Picture a job site. An excavator is moving earth, and nobody’s sitting in the cab. Since mid-August, a startup called Bedrock Robotics has had machines like this doing real, paid work.
On Monday, 26 October, Bedrock announced it had raised $600 million to put a lot more of these machines to work. If you’ve been following physical AI mostly through humanoid robots, autonomous excavators might be the more interesting story right now. Here’s why.
So what actually happened?
Bedrock Robotics, a San Francisco startup that started in 2024, announced a $600 million Series C round. That puts its value at $4.1 billion after the new money. CapitalG and Emergence Capital co-led the round, which Bloomberg first reported.
New investors include the US Innovative Technology Fund, Bessemer Venture Partners and Temasek. Older backers like 8VC and Nvidia came back too. All told, Bedrock has raised nearly $1 billion. That’s a lot of money for a company most people have never heard of.
The jump is steep. Back in February, Bedrock raised $270 million at a $1.75 billion valuation. So its value more than doubled in about eight months.
The founders know self-driving tech well. They’re former Waymo engineers, and CEO Boris Sofman previously led Waymo’s autonomous trucking work. That’s the same company now racing to grow its robotaxi fleet, so the team knows driverless tech from the inside.
How do autonomous excavators actually work?
Here’s the clever bit. Bedrock doesn’t build excavators. Instead, it adds a self-driving system to machines contractors already own.
Its kit, Bedrock Operator, targets excavators in the 20 to 80 ton range and similar heavy equipment, according to Crypto Briefing. The idea is simple: keep the expensive iron, upgrade the brain.
That retrofit approach matters. A contractor doesn’t have to bet on a whole new fleet. They can try autonomy on one machine, see if it works, then add more.
Under the hood, it’s still the same hydraulics and moving parts. If you’ve ever wondered how a robot actuator turns a command into motion, a big excavator is basically that idea at a much larger size. The new part is the software deciding what to dig, and when.
Longer term, Bedrock wants mixed fleets of machines that coordinate across a job site. Think of several different machines sharing one plan instead of each one waiting on a human. Honestly, that’s the part that’d change construction most, but it’s also the furthest away.
Is any of this doing real work yet?
Yes, and that’s why investors are paying attention. Since mid-August, Bedrock says its autonomous excavators have completed paid jobs on active commercial sites.
One early customer is Haydon, an employee-owned contractor in Tempe, Arizona. It ran fully autonomous earthwork at a remote site in Texas. Next, it plans to test trench work and larger machines, and to use Bedrock-equipped machines on data center construction.
“For us, the biggest win is a safer, more predictable job site,” said Katie Haydon, the company’s CEO and chair.
The customer list keeps growing, too. New names include Wagner Construction in Minnesota, Smith Trucking Company in Florida, Native Excavating in Colorado and Clarkson Construction in Missouri. Bedrock says customer volume has more than doubled since 2025.
That matters more than any demo video. Robots that look great on stage are easy to find. But robots a contractor will pay for, again and again, are much rarer.
Why investors are betting on autonomous excavators
Short answer: there aren’t enough people to run these machines. Bedrock’s announcement says employers need to fill about 1.7 million construction openings a year through 2035. It also says 87% of construction firms have openings for hourly craft roles, and 88% say those roles are as hard or harder to fill than a year ago.
So the pitch isn’t about replacing crews. It’s about keeping jobs moving when you can’t hire. Sofman put it this way: “We have proven that fully autonomous excavators can perform real work of value.”
Investors see something bigger, though. Emergence Capital’s Jake Saper said, “That’s not a software problem, it’s a physical one, and Bedrock is building for it.” CapitalG’s Derek Zanutto went further: “We cannot afford to take decades developing advanced robotics for this critical industry.”
Bedrock isn’t alone, either. Gravis Robotics raised $200 million in August, Eka Robotics raised $150 million last week, and Formic raised $50 million, according to The Next Web. Money is clearly flowing toward robots that do dirty, physical jobs.
Okay, so what could go wrong?
Plenty. A job site isn’t a highway. The ground shifts, the weather turns, and people walk around the machines all day. Dealroom lists variable terrain and weather, safe operation alongside human crews, and working across different equipment brands as key risks.
We’d add one more: trust. A contractor who has a bad day with a robot won’t care about the valuation. That’s probably why Bedrock is putting money into field operations and just hired Amy Jones Satrom, who’s worked at Amazon, Nuro and Uber, as its VP of operations.
Our honest take? Excavators look like a safer near-term bet than humanoids like Tesla’s Optimus. The task is narrower, the machine already exists, and customers are already paying for the work.
There’s also the money question. A $4.1 billion valuation assumes a lot of growth, so Bedrock still has to show it can support machines spread across many states.
What autonomous excavators could mean for the rest of us
If you’re not in construction, why care? Because almost everything you use sits on dug-up ground: roads, homes, data centers, power lines.
For emerging markets like Armenia, it’s an open question. Retrofits sound appealing, since contractors wouldn’t need to buy new machines. But lower labor costs, older equipment and tricky mountain terrain could slow things down. Then again, is a shortage of skilled operators really just an American problem?
In fact, the real test might be local. Contractors anywhere will want to see a machine work on a site they know before they rely on it. So the first local pilot, whenever it comes, will tell us far more than any funding round.
So here’s what we’re watching: will autonomous excavators stay a US story for years, or will retrofit kits travel fast enough to show up on job sites near you?