The AI boom was already straining the supply chain, and agentic AI is apparently pushing it to a new level of dysfunction.
Speaking at Dell Technologies World 2026 in Las Vegas, Dell Technologies CEO Michael Dell said the rise of autonomous AI systems is making it harder for supply and demand to settle into any kind of balance.
“If it had not been for this agentic movement, we would probably see much better demand-supply equilibrium in the second half of next year,” he said.
That balance is getting thrown further out of whack. Demand for AI infrastructure keeps inching gradually upwards, but the supply side doesn’t move nearly as fast. Building semiconductor fabs still takes years, and AI demand sure isn’t slowing down to wait for them.
Agentic AI just adds to it. These systems are built to act on their own, which usually means more compute, more storage, and everything working together more tightly. Scale that across industries and regions, and the strain starts to add up pretty darn fast.
Beyond chips: costs spread across the stack
That strain is no longer limited to semiconductors. Dell pointed to rising costs in raw materials and logistics, with geopolitical tensions adding further volatility to an already complicated picture.
The result is a supply chain where constraints show up in multiple places at once – hardware availability is one piece, but so is the cost of moving that hardware and the materials that go into it.
Dell said securing supply remains a top priority, and scale has helped the company navigate some of these challenges. Still, though, the approach is becoming more proactive. It has to.
“For our important customers, we are able to create a framework to work with them where they are assured access to the supply they need,” Dell said.
Dell pushes partners and customers to plan earlier
That framework starts earlier in the buying cycle, with customers being asked to plan further ahead, lock in demand, and work through allocation before shortages become an issue.
MSPs planning AI infrastructure projects should treat supply availability as a strategic constraint, not a procurement detail.
If agentic AI adoption continues to drive demand for compute, storage, and supporting hardware, providers may need to forecast customer needs earlier, secure capacity sooner, and set clearer expectations around deployment timelines.
Dell also used the conference to roll out new partner incentives and program changes tied to this evolving market.
“Customers are not asking whether to modernize their infrastructure, rethink their cloud strategy, or invest in cyber resilience,” Chief Partner Officer Denise Millard said. “They are asking how fast they can do it and who they can trust to help them get there.”
Supply falls short of demand
One of the more telling comments from Dell was how widespread the demand has become.
“One of the interesting things about this is how universal the demand is; it’s in every country we are allowed to sell in,” he said.
That’s all fun and games, but demand being this widespread makes it messy. It’s not one region causing the squeeze; it’s a lot of them, all at once.
Global AI demand complicates allocation
On top of that, there are still regional rules to deal with, especially in parts of Asia where access to more advanced AI systems can be super tight.
Dell said it’s continuing to supply customers within those limits, including for bigger AI buildouts.
So you end up in this spot where demand is the hare that just keeps hopping further and further ahead, and supply is the tortoise, almost comically trying to catch up.
If you’re building out AI right now, timing starts to matter in a very real way. Not just what you’re buying, but when you can actually get it.