AI breakthrough could hit H1 2026
Morgan Stanley warns a major AI breakthrough could arrive in the first half of 2026 — and the surge in compute required could strain power grids and trigger large labour‑market disruption, analysts say warns. Professors are already reporting AI tools are eroding students' ability to think independently, while other analyses argue highly productive workers may be relatively insulated as automation raises the premium on creative, value‑added skills reported argued.
The note was authored by Stephen Byrd, Morgan Stanley’s Global Head of Thematic and Sustainability Research, and frames AI progress through an internal “Intelligence Factory” model that links compute scale to firm-level value creation and infrastructure needs. pod.wave.co Morgan Stanley’s analysis, as reported, projects a net U.S. power shortfall of roughly 9–18 gigawatts through 2028. finance.yahoo.com Other coverage of the same Morgan Stanley work has quoted a larger estimate of about 44 gigawatts (near a 20% shortfall) through 2028, showing media reporting varies on the magnitude. uk.investing.com To keep pace with demand, operators are already converting Bitcoin‑mining sites into high‑performance computing centers, running temporary natural‑gas turbines, and deploying fuel cells for colocations, according to reporting tied to Morgan Stanley’s briefing. finance.yahoo.com Separately, the firm flags memory and data‑center infrastructure as emerging supply bottlenecks for next‑generation models. trustfinance.com Multiple industry accounts referenced plans at leading labs to scale training compute by roughly 10x over recent cycles, a jump commentators have linked to Elon Musk’s public claim that a tenfold increase in compute can materially boost model capabilities. ainvest.com Morgan Stanley’s modelling also estimates roughly $920 billion in annual net benefits to S&P 500 firms from AI adoption, while flagging the potential for widespread task automation across many jobs. investing.com Independent research from Brookings finds that more than 30% of workers could see at least half of their occupation’s tasks disrupted by generative AI, and academic work shows productivity gains often accrue disproportionately to the most‑productive agents. brookings.edu Faculty surveys and university actions already document classroom consequences and policy responses: an Inside Higher Ed survey of 1,057 U.S. faculty found about 90% believe generative AI diminishes students’ critical thinking and 95% expect rising overreliance, while College Board polling of 3,000+ faculty reported near‑universal concern. insidehighered.com Colleges are responding by rewriting honor codes, redesigning assessments, and rolling out ad‑hoc AI policies as campus surveys report student AI adoption rates around 80–85%. kanw.org
