AI This Week

China now mandates its chipmakers use at least 50% domestically made equipment to expand capacity in a major move to build a self-sufficient semiconductor supply chain. The undocumented rule forces companies seeking state approval for new factories to prove they meet the threshold. This policy directly advances Beijing’s goal of reducing reliance on foreign technology, especially after recent U.S. export restrictions. The directive is compelling Chinese manufacturers to select local suppliers over available foreign options. While authorities offer some flexibility for the most advanced production lines where domestic tools are scarce, the long-term objective points toward 100% domestic sourcing, significantly altering the global market.

Microsoft CEO Satya Nadella is overhauling his leadership team in a major move to forge the company’s AI strategy. The shake-up signals a dramatic pivot as the tech giant plots a course extending far beyond its deep and defining partnership with OpenAI. After investing billions to weave OpenAI’s technology into products like Copilot and Bing, Microsoft is now building a new foundation. This reorganization aims to diversify its AI assets and reduce its strategic reliance on a single, vital ally, fundamentally reshaping its approach to the AI race. The changes suggest Microsoft is preparing to secure its own long-term dominance.

Meta Platforms joins the year-end AI spending frenzy, agreeing to acquire Manus, a Singaporean startup building autonomous AI agents. The price tag tops $2 billion. This deal culminates a flurry of activity from tech titans, including SoftBank’s recent data-center investments and Nvidia’s pact with Groq, as they scramble for dominance. Manus hit a $100 million revenue run rate with shocking speed, reshaping how consumers use AI. Analysts are already comparing the move to Meta’s landmark acquisitions of Instagram and WhatsApp, signaling a deep bet on the future of AI interaction. Meta plans to integrate the technology across its product lines while keeping the existing service live.

Groq and Nvidia have struck a significant non-exclusive licensing agreement for Groq’s AI inference technology. The deal is designed to expand global access to high-performance, low-cost inference solutions. But this is more than just a technology pact. In a stunning move, Groq Founder Jonathan Ross, President Sunny Madra, and other key team members will join Nvidia to help scale the licensed technology. The announcement confirms Groq will continue operating as an independent company. Simon Edwards steps in as the new Chief Executive Officer. Customers can expect no service interruptions, as its GroqCloud platform will remain fully operational throughout the transition.

NOAA unveiled AI-driven forecasting systems that promise quicker, more accurate predictions at a fraction of computing cost. Officials switched them on Wednesday for use. The new models complement, not replace, physics-based staples like the Global Forecast System and GEFS. Trained on decades of analysis data, they cut compute needs by 91% to 99% and can extend forecast skill by up to a day. AIGFS produces a 16‑day outlook using just 0.3% of GFS resources and finishes in about 40 minutes. AIGEFS adds probabilistic guidance, while a Hybrid‑GEFS blends AI with traditional ensembles to capture uncertainty. NOAA still works on hurricane guidance and outcome diversity. Leaders cite lower costs but note the heavy energy footprint of model training.