Carbon Robotics' 275 Can't Size Its Cut. Call Perception Engineers First
Carbon Robotics cut jobs without a headcount. Here's how to rebuild the cohort from public signals and which perception and field roles to contact first.
If you hire robotics or physical-AI talent, Carbon Robotics just put a group of people on the market that you can't size. The company confirmed to GeekWire on October 19, 2026 that it had cut jobs across several departments in the US and abroad, but it gave no total and no list of affected offices. The only number it offered ("currently employs 275") gets read as both the before and the after figure. So the useful question is which of these people are worth reaching first.
My answer: skip the generic ML engineers. Go for the perception engineers who have worked on farmland and the field-robotics engineers who kept machines running in Europe and Australia. Both pools are tiny, and outside Carbon, almost none of these people live in Seattle.
What did Carbon Robotics actually announce?
Carbon Robotics confirmed job cuts across several US and international departments. VP of marketing Brett Goodwin said the company needed to "realign costs due to broader economic conditions." There was no headcount, no severance terms, and no list of overseas offices.
Here's the background that matters for sourcing:
- The product. The LaserWeeder is a perception-heavy machine. It runs 24 NVIDIA GPUs and processes 4.7 million high-resolution images per hour.
- The backers. NVentures, NVIDIA's venture arm, led an $85M round in 2024. Total disclosed funding is about $177M in GeekWire-derived coverage. Other trackers say $229.9M (StartupIntros) or $276M (the open professional graph).
- The business. CEO Paul Mikesell, who co-founded Isilon (sold to EMC for $2.5B), told AgFunder the company has passed $100M in revenue and wants to IPO, with no date set. Carbon is No. 8 on the GeekWire 200 and won Next Tech Titan at the 2026 GeekWire Awards.
- The footprint. Startuply reports offices in the Netherlands, Australia, Brazil, France and Canada. Manufacturing is in Richland, WA, which is also where the AutoTractor's Remote Operations Control Center sits. BuiltIn has carried Field Robotics Engineer listings in Vienna.
This is an NVIDIA-backed company past $100M in revenue making cuts anyway, in a city where, by Seattle Times estimates, local tech firms have cut roughly 10,000 jobs since last year.
Why the 275 figure can't size the Carbon Robotics layoffs
The 275 figure can't size the cut because nobody agrees on what it refers to, and third-party headcounts for Carbon run from 208 to 306. What you can use is the spread between them, as long as you're clear about how speculative that is.
One write-up says Carbon employed 275 "before these reductions." Another says 275 is the count after. MyNorthwest sticks to the company's wording: it "currently employs" 275. Here's how the numbers line up:
| Source | Figure | Note |
|---|---|---|
| Company statement (via MyNorthwest) | 275 | Before or after cut is unclear |
| Open professional graph | 306 | Dated July 31, 2026 |
| Startuply | ~260 | Third-party estimate |
| BounceWatch | 237 | Third-party count |
| BuiltIn (older) | 208 total, 67 product/tech | ~32% technical |
| Refolk index, "Carbon Robotics" in headline | 195 | ~117 currently at Carbon (derived) |
Suppose the July figure of 306 is right and 275 is the post-cut number. Then about 31 roles went, roughly 10%. If 275 is the pre-cut number, the total could be anything. Treat the 10% as one possible reading. It isn't a finding.
The spread is the real warning: 98 people separate the lowest third-party count from the highest, about three times the 31-role reading. Size the cohort from profiles, not from the press release.
How do you rebuild the affected cohort from public signals?
Since the press release won't give you a list, build it from profile changes: new end dates, open-to-work flags, and stale headlines. Start with the people who still name Carbon in their headline and work out from there.
In Refolk's index of professional profiles, 195 people mention "Carbon Robotics" in their headline. In a 25-profile sample, 15 (60%) still list Carbon as their current employer, and only 10 (40%) are in the Seattle area. The other profiles are in Taiwan, Bristol, Zurich, Pune, Noida, Houston and Broomfield, CO. That spread is what you'd expect from a cut that reached teams abroad.
Here's how to rebuild it:
- Pull everyone tagged to Carbon, current and former, on LinkedIn and GitHub. A headline mention is the widest net.
- Flag recent end dates. Any October 2026 end date is a likely leaver. If a profile changes from "at Carbon Robotics" to "ex-Carbon Robotics," that counts too.
- Watch open-to-work flags and headline edits over the next few weeks. People rarely update on the day of the news.
- Split by function. Perception, embedded, autonomy, field and service, manufacturing (Richland), and sales. Each group goes to a different hiring manager.
- Split by geography. Seattle, Richland, and the international offices need separate outreach plans (more on that below).
Steps 1 through 4 are tedious when you do them by hand across three sources. That's the gap Refolk closes: you describe the cohort in plain English ("ex-Carbon Robotics perception engineers who left in the last month") and get back a ranked list drawn from GitHub, LinkedIn and the open web, without stitching boolean strings together.
Why perception engineers are the first call
Perception engineers come first because the US pool is tiny and none of it is concentrated in Washington state. A perception engineer Carbon let go in Seattle is coming out of a local market that has almost no one else like them.
A perception engineer builds the systems that turn camera, lidar or radar data into an understanding of the scene: what's a weed, what's a crop, where the row ends. At Carbon, that's the core of the product. Mason Fontes, a Carbon engineer, walked BuiltIn through the LaserWeeder's computer-vision stack, and it runs at a scale (4.7 million images per hour on 24 GPUs) that few teams anywhere work at.
Here's who employs those 95 people, according to Refolk's index:
- Zoox, Anduril, Aurora: 3 each
- Autonomous Solutions Inc. (Logan, UT), Zipline, Saronic: 2 each
US perception talent clusters around autonomous vehicles and defense. Ag tech robotics engineers who do perception are a subset of a subset: Autonomous Solutions Inc. is the closest ag-autonomy employer in the index, and it has two.
That tells you who you're up against. When a Carbon perception engineer goes on the market, Zoox, Aurora, Anduril and Saronic are the employers competing for the same person. If you're trying to hire robotics engineers in Seattle and can't match AV or defense pay, lead with what those companies can't offer: unstructured outdoor environments, crops that look different every week, and a product that ships to paying customers.
Field robotics engineers: the rarest profile, and the easiest to misread
Field robotics engineers are the rarest profile in this cohort, and they're easy to overlook because their title gets mixed up with general industrial service work. Only about 28% of US field and deployment titles are actually about robotics.
A field robotics engineer gets autonomous machines running, and keeps them running, at customer sites. The job sits between engineering and operations, and most sourcing searches miss it because they filter to one side or the other.
Refolk's index has 135 US profiles with field or deployment robotics titles. In a 25-profile sample, only 7 had a robotics-specific title: 6 "Robotics Field Engineer" and 1 "Field Robotics Engineer." The rest were general field service engineers at ABB, Ford, Walmart and similar companies. Scaled up, that's roughly 38 robotics-specific profiles nationwide. Bear Robotics and Locus Robotics are the top employers of robotics field engineers in the index.
| Pool (US) | Profiles | Robotics-specific |
|---|---|---|
| "Perception Engineer" titles | 95 | All of them |
| Field/deployment robotics titles | 135 | ~38 (derived, 28%) |
| General field service in that set | ~97 (derived) | No |
Carbon's field staff bring something even those ~38 usually don't: deployment experience on farms across several countries. The company has posted Vienna-based field roles, runs an EU footprint, and ships an iPad operator app in seven European languages. Someone who has kept a 24-GPU laser machine running through a Dutch planting season is very hard to find in the US market.
A field engineer who kept a LaserWeeder alive through a Dutch planting season is not a service tech. Don't search like one.
The practical trap is searching for "field engineer" and getting buried in industrial service profiles from ABB, Ford and Walmart. When I need this profile, I describe the work instead of the title: "field engineers who deployed autonomous agricultural or warehouse robots at customer sites."
Don't assume the cut was mostly engineers
The Carbon Robotics layoffs probably hit sales and field operations harder than core autonomy. Check that before you build a pipeline that assumes the cohort is mostly engineers.
Carbon has $100M+ in revenue and NVIDIA backing, and demand for physical AI didn't stop it from cutting. The likeliest explanation is the agricultural sales cycle. Per CNBC, Carbon's machines can cost over $1M, and some departments are tied to seasonal sales. Those are the teams that get trimmed first when a company "realigns costs," and Goodwin's phrasing (several departments, broader economic conditions) doesn't point only at engineering.
What that means for you:
- Engineering hiring managers: expect a smaller pool than the headline suggests. BuiltIn's older data put product and tech at 67 of 208, about 32%.
- Revenue leaders at ag tech or robotics companies: this may be the more useful cohort for you. Salespeople who've closed seven-figure equipment deals with growers are rare too.
- Ops leaders: the Richland manufacturing and remote-operations staff are a distinct group with hardware and fleet experience.
Search the abroad cohort separately
Carbon's international leavers will probably fit European and other non-US ag-robotics companies better than Seattle startups. Source them in a separate pass with separate messaging.
Refolk's index shows Carbon-tagged profiles in Zurich, Bristol, Pune and Noida, and the company's reported offices cover the Netherlands, Australia, Brazil, France and Canada, plus field listings in Vienna. Most of these people aren't going to move to Seattle.
How to work this pass:
- Group by country, not by function. Notice periods, employment law and pay benchmarks differ by country.
- Match to the local ag-robotics scene. A Netherlands-based field engineer fits EU agricultural automation companies far better than a US startup.
- Lead with languages and site experience. Carbon ships its operator app in seven European languages, so multilingual farm-deployment experience is the asset worth naming in your first message.
FAQ
How many people did Carbon Robotics lay off?
Carbon Robotics hasn't disclosed a number. It told GeekWire only that it "currently employs 275" people, and reports disagree on whether that's before or after the cuts. If a July 2026 third-party count of 306 is accurate and 275 is the post-cut figure, about 31 roles (roughly 10%) were cut, but that's a speculative reading, not a confirmed total.
Where can I hire robotics engineers in Seattle after the Carbon cuts?
Start with the Carbon cohort itself, though only 10 of a 25-profile sample of Carbon-tagged profiles in Refolk's index are in the Seattle area. Beyond Carbon, the local perception pool is thin: none of the top regions among the 95 US "Perception Engineer" profiles in the index are in Washington state, so expect to recruit people to relocate or to hire remote.
Which Carbon Robotics roles are hardest to replace?
Field robotics engineers and perception engineers. Only about 28% of US field and deployment titles are robotics-specific, roughly 38 profiles nationally, and US perception engineers cluster at Zoox, Aurora, Anduril and Saronic. Carbon staff who combine either skill with multi-country farm deployment are rarer still.
Try it on the search you came here for
Stop building boolean strings. Just describe the person.
Type one sentence. I plan the search, read GitHub, public LinkedIn and Crunchbase records, and the open web as it is right now, and hand back a ranked list with the reason next to every name.
01Describe them
One plain sentence. Role, city, stack, stage, whatever matters to you.
02I read the web live
GitHub, public LinkedIn and Crunchbase records, the open web. Not a database that went stale last quarter.
03You read the shortlist
Ranked, with the reasoning under every name. Open a profile, ask a follow-up, narrow it down.
- Staff backend engineers in NYC who shipped Rust in production
- Series A fintechs in SF under 50 people, growing headcount this year
- Maintainers of fast-growing Rust web frameworks on GitHub
- No boolean, no filters, no seat to buy. One box.
- Read at search time, so a profile updated yesterday counts today.
- Every step visible as it runs, every name with its reason.
500 free credits on sign-up. No card, no demo call. See real searches.