A major new player is entering the AI computing race, but one of the industry's leaders isn't worried—in fact, they see it as a good sign.
Recently, investment giant Blackstone and Google announced a joint venture to create a new cloud service based on Google's TPUs, specialized chips for AI. This news initially made investors nervous about competition for companies like CoreWeave, which specializes in cloud services using Nvidia's GPUs. However, CoreWeave quickly reframed the narrative, arguing this new venture is actually evidence that the demand for AI computing is growing faster than anyone can supply it.
CoreWeave's confidence stems from a few key factors. First, they have the numbers to back it up. The company revealed a staggering revenue backlog of nearly $100 billion, with over $40 billion in new deals signed in just one quarter. This, they argue, shows that customers are actively choosing and committing to Nvidia's GPU-based systems for the long term. The market's immediate reaction was a slight dip in CoreWeave's stock, prompting the company to assert its strong position.
Second, there's the technology ecosystem. Most AI applications today are built using Nvidia's CUDA software platform. Developers are familiar with it, and a vast library of tools is optimized for it. Switching to a different architecture like TPUs isn't a simple flip of a switch; it requires significant effort. Other independent cloud providers have echoed this, stating that around 99% of their customer demand is specifically for Nvidia GPUs.
Finally, CoreWeave has been preparing for this growth. The company has secured billions in financing, including loans backed by its valuable GPU hardware. This capital is crucial for building the data centers needed to turn its massive backlog into actual revenue. Their deep partnership with Nvidia also ensures they get early access to the latest and most powerful chips, keeping them at the cutting edge.
So, while the Blackstone/Google TPU cloud is a significant development for the future, its impact won't be felt immediately, as the first data centers are planned for 2027. For now, CoreWeave's message is clear: the AI pie is getting bigger for everyone, and their slice, built on a foundation of strong customer demand for GPUs and solid financing, remains secure.
- Glossary
- TPU (Tensor Processing Unit): Google's custom-designed chip specifically for accelerating AI and machine learning tasks.
- GPU (Graphics Processing Unit): A processor originally for graphics, now widely used for its parallel processing power in AI, especially those made by Nvidia.
- Backlog: The total value of confirmed customer orders that have not yet been fulfilled or recognized as revenue.