Artificial intelligence startups are raising unprecedented amounts of money—we're talking billions—to build sophisticated "brains" for robots destined for oil rigs, construction sites, and everywhere in between. If you thought AI job disruption was just a white-collar problem, think again.
What's Actually Happening?
The concept is straightforward but ambitious: create AI software that understands physics and real-world conditions so robots can adapt to changing environments on the fly. These aren't necessarily humanoid robots—functionality matters more than form. If a robot has the physical capability to perform a task, this AI could give it the flexible knowledge to actually do it.
We're talking about plumbing, electrical work, welding, roofing, auto repair, cooking—you name it. Imagine C-3PO and R2-D2, minus the personality quirks.
The Technical Debate
Interestingly, there's no consensus yet on the best approach to robotics AI:
The Data-Hungry Approach: Big Tech companies and startups are collecting massive amounts of real-world data to train their AI models.
The World Models Approach: Some researchers are using "world models" trained on simulated physical data. These are more cost-effective since they rely on fundamental principles like gravity. This method has gained traction thanks to advocates like Yann LeCun, Meta's former chief AI scientist who recently launched AMI Labs.
Following the Money
The investment dollars tell a compelling story:
Waabi raised up to $1 billion last week—potentially the largest funding round ever for a Canadian startup. The Toronto-based company is initially targeting robo-taxis and self-driving trucks. "It's obvious that the physical AI moment is here," says founder and CEO Raquel Urtasun. "Autonomy is the first application where scale is going to happen."
Skild AI in Pittsburgh recently secured approximately $1.4 billion at a $14 billion valuation. Their ambitious tagline says it all: "Any robot. Any task. One brain."
FieldAI raised nearly $400 million last month to focus on industries that are "dirty, dull, or dangerous"—think energy and logistics. Their software could enable robots to help construct data centers, creating a somewhat surreal scenario where AI builds the infrastructure for more AI, potentially leaving humans watching from the sidelines.
The Uncertain Future
Predicting how many blue-collar jobs might become obsolete—or over what timeframe—is genuinely impossible right now. Even when AI-powered robots can outperform humans, the costs of hardware and implementation might outweigh efficiency gains.
At least for now.
Two Perspectives
The optimists argue we won't see net job losses, regardless of job type, because new technologies historically create new labor demands.
The skeptics warn that history isn't always a reliable guide. AI represents a more fundamental shift than any previous technological change we've experienced.
The debate continues, but one thing is clear: the impact of AI isn't limited to office workers. The physical world is next, and blue-collar industries should be paying very close attention.