Wall Street closes at a record for the first time since end of January
The AI trade has moved in waves—semiconductors, then software, then cloud infrastructure. Each one rewarded early investors and punished latecomers. The next wave is already building, and it has nothing to do with data centers or chatbots. It’s edge AI, the technology that puts artificial intelligence directly inside machines. Keith Kaplan, CEO of TradeSmith, has been tracking this shift closely, and he sees three companies across very different risk profiles positioned to ride it.
Why AI Can’t Stay in the Cloud
Most people still picture AI as a conversation with ChatGPT: type a question, wait for a server miles away to respond, get an answer. That model works fine for text. It falls apart the moment AI has to operate in the physical world.
A Tesla at highway speed has roughly 100 milliseconds to spot a pedestrian, read a signal, and decide whether to brake—a round trip to a cloud server takes too long. John Deere combines make thousands of decisions per minute in fields with no cell signal. Military drones processing targeting data can’t rely on a connection an enemy jammer that could cut out at any moment. In every case, the AI has to already live inside the machine. That’s edge AI—and the companies building chips, rugged hardware, and thermal infrastructure for it are facing a demand wave most investors haven’t priced in yet.
The edge AI market was roughly $11.8 billion in 2025 and is projected to approach $60 billion by 2030, representing nearly 37% annual growth. But the real story may move faster than projections suggest. Unlike the cloud AI boom, which was driven by a handful of hyperscale spenders, edge AI demand could spread across every autonomous vehicle, smart factory, hospital, satellite, power grid, and eventually every home—billions of devices across thousands of industries.
And that’s the scale of what’s coming.
1. Honeywell: The 140-Year-Old Edge AI Cornerstone
Honeywell is nobody’s idea of a hot AI stock, and that’s exactly the opportunity.
Founded in 1885 with a market cap around $147 billion, this is a profitable, dividend-paying industrial giant that most investors completely overlook when scanning for edge AI exposure.
But they shouldn’t.
Honeywell builds the rugged devices that edge AI actually runs inside—on factory floors, oil rigs, and power plants. Their enterprise-grade cybersecurity is baked in at the hardware level. Their industrial networking infrastructure connects edge devices to each other and to central systems. An estimated 35% of Honeywell’s total revenue is already tied to edge AI applications.
The growth lever here is an upgrade cycle that’s already in motion. Honeywell doesn’t need to win new customers—it needs to modernize thousands of existing sites with AI-capable infrastructure. The installed base is massive, and the switching costs are high. For investors who want serious edge AI positioning backed by a balance sheet that has survived over a century, Honeywell is a cornerstone holding in this space.
2. Vertiv: The Thermal Backbone of Distributed AI
Vertiv has already been a rocket ship, growing from a $10 billion company to a market cap just over $100 billion.
The stock is up sharply over the past year, and its recent addition to the S&P 500 brings a fresh wave of index fund buying. The question investors are asking is whether the run has room to continue.
The case for more: the cloud data center boom involved a few hundred hyperscale facilities. Edge AI means millions of distributed nodes, each one requiring power management and thermal solutions—Vertiv’s core business.
And here’s the key insight most people miss. Edge devices are small, densely packed, and deployed in harsher environments than climate-controlled data center racks. The thermal challenge at the edge is just as demanding as it is in a centralized facility, but in a much harder setting.
As edge nodes spread from factories to hospitals to substations to vehicles, Vertiv’s addressable market scales directly with that buildout. The valuation isn’t cheap—the P/E ratio reflects high expectations—but the demand trajectory could justify it if edge deployments accelerate on the timeline bulls expect. Think of Vertiv as the picks-and-shovels play for the physical infrastructure of edge AI.
3. One Stop Systems: A Speculative Pure Play on Defense-Grade Edge AI
One Stop Systems is a different animal entirely. With a market cap around $250 million, this is a thinly traded micro-cap that is not for the faint of heart.
The company is not yet profitable on a net income basis. It is, however, as pure an edge AI play as exists in public markets.
One Stop Systems builds ruggedized, high-performance computing platforms engineered to survive conditions that would destroy conventional hardware—extreme heat, shock, vibration, and g-forces. Their sweet spot is defense. They hold contracts with the U.S. Navy’s Poseidon program, where lifetime revenue has already exceeded $65 million, and they’re building infrastructure for U.S. Army combat vehicles. A newer autonomous construction and mining equipment contract carries a $10 to $15 million five-year pipeline.
There is no cloud business here, no consumer division, no software subscription to fall back on. This is all edge, all the time. The company has guided for 20% to 25% revenue growth in 2026, and the path to profitability is visible if the contract pipeline continues expanding.
The risk is real—small-cap, speculative, volatile. But if defense-focused edge AI scales the way geopolitical trends suggest it could, this is the kind of name that can move a portfolio.
The Uncomfortable Phase Is Where the Returns Are
Edge AI is not a trade. It’s the deployment of artificial intelligence into the physical world, and it’s just getting started. The demand shock hasn’t arrived yet. The infrastructure is being built, the models are being trained, and deployments are in their earliest stages.
Investors who bought NVIDIA early made extraordinary returns, but they did it when the thesis felt uncomfortable. That’s roughly where edge AI sits today—early, volatile, and full of conviction-testing moments. The difference is that the demand here won’t be concentrated in a handful of hyperscalers. It will be distributed across every industry that operates in the physical world. Autonomous vehicles, smart manufacturing, AI-driven medicine, and defense robotics aren’t pausing for rate hikes or tariffs. The buildout is happening, and the companies enabling it are worth watching closely.
