Reservoir Brings Its Farm Robot Proving Ground to Washington Orchards
A farm robot that works in a California vegetable field still has to prove itself in a Washington apple orchard, where the trees are trained into narrow walls, the fruit bruises easily and the season swings from frost to heat. Reservoir, the agricultural robotics hub that got its start in California's Salinas Valley, now wants to give those machines that chance.
On Monday the company announced Reservoir Farms Washington, an on-farm testing site in the Yakima Valley set to open in spring 2027. It is Reservoir's first site in the Pacific Northwest, and it goes straight at the crops robot makers find hardest: tree fruit, blueberries, wine grapes and hops.

What Reservoir Is Building in Yakima
Reservoir calls its farms technical communities, places where startups build, test and refine machines on working land next to researchers and growers. The Washington site is being developed with the Washington Tree Fruit Research Commission, the Tri-City Development Council and the North American Blueberry Council, and Reservoir is exploring a partnership with Washington State University's Prosser Irrigated Agriculture Research and Extension Center. The Tri-City Development Council, known as TRIDEC, is providing financial support.
Four startups are already on board. Bonsai Robotics and Sky Sense, which already build technology for tree fruit, will join as members, along with early-stage companies Cordon and Rootline. Reservoir says it has more than 30 startups enrolled across its locations, and associate-tier startups will get free farm access in Washington.
The first work will focus on tree fruit and blueberries, with testing in wine grapes and hops to follow and row crops planned later. "When you put builders on real farms with growers who will tell them the truth," founder and CEO Danny Bernstein said in the announcement, "rugged physical AI goes from prototype to production much faster."
Why Washington's Crops Are Hard to Automate
Washington is the country's leading producer of apples, pears and sweet cherries. Citing 2025 USDA figures, Reservoir says the state grew about two-thirds of America's commercial apples, a crop worth nearly $1.9 billion, along with about 70% of U.S. sweet cherries and 44% of pears. It also led the nation in cultivated blueberries, with 192.4 million pounds, about a quarter of the U.S. crop. The Yakima Valley alone hosts more than $3 billion in high-value crops.
Those are precisely the crops that resist automation. Apples and cherries hang at different heights inside dense canopies, blueberries are small and delicate, and every variety ripens on its own schedule. Sweet cherries, for example, must be picked within a short window of a few weeks, and a machine that bruises the fruit is worse than no machine at all. "Tree fruit, blueberries, wine grapes and hops are some of the hardest crops in the world to automate," Bernstein said, "and labor is now the biggest cost Washington growers face."

The Labor Math Behind Farm Robots
The labor numbers show why growers are paying attention. According to Washington State University, labor makes up 50% to 57% of orchard variable costs for Washington apples. The number of H-2A seasonal worker positions certified in Washington nearly doubled from 2017 to 2025, to more than 34,500, while the number of Washington farms with orchards fell 12% over a similar period.
Growers, though, are wary of machines designed somewhere else. "Our growers are spending more on inputs every season, and they will not adopt technology that was not developed or tested locally," said Ines Hanrahan, executive director of the Washington Tree Fruit Research Commission. "A machine that hasn't been through a Washington growing season isn't ready for one."
Her commission wants machines shaped by local conditions instead. "Building here, with growers involved from day one, is how we make sure our industry is robot ready, and the robots are ready for our industry," Hanrahan said. Local business leaders see a regional payoff too. "Reservoir gives agricultural technology companies a reason to build here," said Karl Dye, president and CEO of TRIDEC.
A Salinas Idea Goes North
Reservoir was founded in 2024 and opened its first farm in the Salinas Valley, about an hour south of San Jose, with innovation barns and 24 acres of commercial test fields. It has since added sites in California's Wine Country and Central Valley and in Yuma, Arizona, which began as a pilot with the University of Arizona. Washington is its second expansion outside California.
The model pairs a hands-on testing platform with a venture fund, Reservoir VC, and it has drawn backing from established players. John Deere committed $10 million over three years in August, and Western Growers announced a $1.5 million, three-year partnership earlier this year.
For Silicon Valley, the expansion is a reminder that physical AI needs safe places to fail. Bonsai Robotics, one of the first Washington members, is based in San Jose and builds vision-guided autonomous machines for orchards. Software can be tested in the cloud, but a picking robot has to meet real trees, mud, dust and weather before anyone will buy it.
Will Robots Be Ready for the Next Harvest?
The outlook is promising. If startups can prove their machines over a full Washington season, growers get evidence instead of sales pitches, and investors get a clearer signal about which farm robots will actually sell. A machine that handles Washington's frost, heat, dust and steep orchard rows is more likely to work in other tree fruit regions around the world. Shared test farms also lower costs for young companies that cannot afford their own orchards. Naidu Rayapati, director of WSU's Prosser center, said that bringing startups together with faculty "could give growers valuable insights within a single growing season and rigorous validation over multiple seasons."
There are real concerns as well. Farm robots remain expensive, and many family orchards run on thin margins. Machines that work on new high-density trellised blocks may struggle in older orchards with larger, irregular trees. Reliability matters more on a farm than in a demo, because a robot that breaks down in the middle of cherry harvest can cost a grower a crop that ripens in days. Data ownership is another open question when startups collect detailed information from growers' fields.
There is a broader debate about jobs. Farmworker advocates worry that automation will cut seasonal work that many families depend on, while supporters argue that robots will fill jobs growers already struggle to staff and create better paid technical roles. Dye framed it as an opportunity: "the next generation of ag jobs will be in designing, running and servicing the technology." How quickly that happens, and who gets trained for those jobs, is still uncertain.
In the end, it comes down to scale. A testing site can show that a machine works, but growers will need service networks, financing and parts available during harvest before they commit. Several farm robotics startups have struggled to move from pilots to fleets, and the industry still has few products running at commercial scale in tree fruit.
The signals to watch are specific. Which startups sign up before the spring 2027 opening, and will WSU formalize its partnership? Will any robot that tests in Yakima reach commercial sales in Washington orchards within a few seasons? Will labor costs keep rising? And will other agricultural states copy the shared test farm model?
From Bees to Bots
At Sparknify we keep coming back to the question of how technology can help feed people. Earlier today we looked at A Century of Records Shows Most of Europe's Wild Bees Losing Ground, a reminder that California's farms depend on a natural workforce that is also under strain. Reservoir's bet is that the other workforce, the human one, will need help from machines that have earned their place in the field.
The robots that win in Washington will not be the cleverest in the lab; they will be the ones that survive a full season in the orchard.
Subscribe to Sparknify for Silicon Valley tech news and the people behind it.













Comments