CES 2026: Boring AI Is Winning, Fast Money Is Circling, and Most Leaders Aren't Ready

1. AI is winning in the back office. 

IBM's talk "AI of ROI" was the most important session I attended at CES—and the most boring. No flashy demos. No agentic breakthroughs. Just chatbots. Ask IT. Ask HR. Order a laptop. Onboard an employee. Submit your vacation. Mundane stuff I've done dozens of times for my clients over the last 16 years. And it saved IBM $4.5 billion. That's the point. The AI that's actually winning isn't on stage—it's in the back office, behind the firewall, where nobody sees it fail. It's a known use case, widely available internal data, a safe way to iterate, and an immediately visible ROI. It is also a low-risk way to train your entire organization on AI, and a smart way to make the transition to AI-driven work inevitable without betting the company on a demo that might blow up in your face.

2. Most leaders aren't ready. 

There's a massive disconnect between what corporate leaders think AI can do and where the industry is actually marching. The change is too fast. The technology is too strange—probabilistic instead of deterministic means a completely different set of constraints, strategies, and execution paths. Most leaders I talk to aren't skeptical. They're terrified. They're still learning basic prompting strategies, still trying to wrap their heads around what's even possible. They're not ready to build with this tool yet, much less utilize all of its capabilities. But the world isn't waiting. Competitors are moving fast. That's what terrifies them most—not the technology itself, but the growing suspicion that everyone else has figured it out and they haven't. For many traditional companies, nobody's driving the ship.

3. McKinsey is throwing college kids at your toughest problems. 

McKinsey is hiring college kids and sending them straight into client engagements. I suspect the other big consultancies are doing the same. These are kids fresh out of school. They've never seen a project fail. They don't know how hard it is to actually ship something that makes money. McKinsey admitted as much—they're growing like crazy, and there's no end in sight, so they're throwing bodies at the problem. But bodies aren't experts. These kids don't bring 16 years of experience. They don't bring 34 real products shipped. They bring PowerPoints. And the consultancy model isn't designed to make you independent—it's designed to create dependency, milk the engagement, and deliver strategy instead of working on delivering real working AI applications. That won't save you. The only way forward in AI is through real use cases, real data, and rapid iteration. AI models are commoditized. The technology isn't the differentiator anymore. But make no mistake—AI is a cruise missile aimed at the heart of your business. The question is whether you're steering it or standing in front of it.

4. Chinese-American money is writing its own rules. 

Chinese VCs are pouring billions into American AI companies—and they're not fooling around. These aren't exploratory bets. They're strategic plays backed by something no American VC can offer: Chinese manufacturing capacity. An idea can become a physical machine in weeks, not months. When I say I can ship a beta in 4-6 weeks, I mean software. These new conglomerations are doing the same for hardware. They're funding the best ideas worldwide, pairing them with infrastructure to make physical products at impossible speed. And they're not building one-off products—they're building platforms. Robotic lawnmowers. Delivery robots. Autonomous floor cleaners. Wearables. Self-driving vehicles. AI is crossing the digital-physical barrier faster than ever. Coding used to be the bottleneck—vibe coding solved that. Now, Chinese manufacturing is doing the same thing to physical products. Korea is likewise deep in the AI know-how game. Japan, strangely, is largely absent. But here's the real lesson: money knows no borders. It never has. The results these international conglomerations are producing are considerably better than any single country achieves alone. That's not politics—it's physics. And the balance is already shifting.

5. AI VCs are weaponizing distressed companies. 

Here's a brand new trend that should terrify every traditional company: VCs are buying distressed businesses—hospitals, legacy enterprises—and turning them into AI laboratories. They pair these acquisitions with their portfolio startups. The startups have solutions but no buyers because traditional companies won't touch unproven AI. So the VCs eliminated the sales cycle entirely. The acquired companies become captive use cases—real data, real workflows, real problems to solve. The startups iterate fast, validate with internal customers, and once it works? Scale it a thousandfold. This is a completely new playbook, and it's moving faster than anything I've seen in 16 years. Billions in fast money are pouring into these "AI incubator factories." And here's what keeps me up at night: they're using my methodology: Real use cases. Real data. Rapid vibe coding of mini AI Applications. Rapid deployment. Standard models. Immediate customer testing. Same-day iterations. The exact recipe I spent 16 years developing—they're running it at industrial scale. The early results are already showing. If you're not doing the same thing in-house, this isn't a "competitive disadvantage." It's an extinction event. Picture an army with diesel engines and computers running circles around cavemen with stone hammers. That's the AI gap that's opening right now.

The way forward is clear.

What continues to work today is the same thing that's worked across my 16 years and 34 shipped products: real use cases, a thin slice of real data, shipped quickly as a working AI application, and put in front of customers for rapid iteration. I call it Snowball Sprint. VCs are increasingly calling it "business as usual".

Using this methodology, your team should be able to ship a small AI feature beta in about six weeks. And you should be building this capability in-house. McKinsey's kids can't do that for you—their model is specifically designed to keep you dependent, not to teach you how to fish. I teach your whole team how to do this themselves after I'm gone, using a real project—you get to ship the proof that it works and keep the training that sticks.

Most companies are still overbuilding—spending 6-9 months and millions of dollars scaling infrastructure for the wrong use case. That's why 85% of AI projects fail. You can be the exception. Winner or tourist—which one are you?

If you're ready to build this capability in-house, let's talk: https://calendly.com/snowballsprint/30min

Greg

P.S. Full report: VC Firms Buying Distressed Companies While Betting on AI. https://claude.ai/public/artifacts/04882a1d-5cc8-4b5e-a65f-281b4707103c

Happy to discuss what that means to your AI strategy: https://calendly.com/snowballsprint/30min