Cathie Wood Says Wall Street Is Missing The Next Big AI Trade — And It's Not Nvidia

For nearly two years, the artificial intelligence boom has largely centred on one trade: graphics processing units (GPUs). That made Nvidia one of the clearest winners of the AI race, as demand for chips surged across data centres, cloud providers, and large model developers.

Now, investor Cathie Wood says Wall Street may be overlooking another part of the AI buildout. Speaking about the next phase of artificial intelligence infrastructure, Wood pointed to CPUs and legacy technology companies that may benefit as AI workloads evolve beyond training models. Her argument comes as AI begins moving into inference, automation, and agentic systems.

Why CPUs Are Back in Focus

Wood referenced comments made by OpenAI CFO Sarah Friar, who said investors may be surprised by how agentic AI could increase CPU usage. According to Wood, current AI infrastructure often operates at roughly four to five GPUs for every CPU. She said that ratio could move closer to one-to-one in the future.

If that shift happens, it would mark a major change in how AI systems are built. GPUs remain critical for training large-scale AI models because they handle heavy parallel computing tasks. But inference workloads operate differently.

Inference includes real-time responses, task execution, workflow handling, and autonomous AI agents interacting across software systems. Those tasks often rely on traditional computing functions where CPUs remain important. Wood described this as a possible overlooked trade inside the AI boom.

The Stocks Cathie Wood Highlighted

Wood pointed to several older technology firms that could benefit if AI infrastructure spending broadens.

• Intel – Intel has been working to regain momentum through foundry expansion and AI chip development. A stronger CPU demand cycle could support its relevance in the AI market.
• Cisco Systems – Cisco is tied to AI networking demand. As data movement between AI systems increases, networking infrastructure may become more important.
• Corning – Corning's fibre and connectivity products are linked to data-centre expansion, which remains a core part of AI infrastructure growth.
• Flex – Flex plays a role in manufacturing and supply-chain deployment tied to AI hardware systems.
• Akamai Technologies – Wood also highlighted Akamai, noting that AI demand may extend into edge computing and cloud infrastructure beyond semiconductor makers.

AI Spending Is Broadening

Wood's broader point is that AI may no longer be a narrow semiconductor trade. The first stage of AI growth was heavily focused on model training, chipmaking, and hyperscale infrastructure. As AI expands into enterprise automation, inference, networking, and system integration, spending may spread across older industrial and infrastructure firms. That could explain renewed market attention on companies once associated with earlier technology cycles.

What This Means for Investors

Wood's view does not suggest GPUs are losing importance. Nvidia remains central to AI training and high-performance computing. But her argument highlights a broader shift. If AI adoption increasingly depends on inference and autonomous systems, CPU-heavy and infrastructure-linked businesses may also capture part of that demand.

For investors, the key question is whether AI growth remains concentrated in chip leaders or expands across the wider technology supply chain. That is where Wood believes Wall Street may be underestimating the next trade.