The mind-boggling valuations of AI companies (www.theguardian.com)

AI’s money storm just got bigger: Nvidia briefly hit a $5 trillion valuation as Microsoft and Apple topped $4 trillion, while Alphabet posted its first $100 billion quarter and bumped planned capital spending to roughly $91–93 billion for the year. The market’s frenzy extends beyond public companies: OpenAI — eyeing a possible $1 trillion IPO — has struck massive cloud and investment deals (Nvidia ~$100B, Microsoft ~$250B of Azure spending commitments, Oracle ~$300B, AWS ~$38B) with headline figures reported around $588 billion in future spend. Together these moves signal enormous, coordinated investment in chips, datacenters and cloud capacity that underpins modern AI. That scale matters technically and economically. Large language models are ballooning into the hundreds of billions — sometimes trillions — of parameters, demanding prodigious compute, power and real-world infrastructure (witness the sprawling Tahoe‑Reno data‑center complex). But the boom carries risks: many enterprise AI pilots fail (MIT puts that rate near 95%), value concentration and intertwined deals raise systemic fragility concerns, and vast capex bets may outpace proven product-market fit. For AI/ML communities this means fierce demand for optimized hardware, energy-efficient model engineering, and careful evaluation of deployment economics, while policymakers and engineers should watch for cascading financial, supply‑chain and regulatory consequences.