Follow the Commitment: How CTA’s Brian Comiskey Tells a Technology Trend From Hype

Artificial intelligence will probably be present in every section of the CES 2027 show floor. Brian Comiskey is precise about why: “not because people are latching on to a buzzword, but because it is actually making a direct impact.”

Telling the buzzword from the impact is, in effect, his job. When I asked him in Paris how decision-makers can separate real signals from hype, I expected a short list of indicators. Instead, he described how his team actually works, which turned out to be more useful than any list.

Two jobs, one question

Comiskey, whose “fun title” is sometimes futurist, holds two roles at the Consumer Technology Association. The first is overseeing CTA’s partnership with Nasdaq on thematic stock indexes, which power exchange-traded funds tracking sectors such as AI, cybersecurity and cloud computing. That work forces a blunt question about every company in scope, whether already listed or likely to go public: “Are they actually an AI company or not?” The second role is tracking the trends that will shape CES and explaining them, which he describes as being “almost like an educator.”

Both roles rely on the same discipline. I would summarize it in four moves: read, narrow, rebuild and reconcile.

Read, narrow, rebuild, reconcile

The first move is reading. Comiskey follows what consultancies, think tanks and peers are watching, to see where a consensus is forming. The consensus does not tell him whether a trend is real. It tells him whether he is looking at “the right universe to start evaluating.”

The second move narrows the lens to companies. His team examines R&D budgets, patent portfolios, financial statements where they are public and investor decks where they are not. The work is collective by design. “I can’t be an expert on every single technology,” he said. The last input he mentioned is the most telling:

“Merger and acquisition activity shows you what conviction looks like in a space.”
Brian Comiskey, CTA

The logic is easy to state. Corporate language is cheap; commitment is not. Any company can describe a product as AI-powered at almost no cost, whereas redirecting R&D, building a patent position or acquiring a company puts capital and credibility at risk in ways that are hard to reverse. None of this proves that a technology will succeed, since entire sectors can make the same wrong bet. It does separate enthusiasm from decisions.

The third move widens the lens again. “You’ve got the general segment of the universe, then you’ve narrowed in on the companies,” he explained. “You take that step back again and you say, all right, what do those companies build up into larger trends?”

The fourth move is the one most trend forecasting skips: checking that bottom-up picture against the consensus he started from. “Then you do a reconciliation process of what you’ve read before,” he said. In my experience, the gap between the two is where the useful information sits. When the consensus runs ahead of the evidence, you are probably looking at hype. When the evidence runs ahead of the consensus, you may be looking at a signal most of the market has not yet recognized.

A test case: Mobileye

Comiskey offered an example without presenting it as one. Among the physical AI companies he is watching most closely is Mobileye, best known for camera-based driver assistance and for developing its own EyeQ chips. In June, Mobileye announced plans to go beyond supplying self-driving technology and to own and operate a robotaxi service, starting with a fleet of about 100 vehicles in a U.S. city in 2027. In February, it had already completed its acquisition of Mentee Robotics, a humanoid robotics company. For Comiskey, the robotaxi move is “the clearest example of physical AI leading into a mobility application,” rather than mobility companies moving into physical AI.

Read through his method, both moves are commitment signals. A component supplier that decides to run fleet operations and rider services accepts costs and risks it never carried before, and an acquisition buys capabilities that could not be built quickly enough in-house. Whether the robotaxi business succeeds is a separate question. The signal is that management has made decisions that are expensive to reverse.

Where the evidence hides

The framework CTA presented in Paris sorts technologies into three pillars: Foundation, Transformation and Exploration. Energy sits in the foundation, alongside semiconductors, cybersecurity, cloud, data, vertical AI and agentic AI. Comiskey’s explanation is blunt. When we think about token usage from AI, he said, “it really is actually a calculation of energy.”

He also stresses that the relationship runs in both directions. AI needs power, but AI is also becoming a way to balance grids and route energy to where it is needed. Activity around that constraint is already visible in his examples: nuclear power back on the table for always-on generation, more energy-efficient chip architectures, including neuromorphic designs, and better liquid cooling in data centers. The lesson for anyone reading signals is general. When a constraint starts to bind, a market forms around it.

Convergence is the other place where evidence hides. Comiskey calls the 2020s “the intelligence decade” and the 2030s “the quantum decade,” and argues that neither fully arrives without the other. As he points out, cloud platforms already let organizations simulate quantum computing environments to prepare for the disruption quantum will bring to security. Quantum, in turn, could accelerate the calculations needed to make nuclear fusion practical, and abundant clean energy is precisely what AI data centers will need. He sees this triangle of AI, quantum and energy as the convergence that is not discussed enough, alongside another underrated pairing: AI and health.

His personal enthusiasm at the moment is space technology, and it shows how he holds excitement and evidence together. He links part of the new momentum to SpaceX’s public listing and to Starlink, and he is drawn to the idea of data centers in orbit. In the same breath, he lists what still has to be solved: radiation shielding, capacity, transmitting the output back to Earth and even choosing the right orbit.

The listing illustrates his method as well as anything could. It drew enormous attention, and it also required SpaceX to publish financial statements that anyone can now examine. Hype and evidence arrived through the same door.

Hype and reality are not opposites

That is the core of Comiskey’s approach, and the reason it travels well beyond CES. A technology can be important and overhyped at the same time, as the internet, cloud computing and generative AI all were. The useful question for an executive is therefore not whether people are talking too much about a technology, but whether the evidence beneath the conversation is getting stronger. Five signals, read together, help answer it.

1.     Resources. Are R&D budgets and specialized hiring moving toward the technology?

2.     Conviction. Are companies acquiring capabilities they cannot build fast enough themselves?

3.     Protection. Are patent portfolios forming around it?

4.     Dependencies. Are the constraints around it, such as energy, components or regulation, starting to ease?

5.     Deployment. Is it moving from demonstrations into workflows where customers measure value?

None of these signals is decisive on its own. Together, they show whether an ecosystem is forming, and they suggest the best way to walk CES. Comiskey insists that innovation increasingly needs to “feel tangible,” and CES Foundry is built around demonstrations of how businesses and consumers can adopt the technologies on display. The question worth carrying through the halls is which products have evidence accumulating behind them, whatever they look like.

Why the capability has to live inside the company

When I asked what CEOs should do over the next twelve months, Comiskey’s first answer was curiosity. “Never stop learning about the technology yourself,” he said, which means going to the sources and experimenting. His second answer came from studying who handled earlier waves well, in cloud and in the internet. Outside expertise can help, and consultants “have good ideas.” His real recommendation, though, was structural:

“Build that expertise and have it be dedicated in your house if you can, because it keeps you honest.”
Brian Comiskey, CTA

In-house expertise, he added, is what allows a company to write an actual technology adoption strategy. As someone who advises companies for a living, I agree with him. External advice is most valuable when someone inside the organization is able to challenge it, and the method described here cannot be outsourced indefinitely.

Reading R&D allocations, deal activity and deployment evidence is a standing capability, not a one-off study. Predictions are fragile. The durable advantage is noticing, before competitors do, when the evidence has changed.

The future is always visible somewhere on a technology show floor. The skill is knowing which parts of it have started to become a market.

Part 1 of this conversation: what changes when AI moves from assistant to actor, and why Comiskey’s answer to the trust problem starts with zero trust.

About this conversation. Brian Comiskey is Vice President, Innovation & Trends at the Consumer Technology Association (CTA), owner and producer of CES. This interview was conducted by Stéphane Gervais in Paris on 16 September 2026, following CES Tech Trends x FDDay. Quotes have been lightly edited for clarity and length. CES 2027 takes place in Las Vegas from 6 to 9 January 2027.