GPUs

Arm's new CSS for Mobile 2 platform pairs neural accelerators with GPUs to handle persistent AI agents within smartphone power budgets, signaling a shift toward on-device AI workloads over cloud inference.

The chip giant deployed over $40 billion in 2026 to build influence across the AI stack, from frontier labs to cloud providers—a tenfold surge that rewrites how platform companies entrench market power.

AMD claims 50 percent more memory and superior precision performance in new data center GPUs, but execution on ROCm software and enterprise adoption remain unproven.

Earnings reveal Nvidia's strategic pivot toward specialized systems and storage optimization as gigawatt-scale data centers become the new competitive battleground.

SpaceX's AI infrastructure expansion may prompt Nvidia to prioritize its chip allocations, tightening GPU supply for neocloud providers CoreWeave and Nebius.

The long-term agreement secures critical high-bandwidth memory for Nvidia's H100 and B200 GPUs, tightening its grip on AI accelerator supply chains.