The geopolitical side is getting complicated. Nvidia is under congressional scrutiny for allegedly helping optimize DeepSeek's V3 model, with fears that civilian AI technology is flowing into China's military. China conditionally approved DeepSeek to purchase Nvidia's H200 chips, adding more tension to an already strained relationship.

CATL, the world's largest battery manufacturer started commercial production of sodium-ion batteries this year. Sodium is more abundant and cheaper than lithium, which is currently mined in just a handful of countries. While sodium-ion batteries have slightly lower energy density today, costs are expected to drop significantly as production scales. CATL's Naxtra sodium-ion line is already heading into vehicles in 2026.

The four major hyperscalers; Google, Microsoft, Amazon, and Meta are spending nearly seven hundred billion dollars on AI infrastructure in 2026. That's sixty-five percent more than last year, which was already a record. And it's triple what they spent just two years ago.

The broader medical AI market is accelerating. AI is now diagnosing cancers from imaging scans with accuracy matching or exceeding specialist radiologists. Startups are building AI copilots for psychiatry that help therapists with session notes, treatment planning, and patient monitoring.

Big Tech is going nuclear to power AI data centers. Meta, Microsoft, and Amazon signed massive nuclear energy deals as data center electricity demand surges. SMRs are the technology making it possible.

Microsoft revealed that roughly forty-five percent of its six hundred twenty-five billion dollars in expected future cloud contracts comes from OpenAI. That single relationship represents over two hundred eighty billion dollars in projected revenue.

The competitive dynamic is fascinating. Microsoft's closest relationship is with OpenAI 45% of its five hundred twenty-five billion dollars in cloud backlog is reportedly from OpenAI. That's a concentration risk. If OpenAI struggles, Microsoft struggles. Google doesn't have that dependency. Its AI capabilities are native and owned outright.

Juniper Research forecasts twenty-one billion dollars in operator investment in AI-integrated network technology this year alone. Operators report saving thousands of person-years of work annually and cutting network downtime by tens of thousands of hours per month just from early AI network automation.

Shield AI, which builds autonomous AI pilots for military aircraft, is another standout. Their Hivemind software can fly fighter jets and drones without GPS or communications critical for contested environments. The company raised five hundred million at a five billion dollar valuation.

Personalized CRISPR saves a baby's life, Roche sequences a full genome in four hours. AI cutting drug development from twelve years to five, and Colossal Biosciences bringing back dire wolves. Biotech's biggest breakthroughs are here right now.