- CoreWeave shares jump 8% after unveiling serverless AI tools for faster, cheaper model training.
- New service automates GPU scaling, cutting costs by 40% without impacting model quality.
- Strong AI demand fuels CoreWeave deals with OpenAI, Meta, and plans to acquire Core Scientific.
Shares of CoreWeave surged as much as 8% on Wednesday following the announcement of new tools aimed at helping developers create artificial intelligence agents.
The cloud provider, which specializes in AI infrastructure, unveiled a serverless reinforcement learning service designed to simplify the process of training AI models.
New AI tools lower costs and boost efficiency
CoreWeave’s latest offering automatically adjusts computing power, eliminating the need for developers to manually add or remove resources.
According to the company, tests showed that using the service allows AI models to be trained up to 40% faster and at lower cost compared to running Nvidia H100 graphics processing units locally, with no compromise on model quality.
Reinforcement learning, the technique supported by the new service, is a long-established AI method that improves outcomes through iterative trial and error.
By automating infrastructure management, CoreWeave aims to make this approach more accessible and efficient for developers.
This launch builds on the company’s previous efforts to expand its AI offerings. Five months ago, CoreWeave acquired Weights and Biases for $1 billion.
The startup provides software for training and evaluating AI models, complementing CoreWeave’s existing business of renting Nvidia GPUs to companies requiring high-performance computing for AI projects.
Rising demand for GPUs and cloud AI services
The demand for GPUs to power AI projects has surged, prompting companies to secure resources either in the cloud or within their own data centers.
CoreWeave competes with major cloud providers, including Amazon Web Services, but its focus on AI infrastructure gives it a niche in the rapidly growing market.
Recent agreements illustrate this trend. CoreWeave said two weeks ago that OpenAI had agreed to expand a multi-year deal with the company by up to $6.5 billion.
Last week, Meta committed to spending $14.2 billion on the platform.
The strong uptake reflects the broader rush among companies to secure GPU capacity for AI development.
In addition to service expansions, CoreWeave announced in July plans to acquire data center infrastructure provider Core Scientific for $9 billion.
Some Core Scientific shareholders have expressed dissatisfaction with the terms of the deal, recommending it be voted down, but CoreWeave’s CEO Mike Intrator emphasized that the company will not revise the acquisition offer.
Public market performance and outlook
CoreWeave, headquartered in New Jersey, went public on Nasdaq in March.
Its recent stock performance reflects investor enthusiasm over the company’s growing presence in the AI sector and its ability to provide cost-efficient infrastructure for large-scale AI model training.
The new serverless reinforcement learning service, along with past acquisitions and high-profile client agreements, positions CoreWeave as a significant player in the AI cloud market.
By combining scalable GPU resources with developer-friendly software, the company seeks to capture increasing demand from organizations looking to implement advanced AI solutions.
As competition intensifies among cloud providers and AI infrastructure firms, CoreWeave’s approach may offer an edge by reducing costs, streamlining operations, and attracting high-profile AI clients.
Investors will be closely monitoring how the company executes its expansion plans, particularly the Core Scientific acquisition, while continuing to meet demand in the fast-moving AI sector.