Signals Inbox·August 22, 2026·Robotics

Gravis or Bedrock: who's ahead today?

Gravis Robotics is narrowly ahead today on platform breadth, mixed-fleet compatibility and distribution, but Bedrock Robotics has taken the sharper technical milestone with fully operator-less excavators on commercial jobsites.

Q1Why are Gravis Robotics and Bedrock Robotics suddenly being compared so much?

Gravis Robotics and Bedrock Robotics are now direct rivals because both want to turn contractors’ existing heavy equipment into autonomous machines and eventually control the software layer running mixed construction fleets.

The comparison became much harder to ignore lately. Bedrock has moved autonomous excavators onto commercial projects with nobody in the cab, while Gravis has just raised $200 million from SoftBank to accelerate a much broader international rollout. Both updates arrived within days of each other.

Their basic pitch is unusually similar. A contractor should not need to scrap a Caterpillar, Deere, Volvo or Develon fleet and buy proprietary robotic equipment. Gravis bolts its Rack, sensors and software onto existing machinery. Bedrock installs its Operator sensor-and-compute system onto excavators already used by contractors.

There is also a bigger ambition behind the retrofit kits. Gravis describes its software as an operating system for mixed fleets. Bedrock talks about moving from autonomous excavators toward coordinated fleets of excavators, dozers, loaders and trucks.

So we are really looking at two companies trying to own the intelligence layer between the contractor and the machine. Gravis has spent longer widening that layer across equipment, countries and workflows. Bedrock is pushing particularly hard on full autonomy in large U.S. earthmoving projects.

Q2Why is it still hard to say whether Gravis Robotics or Bedrock Robotics is ahead?

Gravis Robotics currently looks more mature as an autonomy platform, while Bedrock Robotics has stronger fresh evidence at one crucial technical milestone: commercial excavation with an empty cab.

That split creates two very different ways to score the race.

If we care about machine compatibility, product breadth, international deployment and routes to market, Gravis comes out ahead. Its technology has already been fitted across numerous equipment brands, works in several operating modes and has reached construction and quarry sites across multiple regions.

Bedrock is narrower today. Its production system is still centered on excavators, and Business Insider recently described its current autonomous deployment as three commercial sites.

Yet those three projects carry unusual weight. Removing the operator from the cab is the line between "automation helps an operator" and "this machine can perform part of the job as autonomous labor." Bedrock has now crossed that line publicly on active customer projects.

So Gravis gets more weight on breadth, while Bedrock gets more on the latest full-autonomy proof. The winner depends on which of those turns out to be harder to scale.

Q3Who has done more real construction work so far, Gravis Robotics or Bedrock Robotics?

Gravis Robotics has the broader real-world operating history, although Bedrock Robotics has just produced the stronger example of fully operator-less commercial excavation.

Gravis has already taken its system well beyond controlled demonstrations. Taylor Woodrow used a Gravis-equipped Yanmar excavator on an active Manchester Airport project, where the machine autonomously drove to work areas and dug georeferenced trenches. Holcim separately tested a 23-tonne Develon excavator fitted with Gravis technology at its Lee Moor quarry in the UK.

Those projects cover very different work. One is live civil construction; the other is repetitive quarry material handling. Gravis has also disclosed deployments in the U.S., continental Europe, Latin America and Asia.

Bedrock's earlier large deployment with Sundt was impressive but still supervised. On a 130-acre manufacturing site, Bedrock said its system moved more than 70,000 cubic yards of material. A human remained in the cab during that phase.

The newer Bedrock projects go further. Construction Dive recently reported fully autonomous excavators working at a Nevada water-treatment facility with Sundt and on a 1.2-million-cubic-yard civil sitework project with Zachry. Champion Site Prep is also using the system on a multi-million-cubic-yard Texas earthmoving job.

For breadth of field experience, we still give Gravis the lead. For the most advanced commercial autonomy being publicly demonstrated right now, Bedrock has caught up remarkably quickly.

Real-world evidence: Gravis vs. Bedrock

Q4Is Gravis Robotics already a broader product than Bedrock Robotics?

Gravis Robotics is clearly the broader product today because it gives contractors several levels of automation instead of asking every job to jump straight to full autonomy.

Gravis's Slate interface can be used inside the cab as Copilot, outside the machine to command autonomous tasks, or remotely to supervise and operate equipment. The operator can define a trench or earthmoving job from design data, use live 3D cut-and-fill guidance, let the machine perform repetitive digging and then intervene when the site becomes messy.

That sounds less dramatic than an empty excavator cab, but it fits how construction actually works these days. A crew might have hours of repetitive digging followed by twenty minutes of awkward work around utilities, trucks or workers. The same machine remains useful through both situations.

Bedrock's commercial message is more concentrated around full machine autonomy. Its Operator retrofit lets a site manager assign work, after which the excavator perceives the environment, plans movements and performs excavation autonomously. Bedrock ultimately wants that system to extend across the wider fleet.

Gravis already covers more of the transition between manual equipment and autonomous equipment. We think that gives it an easier path into contractors who want productivity now without redesigning the whole jobsite around autonomy.

Q5Which company works better across mixed construction fleets, Gravis Robotics or Bedrock Robotics?

Gravis Robotics has a clear lead in mixed-fleet compatibility, which is one of the strongest pieces of evidence in its favor today.

Gravis says the same core system has been installed on machinery from Caterpillar, Case, Develon, John Deere, JCB, Hitachi, Sumitomo, Yanmar and Volvo. Recent reporting puts its compatibility at more than a dozen makes and models, with machines ranging from roughly 10 tonnes upward.

That range is valuable in construction because large contractors rarely standardize everything around one manufacturer. Dealer coverage, rental availability, project requirements and existing fleet investments all influence what ends up on site.

Gravis has deliberately built around that fragmentation. Its learning-based control software adapts to the hydraulic behavior and physical characteristics of different machines rather than requiring the contractor to buy into one OEM ecosystem.

Bedrock has shown useful flexibility inside the excavator category. Its systems have been installed on excavators from roughly 20 to 80 tons, and the company says the retrofit can be installed in a day. Moving that architecture into other machine categories is now one of the big tests ahead.

Bedrock CTO Kevin Peterson has been quite open about the challenge, telling Business Insider that transferring the system into equipment such as dozers, loaders and dump trucks should work but will be hard.

Gravis has already done more of that compatibility work. This part of the comparison is not particularly close.

Q6Did Bedrock Robotics just take the technical lead with operator-less excavators?

As we saw above, Bedrock Robotics currently has the clearest public proof of fully autonomous excavators doing commercial earthwork without an operator in the cab.

The distinction is important because Bedrock's earlier autonomy program still had humans sitting inside the machines. The new deployments remove that safeguard. According to Construction Dive, a site manager defines the initial work and the equipped excavator can then execute the task while detecting nearby people and objects and stopping when necessary.

Business Insider spoke with University of Michigan construction professor Vineet Kamat, who called the current rollout a limited test. We agree with that characterization. Three sites prove that the technology has crossed an important threshold; they do not tell us yet how well the system works across hundreds of sites, different soil conditions or long periods without support engineers nearby.

Gravis has also demonstrated autonomous operation without continuous manual control, including machines that can work while operators supervise from outside the cab. What we lack is an independently described Gravis deployment matching Bedrock's newest projects closely enough to compare intervention rates, autonomous hours or unit economics.

So Bedrock deserves the technical edge here. We would need much more operating data before turning that edge into a claim that Bedrock has the better autonomy stack overall.

Q7Which company has better proof that autonomous construction actually improves productivity?

Gravis Robotics currently has the better measured productivity evidence, mainly because the Holcim quarry test gives us real output numbers instead of a general claim about autonomous efficiency.

At Holcim's Lee Moor quarry, a 23-tonne Develon excavator fitted with Gravis technology was asked to feed a screener at a target rate of 100 cycles per hour. Holcim reported an average of 133 cycles per hour during the two-day trial, with about 1,500 tonnes of material moved per day and 99% accuracy with minimal spillage.

That is unusually useful evidence for this young market. It tells us what the machine was asked to do, the target throughput and the actual result.

Gravis also advertises productivity improvements of up to 30% versus peak manual operation. We treat that number more cautiously because the comparison comes from the company and jobsite productivity depends heavily on soil, haul cycles, machine size and the human operator used as a baseline.

Bedrock has deliberately focused first on proving that its excavators can perform production work safely. Sundt did not claim a major cost reduction from the earlier supervised deployment, and Bedrock has not yet published a comparable cost-per-cubic-yard or productivity study for its new operator-less sites.

That leaves Bedrock with the more advanced autonomy milestone and Gravis with the better quantified productivity case.

Productivity evidence: what the field data actually shows

Q8Which company has better construction customers, Gravis Robotics or Bedrock Robotics?

Gravis Robotics has the more diverse industry relationships, while Bedrock Robotics has built an unusually strong group of large U.S. civil contractors for such a young company.

Bedrock lists Sundt, Champion Site Prep, Zachry Construction, Austin Bridge & Road, Maverick Constructors, Haydon Companies and Capitol Aggregates among the companies helping develop or use its technology.

These are useful customers for an autonomy startup because they actually manage large earthmoving fleets. Bedrock can learn from experienced operators, test on real projects and work directly on the tasks contractors would eventually pay to automate.

Gravis's relationships cover more parts of the construction equipment chain. Taylor Woodrow has tested the technology on live infrastructure work. HD Hyundai and Develon are working with Gravis around OEM integration. KIBAG has used a Develon-integrated system in Switzerland. Gravis also has relationships with equipment rental and building-material companies.

Gravis's mix looks stronger because selling autonomous equipment at scale will require more than convincing individual contractors. Someone still has to supply machines, install the technology, service the fleet, rent equipment to temporary users and support contractors locally.

Bedrock has excellent access to the people who use the machines. Gravis currently has more access to the wider system through which those machines reach customers.

Q9Does Gravis Robotics have a better distribution model than Bedrock Robotics?

Gravis Robotics currently has the stronger route to scale because it is already using rental companies, OEM relationships and dealer networks instead of relying only on direct startup-to-contractor deployments.

The Flannery relationship is particularly interesting. Gravis is working with the UK's largest operated plant-hire provider so contractors can rent excavators that already carry the Gravis Rack. A government-backed project worth roughly $8 million is also supporting the retrofit of excavator fleets with Flannery.

That removes a major adoption headache. A contractor can try autonomy through a familiar rental channel instead of buying a new machine, signing a long technology contract or becoming an early robotics customer.

Gravis has also said that Develon-integrated systems can be offered through existing OEM dealer networks. If that model works, local equipment businesses can eventually handle part of the distribution and support burden.

Bedrock's approach currently looks more hands-on. Its retrofit is designed for same-day installation, which is attractive, but the company still works closely with selected construction partners to deploy the technology and shape the product.

That direct approach probably helps Bedrock improve faster right now. For scaling into hundreds or thousands of customers, though, Gravis's model looks stronger because construction remains a highly local business built around equipment dealers, rental fleets and service relationships.

Q10Can we actually prove whether Gravis Robotics or Bedrock Robotics is growing faster?

We cannot prove which company is growing faster commercially because neither Gravis Robotics nor Bedrock Robotics publishes revenue, bookings, active-machine counts or comparable customer growth data.

The proxies point in different directions.

Gravis said late last year that its technology was live in seven countries across the UK, Europe, the U.S., Latin America and Asia. Since then, it has added a major U.S. presence, demonstrated Copilot with Hitachi at Conexpo and secured much more capital for its global rollout.

Bedrock's expansion has been compressed into a shorter period. The company only emerged publicly last year, yet its partner list quickly expanded beyond the initial group and its autonomy program progressed from supervised mass excavation into operator-less customer work.

Funding has accelerated for both companies too. Gravis went from a $23 million round late last year to a much larger SoftBank investment. Bedrock emerged from stealth with $80 million and later took its cumulative funding above $350 million.

Those are strong signs of momentum, but funding and geographic expansion are still proxies. Until either company reports commercial revenue, paid machine deployments or renewal data, headlines such as "Bedrock is growing faster" or "Gravis is commercially larger" would go beyond the evidence.

Q11Which company is closer to solving the hardest problem in construction autonomy?

Gravis Robotics is currently closer to solving the adoption problem, while Bedrock Robotics is pushing harder on the harder end-state problem of removing the operator entirely.

A construction robot competes with an extremely flexible human operator. The human can dig, notice a strange patch of soil, wave a truck forward, work around another crew and then switch tasks when the superintendent changes the plan.

That flexibility explains why autonomy usually works first on repetitive, clearly defined tasks.

Gravis has built around that reality. Its system lets the operator stay involved when necessary and hand repetitive work over to the machine when possible. The same interface can provide 3D guidance, autonomous digging, outside-the-cab supervision and remote operation.

Bedrock is betting that machine learning can eventually absorb much more of the human operator's flexibility. Its founders bring experience from Waymo, where autonomous vehicles gradually moved from tightly supervised operation into genuinely driverless service.

Right now, Gravis has chosen the easier commercial path because the product can create value before full autonomy works everywhere. Bedrock has chosen a more demanding technical path, but success there could produce a bigger labor substitution effect.

The next few quarters should tell us whether Bedrock can keep removing humans without sharply increasing support costs or limiting where the machines can work.

Q12Who has more money to keep competing, Gravis Robotics or Bedrock Robotics?

Bedrock Robotics still has the larger disclosed funding base, although Gravis Robotics now has enough capital that financing should no longer constrain the contest in the near term.

Bedrock has raised more than $350 million in total. Its latest $270 million Series B valued the company at $1.75 billion and brought in investors including CapitalG, Valor's Atreides AI Fund, NVentures, Eclipse and 8VC.

Gravis recently secured $200 million from SoftBank after raising $23 million late last year. The latest investment reportedly valued Gravis at about $1 billion post-money.

Using the disclosed figures, Bedrock has raised at least roughly 57% more capital than Gravis's two major publicly announced rounds combined. Bedrock also carries the higher valuation.

We would still avoid reading too much into the valuation gap. Neither company publishes the revenue data required to tell whether $1 billion or $1.75 billion represents an attractive multiple. Investors are mostly pricing future market control.

Practically, the point is simpler: both companies can now afford expensive field engineering, hardware development, machine purchases and long customer trials. Bedrock keeps the financial edge, but capital is no longer a convincing reason to expect Gravis to lose.

Q13Could Caterpillar, Built Robotics or another company make the Gravis-Bedrock race irrelevant?

Gravis Robotics and Bedrock Robotics are leading an interesting retrofit race, but neither company currently owns autonomous construction, and Caterpillar remains the scale benchmark that both have to worry about.

Caterpillar already operates hundreds of autonomous haul trucks through its mining business and has decades of experience deploying heavy equipment into harsh environments. It also owns something neither startup can recreate quickly: a global dealer and service network.

Built Robotics attacks the market differently. It has accumulated substantial operating experience by automating narrower, repetitive workflows, particularly around solar and large infrastructure projects. That specialization can beat a more general platform when the customer wants one task automated reliably rather than an entire fleet operating system.

AIM Intelligent Machines and Teleo add other approaches, from broad machine retrofits to supervised remote operation.

The opening for Gravis and Bedrock comes from mixed fleets. A contractor may own Caterpillar excavators, Deere machines, rented Volvos and specialized equipment from several other brands. An independent autonomy layer that works across all of them can solve a problem that an OEM-specific system cannot solve as cleanly.

Gravis currently looks better positioned for that horizontal role because its cross-brand compatibility is already broad. Bedrock could still take the market another way if its full-autonomy performance becomes good enough that contractors care more about capability than breadth.

Q14So who is winning today, Gravis Robotics or Bedrock Robotics?

Gravis Robotics is narrowly ahead today because its autonomy platform already works across more machines, more workflows and more routes to market, while Bedrock Robotics is currently ahead on the freshest proof of fully operator-less excavation.

Gravis gets the overall lead for three reasons.

The first is mixed-fleet compatibility. Construction fleets are fragmented, and Gravis already runs across a long list of major manufacturers. Bedrock still has to prove the same architecture beyond excavators.

The second is deployment flexibility. A contractor can use Gravis as operator assistance, task automation or remote supervision depending on the job. That makes the technology useful even when full autonomy reaches its limits.

The third is distribution. Gravis is already working through equipment rentals, OEM integrations and dealer channels. We think that will become increasingly important once the engineering problem starts giving way to the rollout problem.

Bedrock deserves more credit today than it did even a week ago. The operator-less commercial earthwork changes the comparison quite a lot. If those deployments multiply quickly, maintain high autonomous uptime and expand into dozers, loaders or trucks, Bedrock can overtake Gravis.

Gravis now has its own burden of proof. Breadth only counts if it turns into a large installed base. We want to see the company convert its new capital and distribution relationships into many more machines doing paid work, with customer-level data on productivity and repeat usage.

So our current verdict is Gravis by a narrow margin. Gravis is winning the broader platform and commercialization race; Bedrock is winning the newest full-autonomy milestone and is closing fast.

Gravis vs. Bedrock: current scorecard