Nvidia teams up with a Goldman-backed startup to tackle a major pain point in AI adoption (www.businessinsider.com)

Nvidia has partnered with Goldman-backed startup Spectro Cloud to debut PaletteAI, a platform aimed at simplifying and optimizing AI infrastructure. Announced ahead of Nvidia’s GTC conference, the collaboration integrates PaletteAI with Nvidia’s AI Enterprise suite to tackle a widespread pain point: low GPU utilization. Spectro Cloud says many deployments use only ~30% of GPU capacity; PaletteAI can push that toward ~60%, effectively doubling hardware efficiency. For organizations racing to scale AI, that translates to substantial cost and energy savings, faster model iteration, and a lower barrier to adoption. Technically, PaletteAI acts as a “glue layer” that coordinates disparate hardware and software across data centers and the edge, improving resource scheduling, orchestration, and security while remaining open to non‑Nvidia vendors. By streamlining management of GPUs, clusters, and software stacks, it reduces operational complexity that often stalls AI projects. The move signals growing focus on software-first optimizations to extract more value from existing infrastructure, not just buying more accelerators—an important shift for enterprises balancing performance, cost, and sustainability as AI scales.