Physical AI helps construction giant predict weather delays and save millions

Brandon Barbello, co-founder of Archetype AI, argues that while artificial intelligence has rapidly transformed office work, the real breakthrough lies in “physical AI” systems that can interpret sensor data from construction sites, manufacturing facilities, and other physical environments. This represents a fundamental shift from text-based AI to systems that can process complex, multimodal data from the real world—potentially unlocking massive value in industries that have been largely untouched by the AI revolution.

The big picture: Physical industries like construction, logistics, and manufacturing represent a major portion of the global economy but have captured only a fraction of AI’s value, despite 89% of companies in these sectors planning to use AI.

Why traditional AI falls short: Unlike chatbots that rely on simple text inputs, physical environments generate “messy, multimodal sensor inputs” including video footage, machine telemetry, GPS data, weather stations, and vibration monitors.

• Each signal type traditionally required a separate software tool for interpretation.
• Data could only be analyzed one-by-one, meaning much valuable information went to waste.
• The physical world is fundamentally harder for machines to interpret than digital text.

Real-world application: Archetype AI partnered with KAJIMA, one of Japan’s oldest construction companies, on a multi-year canal widening project in flood-prone Niigata.

• The project involved thousands of workers, tens of millions in equipment, and hundreds of millions in materials.
• Weather delays could trigger cascading cost overruns adding up to millions over the project’s duration.
• Using Newton, Archetype’s foundation model for the physical world, they analyzed over two years of weather logs and footage from 27 cameras (nearly 12,000 videos).

How it works: The Newton AI platform created a unified interface that allowed project managers to see visual summaries of daily operations, flag deviations from work plans, and compare productivity across different weather conditions.

• Teams could better predict weather impacts and quickly reallocate resources.
• The system helped revise schedules proactively to stay on track.
• Managers gained unprecedented insight into construction progress and delay causes.

Knowledge preservation challenge: The construction industry faces both a labor shortage and a widening knowledge gap as veteran engineers and project managers retire.

• Years of planning intuition and problem-solving insight traditionally lived in the heads of experienced professionals.
• Newton can learn from veteran teams’ expertise: why delays happen, what conditions cause deviations, and what distinguishes normal idle periods from red flags.
• Over time, the platform becomes “institutional memory” that teams can build upon.

Future capabilities: Physical AI promises to go beyond preserving existing knowledge to generating new insights previously impossible.

• AI can analyze vast volumes of sensor data to discover hidden patterns.
• Teams could know not just that weather is coming, but which activities it will affect, which teams will be delayed, and which scheduling workarounds work best.
• Construction teams can move “beyond mere observation to predict and address challenges before they start.”

What he’s saying: “The world’s most important problems are physical, not digital, yet these real-world problems are often the hardest for technology to address,” Barbello explained.

• “AI will not just preserve knowledge from construction teams, but ultimately augment them with new insights that weren’t previously possible.”
• “Construction has always been a complex high-stakes endeavor. But now, thanks to physical AI, we have the tools to help teams build smarter and faster than ever before.”