CoreWeave shares jump nearly 16% after results as analyst points to the longevity of NVIDIA A100 GPUs

CoreWeave Inc. shares surged nearly 16% in after-hours trading Tuesday following the company’s second-quarter results, according to Yahoo Finance. The move underscored market focus on whether cloud and AI infrastructure providers can keep deploying cost-effective compute hardware as the AI industry moves through hardware generations.

While the earnings release itself set the day’s tone, the stock reaction also reflected commentary from analyst Daniel Newman. Newman argued that “old” NVIDIA GPUs can still be monetized long after their introduction, pointing to an extended NVIDIA A100 contract discussed in the segment.

NVIDIA’s A100 is a data-center accelerator used to train and run AI models. In the context of the report, the key point was not just that A100 units are being used, but that contract arrangements around them can extend, allowing an operator like CoreWeave to continue generating revenue with existing installed technology rather than replacing hardware on a fixed schedule.

Newman’s framing suggested that customers and operators are willing to keep systems in service even as newer GPU families arrive. If true at scale, that approach can affect demand timing across the supply chain, including component vendors and the makers of the platforms that run large-scale AI workloads.

For CoreWeave, the immediate implication is that its financial performance and guidance depend not only on overall AI spending, but also on how effectively it can match workloads to available GPU capacity. Long-lived contracts for earlier-generation accelerators could support utilization and revenue over time, potentially smoothing the operational challenges that come with frequent technology refresh cycles.

Sector-wide, investors have increasingly treated hardware utilization and contract durability as leading indicators in AI infrastructure businesses. The commentary described in the Yahoo Finance article fed into that narrative by emphasizing that procurement and deployment are not strictly “newest model wins,” but can reflect performance-per-dollar and existing infrastructure economics.

Still, the report did not provide granular details in its brief market summary, such as the contract size, the exact duration of the A100 arrangement, or how much of CoreWeave’s revenue is tied to A100 deployments versus newer NVIDIA platforms. Without those figures, it is not possible to quantify the contribution of older GPUs to CoreWeave’s future results based solely on the headline market reaction.

Heading into the next reporting window, investors will likely look for more clarity on how CoreWeave’s GPU fleet is composed, what portion of revenue is linked to specific NVIDIA GPU generations, and whether management expects sustained demand across both legacy and newer hardware. The durability of those arrangements, and the pace of any incremental spending, could determine whether Tuesday’s after-hours jump extends into the regular session.