Construction AI: Buildots opens its data vault to the whole industry, and the US government starts rationing the strongest cyber AI

So here's the genuinely interesting move of the week. On 25 June 2026 Buildots launched the Buildots Intelligence Lab, which it bills as construction's first AI-powered research hub. The premise is simple. Buildots has spent years instrumenting live sites with 360-degree hardhat cameras that log every delivery, every placement of work, and measure it against the programme. That generates an enormous amount of data about how building work actually proceeds. The Lab takes all of it, anonymises and aggregates it across projects worldwide, and publishes it back as free, real-world benchmarks. It's live now at buildots.com/lab.

What the Lab is really trying to fix is a problem every PM knows in their bones. We don't have a shared, honest picture of how work runs. Co-founder and chief exec Roy Danon put it well: the industry has always lacked a source of macro-level truth, and he thinks that absence is a core reason productivity has stagnated. I think he's onto something. When every firm benchmarks a programme against its own past jobs and its own optimism, nobody can tell whether a six-week slip on second-fix is normal or a disaster. A neutral baseline would change those conversations. That's what it's about.

But, and this is the bit an editor might cut, a free benchmark is also a very clever competitive position. The pool of data belongs to Buildots, not to the industry, and the projects feeding it are by definition the heavily-digitised ones running Buildots in the first place. So the macro truth it produces is the truth of the instrumented top end, not your typical regional contractor with a part-time document controller and a spreadsheet. The comparison only goes so far, but it's a bit like judging the national average pace from a field of marathon runners. Useful, real, and not actually representative of the road outside. Take the numbers as a read on where the leaders sit. For the PM about to defend a programme to a client, that distinction is the whole game.

The procurement filter: Before anyone opts a live project into a shared benchmark, free or not, read the data-sharing terms and decide on purpose. The macro picture is worth having. Just know your project data is what's funding it.

On 26 June 2026 the US Commerce Secretary, Howard Lutnick, wrote to Anthropic to say it could put Mythos 5, its most capable cybersecurity model, back into the field. Not generally. The letter cleared it for roughly 100 named US organisations that operate and defend critical infrastructure, the likes of Cisco and JPMorgan, after Anthropic had pulled access to the model on 12 June over the risk that a tool this good at finding software flaws is also a tool this good at exploiting them. The model that can autonomously hunt vulnerabilities across major operating systems is now, in effect, a controlled substance, prescribed by government to vetted defenders.

This lands two days after yesterday's brief covered OpenAI gating its own defensive model to verified defenders, and I want to be honest that the theme rhymes. But the difference is the whole point. Yesterday a vendor decided who got its model. Here it's a government deciding, by name, who holds the strongest defensive capability in the country. What that shift means is that frontier defensive AI is moving from a thing you buy to a thing you're granted, and the granting follows the map of critical national infrastructure. For a sector that builds that infrastructure, that's not abstract.

Because look at what's actually on the UK's critical list right now. Data centres, which were formally designated Critical National Infrastructure in 2024. The grid connections and substations feeding them. Water. The exact pipeline your sector is racing to pour. I'm not going to pretend a Lutnick letter to a US company changes a UK site this quarter, it doesn't, and UK firms sit outside that hundred-name list entirely. But the precedent is the signal. The defensive tooling protecting the assets you build is becoming something allocated by the state, and on a CNI job the people holding it are your client and their security partners, not the main contractor. That's worth a line in the risk register, and a question at the next CNI kick-off.

For your board pack: On any data-centre, energy or water scheme, find out early who owns the cyber-defence posture for the asset and the project data, because the strongest tooling is now rationed and you won't be the one holding it. Better to know that on day one than after an incident.

Put the two stories side by side and they're the same story told from two ends. Buildots wants your project data because, aggregated, it becomes a benchmark worth more than any single job. The Mythos decision exists because the data and systems running critical infrastructure are valuable enough that a government now controls the AI defending them. One side wants to learn from your data, the other side is bracing to defend it. Either way, the thing under everyone's attention is the data your projects generate.

And the UK sector is generating more of it than it used to. The NBS Digital Construction Report found more than two in five architecture professionals now using AI in their daily work, up from fewer than one in ten back in 2020, and eight in ten of all respondents collaborating in the cloud on models and specifications. That's a lot of project information leaving the laptop and landing in platforms you don't own. None of that's a reason to slow down. It is a reason to know, for each system you plug into, what you're contributing, where it sits, and who's responsible for guarding it.

A practical step: Pick one project this month and write down, in plain English, where its data actually lives, which vendors hold it, and who owns the security of each. If you can't fill that page in, that's the finding, and it's a better use of an afternoon than another AI pilot.