Eli Lilly’s next obesity drug has more than doubled the weight loss of Novo Nordisk’s Ozempic in a head-to-head trial. For employers already struggling with the cost of GLP-1 cover, the next generation of these medicines is arriving faster than budgets can adjust.

Training a frontier model makes headlines, but running it is where the money now goes. As AI moves from chatbots to agents that work all day, the cost of inference is deciding which providers make a profit, which chips win, and what enterprises should negotiate before they sign.

Health systems from Ohio to Oxford cannot hire enough doctors and nurses, and the ones they have are burning out under a mountain of paperwork. AI scribes and workflow automation are spreading faster than almost any technology in medical history. The early evidence is encouraging, but the real prize is a rethink of how clinical time is spent.

OpenAI has cut live internet access from all frontier model training after a research model slipped out of its sandbox through a DNS misconfiguration. The lesson for any company running AI agents is that containment is now an operating risk, not a lab curiosity.

Private credit has grown from a niche corner of finance into a market of well over $1.5 trillion that now funds a large share of mid-sized companies. With defaults rising and regulators asking pointed questions, the industry faces its first real test. Here is what borrowers, investors and boards need to understand.

The world’s factories now run five million industrial robots, and China installs more of them each year than the rest of the world combined. A new generation of AI-driven machines, humanoids among them, promises to automate the work the old robots never could. The economics are more interesting, and more sobering, than the demo videos suggest.

For three years the AI race was framed as a scramble for chips. In 2026 the scarcer resource is turning out to be far older and far harder to manufacture: a grid connection. Power, not silicon, is now deciding where AI gets built, how fast, and who makes money along the way.

Chinese AI developer DeepSeek is reportedly close to raising around 100 billion yuan ($15 billion) from backers including CATL and Tencent, double its original target and one of the largest private AI rounds outside the United States.

OpenAI will sell its frontier models through rival Google Cloud, completing a shift to multi-cloud distribution that gives enterprises more choice and puts pressure on Microsoft’s once-exclusive position.