations) underline that hype isn’t a substitute for robust product-market fit and reliable data and compute engineering.