Expanse raises $5.3 million to cut AI resource waste

The article frames Expanse's seed round around a specific inefficiency: AI workloads often reserve more compute than they use. By predicting the exact GPU, CPU, and memory a task needs before it runs, Expanse aims to minimize that over-allocation and make existing hardware go further. The article cites industry estimates that significant percentages of cloud expenditures are wasted on underutilized resources, which is the problem the software is pitched against.

The broader market snapshot in the same piece shows AI infrastructure still moving stock prices. Super Micro Computer was a standout, up 9.5% to $40.35, while ServiceTitan fell 7.5% to $54.25, close to its 52-week low. Oracle rose 5.2% to $150.59 after announcing Java 27 advancements and a Quorum Health partnership, and Microsoft settled at $497.75, up 1.5%, after alliances with Marvell and Nokia.

Whether Expanse's prediction approach translates into real savings is the open question. Seed-stage funding gives it room to prove the technology, but the article offers no customer deployments or measured savings, so the value hinges on whether accurate predictions hold up in live production workloads.