Intel: Written Off Too Soon
Intel was one of the most obvious investment opportunities when the current U.S. administration took office. The company had — and still has — a tremendous amount of work ahead of it, but the pieces have been falling into place in ways that few anticipated.
The collaboration between Elon Musk and CEO Lip-Bu Tan on the TeraFab initiative has been a meaningful catalyst. Perhaps even more telling is that Google — a company that designs and manufactures its own custom CPUs — has come forward saying it needs Intel processors. When a competitor of that caliber validates your product, it speaks volumes about the shifting landscape.
The Agentic Era Changes Everything
The development that caught most observers off guard is the sudden resurgence in demand for data center CPUs. The rise of the agentic AI era — where autonomous AI agents perform complex, multi-step tasks — has reignited demand across the entire CPU ecosystem. ARM has gotten a boost. AMD is getting a boost. And Intel is getting a boost right alongside them.
This is a paradigm shift. For years, the narrative around AI infrastructure centered almost exclusively on GPUs. But agentic workloads require robust CPU capabilities for orchestration, memory management, and general-purpose compute that GPUs alone cannot efficiently handle. The market had written Intel off too soon, failing to account for the possibility that the next wave of AI would bring CPUs back into the spotlight.
Beyond the business fundamentals, there is a strategic argument: the United States needs a domestic chip champion. It needs a leading-edge foundry capable of manufacturing advanced semiconductors on home soil. Intel is uniquely positioned to fill that role, and the geopolitical tailwinds behind domestic chip manufacturing are unlikely to fade anytime soon.
Nebius: A Cloud Built for the AI Future
Nebius is another name that has delivered an exceptional run. The company has positioned itself as a cloud platform purpose-built for AI development, and that thesis has played out impressively. The core appeal is straightforward: as the AI ecosystem expands, developers and enterprises need infrastructure specifically optimized for training and deploying AI models. Nebius fills that niche.
That said, after such a tremendous rally, investors need to keep a close eye on valuation. The fundamentals remain attractive — a cloud platform tailored for the AI future is exactly the kind of business the market should reward — but price discipline matters. The underlying business is still compelling, but the question shifts from "is this a good company?" to "is it priced appropriately given how far it has already run?"
The Broader AI Infrastructure Ecosystem
Beyond Intel and Nebius, the AI infrastructure investment thesis extends into what might be called "AI GPU REITs" — companies that provide data center power and infrastructure to hyperscalers like Google. These businesses sit at the foundation of the AI boom, supplying the physical power and facilities that make large-scale AI compute possible.
This layer of the stack is often overlooked in favor of flashier chip designers and software companies, but it represents a critical bottleneck. As AI workloads scale, the demand for power, cooling, and physical data center capacity only grows. Companies operating in this space offer a more utility-like exposure to the AI megatrend, potentially with more predictable cash flows than the semiconductor names further up the stack.
Conclusion
The current AI investment landscape rewards those who look beyond the obvious GPU narrative. Intel's resurgence on the back of agentic AI demand and strategic national importance, Nebius's positioning as an AI-native cloud, and the foundational role of data center infrastructure companies all point to the same conclusion: the AI boom is broadening, and the opportunities are more diverse than many investors initially recognized. The key going forward is balancing conviction with valuation discipline, especially after the significant rallies these names have already experienced.