Tesla chip surge begins, factory scale ambitions & more

Enterprises are accelerating AI infrastructure investments while hardware makers expand capabilities, driving a wave of inference‑focused deployments across cloud, automotive, and defense. At the same time, governance and data‑rights issues are emerging as critical constraints on rapid growth.

• NVIDIA announced a next‑generation AI stack that adds open foundation models, agent tooling, and domain‑specific platforms for robotics and drug discovery, extending its GPU ecosystem into inference‑heavy workloads.
• Tesla began construction of a terafab to mass‑produce its AI5 chip, promising a forty‑fold performance boost for Full Self‑Driving and Optimus robot applications.
• Meta signed a $27 billion AI cloud infrastructure deal with Nebius, leveraging Nvidia’s Vera Rubin platform to scale its large‑model services while reducing reliance on internal data centers.
• OpenAI reorganized its Stargate team to prioritize renting external AI compute, aiming to lower costs and improve scalability as its services expand.
• Anduril secured a multi‑year U.S. Army contract valued up to $20 billion for AI‑driven surveillance and autonomous systems, highlighting defense as a major enterprise adopter.
• Moonshot AI introduced Attention Residuals, a technique that improves compute efficiency by 1.25×, offering a low‑cost performance gain for large language models.