A Direct Message From Davos
Imagine being the CEO of a company that relies heavily on a single supplier for a critical component. Now imagine that component is the absolute bedrock of the artificial intelligence boom. That’s a bit like the situation for CoreWeave, a cloud provider specializing in AI workloads. And its CEO didn’t hold back at the World Economic Forum in Davos in January 2026.
The message, directed at Nvidia stock investors, was clear: Nvidia needs to keep expanding its AI capacity, or it risks seeing customers, including CoreWeave, turn to alternatives like AMD. It’s a direct warning that highlights the intense competition and demand in the AI infrastructure space.
The Supply and Demand Dance
For those of us watching the AI world, this isn’t just about corporate rivalries. It’s about the very infrastructure that enables AI agents to learn, adapt, and perform their tasks. If there aren’t enough specialized chips, the progress of AI itself can slow down.
Nvidia has been the dominant force in AI accelerators for a long time, and for good reason. Their GPUs are essential for training and running complex AI models. But dominance doesn’t mean immunity from competition or from the simple economics of supply and demand. When demand for something is astronomical, and supply struggles to keep up, customers start looking elsewhere.
AMD’s Opportunity
This is where AMD enters the picture. While Nvidia has enjoyed a significant lead, AMD has been working to catch up, developing its own AI-focused hardware. The CoreWeave CEO’s statement suggests that if Nvidia can’t meet the needs of its customers, those customers are ready and willing to explore what AMD has to offer. This isn’t just theoretical; it’s a real threat to Nvidia’s market share if capacity issues persist.
For AI agents, having more options for underlying hardware is a good thing. It fosters competition, which can lead to faster innovation and potentially more affordable computing resources. Think of it like needing a powerful engine for your AI car – having two solid engine manufacturers is better than just one.
Nvidia’s Investment and “Circular Deals”
Interestingly, Nvidia itself has made a significant investment in CoreWeave, to the tune of $2 billion. This kind of investment can sometimes raise eyebrows, leading to questions about “circular deals” – where one company invests in another that is also a major customer. Jensen Huang, Nvidia’s CEO, addressed these concerns directly, calling the suggestion that the CoreWeave investment was a “circular deal” “ridiculous.”
From Nvidia’s perspective, investing in a key partner like CoreWeave could be seen as a way to ensure their chips are being used effectively and at scale, further solidifying their position in the AI cloud space. For CoreWeave, it likely provides capital to expand its own infrastructure, which in turn means more demand for those essential AI chips. It’s a complex relationship in a rapidly evolving market.
What This Means for AI Agents
As someone who explains AI agents to non-technical people, this whole situation boils down to one thing: accessibility to computing power. AI agents, whether they’re helping you write emails, manage your calendar, or even create art, need powerful processors to do their work. The more available and diverse these processors are, the better it is for the development and deployment of AI.
A competitive environment, spurred by warnings like the CoreWeave CEO’s, could ultimately benefit everyone who uses or develops AI agents. It pushes chip makers to continually improve their offerings and expand their production. It also encourages cloud providers like CoreWeave to build out more capacity, making it easier for developers to get the computing resources they need to bring new AI agents to life.
The AI space moves at incredible speed, and the battle for the underlying hardware is just as intense. Keeping an eye on these developments helps us understand the foundations upon which our future AI assistants are being built.
🕒 Published: