What the billboards in Silicon Valley tell us about AI

Billboards aren't science, but they are one of Silicon Valley's most expensive datasets. What the roadside messages reveal about the next shift in AI.

Billboards aren't science, but they are one of Silicon Valley's most expensive datasets. What the roadside messages reveal about the next shift in AI.

I was recently back in Silicon Valley after about half a year. And precisely because of that, something I always find interesting there stood out: the billboards.

Billboards are, of course, not science. But in Silicon Valley, they are particularly interesting visual material. They hang literally in the middle of the ecosystem and are aimed at an audience of founders, engineers, investors, and decision-makers. Companies pay a lot of money to succinctly place a single message there. This is precisely why they clearly show which words, problems, and frames are currently considered relevant.

When you return after a few months, you can almost literally see how the conversation has shifted.

The Shift: From Larger Models to Fundamentals

And this time, that shift was striking. Where previously much was discussed about larger models, more intelligence, and new AI capabilities, you now see a much greater emphasis on the fundamentals of AI.

Safe AI. Reliable infrastructure. Robust code. Security. Control over data. Observability. Reproducibility. Understanding what systems actually do.

My impression based on these billboards: the conversation in Silicon Valley is shifting from "what can AI do?" to "how do we ensure we can truly build on it?"

European themes become global preconditions

That's interesting because precisely these themes have long received a lot of attention in Europe. Privacy, data sovereignty, control over infrastructure, and governance were sometimes seen primarily as European limitations. But as AI becomes a more serious part of business processes, they are increasingly becoming technical and strategic preconditions.

What we see at The Automation Group

We also see this at The Automation Group. Organisations are more frequently asking for European models, European infrastructure, and AI solutions that can run within their own cloud environment or on-premise.

No firm conclusion, but an interesting observation

So, no firm conclusion. But an interesting observation from perhaps one of Silicon Valley's most expensive and visible datasets: the roadside billboards.

And if these are a good proxy for where the market is mentally heading, then control over AI currently seems to be one of the most important themes.

Job van den Berg is an AI keynote speaker, tech entrepreneur, and author of five books on AI. He deploys agents himself weekly and gives 150+ keynotes per year on AI agents and agentic commerce.

Counting Cybertrucks in the back seat is not representative research. But it is a perfect example of how statistics, logical thinking, and AI come together: observe, formulate hypotheses, and, most importantly, try to debunk them.

Hallucinations can be controlled. Much trickier is a line of reasoning that sounds logical, is partly supported by research, yet goes further than the evidence permits. On generational differences, statistics, and methodological caution as an AI skill.

The promise of AI agents is to eliminate friction, putting ease first. But precisely in that friction, the contact that forges our social capital always arose. An essay on connection, technology, and the default setting.

Answers are more accessible than ever. The real skill will not be formulating the question, but opening the hatch under the bonnet.