Happenings: I will be joining a panel on AI in agriculture hosted by Co-Bank on August 7th. You can register for the free event here. AgTech Alchemy is hosting an event in Chicago called “The Soy-Bean” on August 14th. You can register for the event here.
Was the dress blue and black, or white and gold? Should you buy or lease a car? Should you rescue or get a dog from a breeder? For many such questions, the answer is “it depends on your unique situation” (I do have to admit I am a strong proponent of rescuing dogs and not fond of breeders.)
A similar version of the debate is prevalent in AgTech.
As an AgTech company should you go direct-to-farmer or through a distribution channel to get your product to market?
The expected and frustrating answer is “it depends” as there is a lot of nuance.
For example, Patrick Honcoop talked about the importance of distribution through channel partners for a startup to scale on his viral LinkedIn post last week. He highlighted a few reasons why most startups who tried to go direct-to-farmer struggled to move forward.
❎ High costs to build and maintain a dedicated salesforce
❎ Lack of local knowledge and relationships
❎ Farmers aren’t waiting for another company knocking on their door
❎ It takes years to become a trusted partner
To his credit, Patrick did lay out a few exceptions to his general heuristic.
- You’re selling a high-ticket, complex solution with enough margin to justify building a network
- You have a very low-cost software product with a lean team and it solves a clear pain point (though very few Digital Ag companies have cracked the “online only" model)
- You can build your business in a small, dense territory that eliminates the proximity challenge
- You’re selling to enterprise customers, not directly to farmers
This advice is not very different from what I give to young startups.
Patrick’s post went viral by AgTech LinkedIn standards, with some absolute opinions expressed, which I am not a fan of. It depends on many factors, including who is your target customer and user.
Matthew Pryor of Tenacious Ventures had provided more nuance to this question a few weeks ago for a specific scenario.
If the startup wants to get acquired by an incumbent, then they should not worry about access to distribution but focus on their technical risk removal, product market fit, and concentrated evidence.
The most successful Agtech acquisitions happen when startups recognize their core strengths and focus relentlessly on building leverage that incumbents can't easily replicate. That's rarely distribution—it's usually deep technical capabilities, validated product-market fit in focused segments, or proven solutions to problems incumbents haven't solved.
I do agree with Matthew on his perspective.
Let us take a specific case of robotics in agriculture, and see how one might think about your Go-to-Market strategy and business model differently. I am picking robotics in agriculture as it is a hot topic right now with recent financial activity from Tric Robotics, 4AG Robotics, acquisition of farm-ng by Bonsai Robotics, and a switch to a service based model by Farmwise (acquired by Taylor Farms)
Go-to-market for robotics in agriculture
There are a few different archetypes when it comes to go-to-market motion for a robotics company in agriculture.
Direct‑to‑farmer (DTF) go‑to‑market
The robotics company sells or rents the machine (or the service that machine performs) straight to the grower, owning the full commercial relationship: marketing, demos, financing, deployment, training and after‑sales support.
There are clearly some advantages and disadvantages to this approach.
Advantages
- Pricing control and full margin capture as you are not sharing it with a channel partner, there are no back-end rebates, and you can preserve cash per unit sold.
- From a product velocity perspective, there is a tight feedback loop for rapid iteration as your team of engineers, product managers, and marketers can hear problems first-hand and push fixes quickly. A direct-to-farmer model is especially important if the robotic technology is still evolving rapidly and needs frequent updates. For example, Palantir has made the term “forward-deployed-engineer” relevant and sexy with their approach.
- The ag robotics company can have direct visibility into agronomic and farm level data, with the optionality to enable AI driven workflows on top in the future.
- It preserves strategic optionality for the robotics company and the freedom to experiment and adopt different business models like RaaS subscriptions or outcome‑based contracts (per acre, per lb.)
Disadvantages
- The biggest drawback to a direct-to-farmer model is a very high cost of customer acquisition, which could be a huge challenge for a small sales organization and put a significant financial burden on a startup's working capital. This is one of the main reasons I advise startups to think really really really hard before deciding to go direct-to-farmer.
- Asking farmers to put trust into a no-name entity can be challenging.
- You have to build the service and parts infrastructure from scratch. This can be especially challenging, when service and support are critical for adoption and repeat sales, especially for ag robotics.
- If your desire is to scale geographically, it will be slow and expensive.
This video put together by Plug and Play is a good primer on AgTech Go-To-Market strategies.
AgTech Go-To-Market Strategy from Plug and Play
What contexts are the best fit for direct-to-farmer for ag robotics?
- Early‑stage products where the technology is still evolving weekly and close farmer feedback drives reliability.
- High‑value specialty crops in dense clusters (e.g., Salinas lettuce, Central Valley orchards) where a small service team can cover many paying acres.
- Regions lacking strong independent dealer networks (parts of Latin America, Southeast Asia).
Tim Bucher, CEO of Agtonomy was quite emphatic about his view:
Farmers don’t have the time to waste on AgTech startups to manufacture machines that take years to deliver and have yet to earn grower’s trust. A more common-sense approach would be to partner directly with established OEMs and IEMs (integrated equipment manufacturers). These collaborations leverage trusted brands and existing dealer networks, ensuring farmers receive reliable, autonomous solutions that seamlessly integrate into their operations and drive profitability.
Distribution / channel‑partner model
The robotics firm signs multi‑year agreements with equipment dealers, custom‑service contractors, input retailers or OEMs, who resell, finance and service the machines.
Advantages
- Quick market reach and imputed trust from the distribution channel.
- Lower customer acquisition cost and service overhead.
- Financing and warranty structure support.
- Unlock OEM co-development money.
- Easier and faster expansion into new regions.
Disadvantages
- Dealer discounts, floor specials and other schemes put margin and price pressure.
- The feedback loops on product performance are muted, slow, and can be opaque. It also results in loss of brand and narrative control.
- Effort required for dealer training, and retraining, which can be challenging, especially for complex ag robotics systems.
- There is a significant risk of mis-alignment of incentives, resulting in improper positioning (or lack of) for your product within the channel’s product portfolio.
What contexts are the best fit for channel distribution for ag robotics?
- Your hardware is mature and reliable, which does not require constant updates and upgrades, and has a long useful life (for example, sprayers, tractors etc.)
- Broad‑acre or global crops where farmer bases are dispersed and uptime expectations mirror mainstream equipment.
- Capital‑intensive scale‑up phases where venture funding alone cannot finance inventories; dealers provide floor‑plan credit and residual‑value guarantees.
- Markets where dealer loyalty is entrenched like with row‑crop producers in the U.S. Corn Belt, Prairie Canada, and Western Europe.
Startups can follow a hybrid sequencing with early years spent on direct pilots with a tight geographic and crop focus with the ability to iterate rapidly and prove agronomic API. In subsequent years, you could pivot to something like a regional hub-and-spoke robotics-as-a-service fleet, followed by a dealer engagement or OEM licensing of your product.
Next let us spend a few minutes to look at different business models for ag robotics. This is not meant to be a comprehensive list but provides the nuance which is often missing in online discussions and on social media.
Business models for robotics in agriculture
Direct sales of ag robots
This involves an up-front sale of the robot with spares and service contracts. This works best for large 6 or 7 figure ag robotics purchases (for example, Carbon Robotics LaserWeeder with a price point about $ 1.2 to $ 1.4 million).
In this model the robotics vendor does get immediate cash flow, but they do have lumpy revenue and have to fund inventory, provide customer support and manage product updates.
If your customers have a huge discretionary budget and can purchase expensive equipment, or if your market is spread across thousands of miles, direct sales can potentially make sense.
Farmers have a lower risk of equipment availability (as they own it), and they are eligible for depreciation, tax-breaks etc. The farmer does have to clear the high cap-ex hurdle, and bears the technology obsolescence risk.
Due to this, it works best with large well-capitalized farms, where the product life is expected to be at least 7 to 10 years, and there is a clear resale value, with access to a trust parts network for service.
Resale values for agriculture equipment have depressed in the last few years, creating downward pressure on first time equipment sales as well. I covered this issue in detail in my SFTW Convo with Ben Voss.
Robots‑as‑a‑Service (RaaS) / subscription
Another model is a service model also called as Robots-as-a-Service, where the grower pays a monthly or per-use fee. For example, Monarch Tractor’s software package is $ 799 / month.
Compared to a direct sales model, the vendor does have access to recurring revenue, tighter data feedback loop, and easy to upgrade software. The capital and utilization risk does shift to the vendor, and requires a dedicated ops team. For example, McKinsey's report (2023) “Trends driving automation on the farm” said
Evolve business models to reduce the up-front capital costs associated with new automation equipment. For instance, models in which farmers pay regular subscription fees or share a portion of their cost savings with a vendor (such as through price-to-performance) can help make new technology more attractive and affordable for farmers.
If you can find a geographically dense market, where the customers already pay for your product as a service, and that market can get you to profitability, then you should consider as-a-service model. If you are established and have a leadership position, you should also be a service if you can be.
For startups, as-a-service allows them to continue to develop their product in parallel while making revenue. For startups with limited brand presence (which is basically any startup), as-a-service model gives a chance to learn with the customer as they build their product, as it is next to impossible to make a completely finished product (especially in ag robotics) and have to deal with recalls and huge customer service issues. The as-a-service model gives the startup the optionality to add additional value on top of existing services, and improve both their top line and bottom line.
The structure is better aligned with the farmers business as costs can align with outcomes, though there is little to no Cap-Ex, and no maintenance burdens. The risk for the farmer is not being able to access the robotic solution when they need it urgently.
Other models
There are many other models which are available to ag robotics providers like custom services provided by a 3rd party which owns the fleet, and sells tasks (for example, weeding, spraying, harvesting etc.) similar to custom application service providers.
In this case, the farmer does not need any new scales, pays only for what they use, though there are significant logistics management challenges for the service provider including weather and scheduling risks.
We have seen this work in custom harvesting in commodity row crops, orchard and vineyard spraying, strawberry harvesting, commodity row crop spraying in tight windows, and in areas with an established custom application culture. For example, Crinklaw Farm Services operates GUSS autonomous sprayers across orchards and vineyards.
Other models include lease or lease to own (not very different from leasing a car), technology licensing to an OEM and then relying on their dealer network.
As with anything else in agriculture, you have to match the context to the economics of the cropping system, existing ecosystems, and cultural norms to choose the right model or come up with a new one.
Key criteria to consider would be farm size and capital access (for example, bigger growers can spread fixed costs, but smaller growers like OpEx flexibility), crop value and labor intensity as the decision is heavily dependent on how tight the ROI window is based on timing and market conditions, seasonality and utilization, regulatory and safety complexity, dealer ecosystem maturity (strong dealer networks favor cap‑ex sale or hybrid franchise, weak networks push vendors toward direct RaaS), technology maturity (with a subscription or a service model farmers are not stuck with obsolete hardware for fast evolving etch), and capital market conditions.
Here are some practical GTM considerations for ag robotics founders
- Prove agronomic ROI first. Even in RaaS, pilots should show hard cost or yield deltas within as few seasons as possible.
- Secure channel partners early as dealers or custom‑service operators already trusted by growers cut acquisition cost and provide service footprint.
- Stage capital carefully. For example, robotics startup FarmWise struggled when fleet financing outpaced utilisation. Also make sure your fleet growth maps to contracted acres.
- Design for upgradeability. Swappable sensor pods or software‑defined implements extend asset life and support hybrid sale + software‑subscription models. Farm-ng (now acquired by Bonsai Robotics) positioned their products as modular, expandable, and easily upgradeable.
- Bundle financing and offer captive or third‑party lease packages. Carbon Robotics partner page lists a plethora of financing partners.
The bottomline is that successful ventures will sequence models to initially refine technology and prove value, and then migrate to equipment sales or licensing once reliability and brand trust are established.
Matching the model to crop economics, regional infrastructure, and capital availability is the difference between becoming the next Blue River–John Deere success and the fate of many shuttered robotics start‑ups.
So unfortunately, the answer depends.