Editor’s Note: This post is co-authored by AFN’s editor Louisa Burwood-Taylor alongside Mark Johnson, who recently led grain trading platform GrainBridge to acquisition and was cofounder of Descartes Labs.
A few months ago, Bayer and Microsoft announced a partnership.
Typically, these cloud ‘partnership’ announcements are nothing more than customer agreements: company X has become a customer of cloud provider Y. After digging a bit deeper, Louisa published an article on AFN introducing the concept that this might be something much more interesting than purely Bayer’s decision to use Microsoft’s cloud computing service Azure.
That article went quasi-viral and spurred a number of responses including from Mark, but also the likes of other agtech newsletter writers, Rhishi Pethe and Shane Thomas. This article tries to unpack some of the responses to give a more complete picture.
What’s the big deal?
Light on details, the deal appears to be a plan for the two parties — both leaders in their spheres — to create a digital infrastructure platform for agtech software companies to build applications off.
Said in another way, the partnership will take care of data processing, storage and behind-the-scenes data management and calibration so that third-party engineers and data scientists (entrepreneurs) can focus on building algorithms to offer the agrifood industry insights from the various data sources at their disposal.
Bayer and Microsoft would not confirm what data sets they would bring to bear for this service; it was implied that any third parties wanting to use the platform would provide their own and they would comply with all data privacy protections etc, but given their huge — and growing — footprints across the industry, it’s hard to believe they wouldn’t offer up use of some of that to potential customers.
Louisa leaned towards believing this platform could become a significant source of new revenue for Bayer – and also possibly for Microsoft through revenue sharing; Bayer has yet to make direct revenues from FieldView. Mark ruminated on the idea that the partnership was a sort of trojan horse for Microsoft to expand and consolidate its growing footprint in agriculture via Great Plains and Microsoft Dynamics.
But we also set about getting feedback from the industry on what the significance of this partnership could be for the industry.
What the poll said
Is this not much more than an announcement regarding Bayer’s plans to use Microsoft’s cloud computing service Azure, or could this be the foundation of a “Meta of agriculture” that controls much of the world’s agricultural data?
Sixty-five percent of people thought the deal was significant, according to a somewhat over-simplified poll Louisa put up on her LinkedIn feed. The rest of the 193 who voted said it was not; and several of the 12,605+ people that saw the poll reached out to say they wished there was a third “jury is still out” option.
Before we dig into the responses, let’s define the problem, solution, and the customer with what we learned and know as a framing.
The problem: Agtech companies seem to build everything tabula rasa as if they were the only company building software for agriculture. There are common types of data (field boundaries, USDA data, prices, etc.) that every company builds software around. This wastes an enormous amount of resources that ought to be going into solving user problems.
The Bayer/Microsoft solution: Our most generous reading is that Microsoft and Bayer are trying to build a platform to enable the rapid development of agrifoodtech applications, by bringing a combination of modern cloud tools and adding a layer of data.
The customer: This is not immediately clear and is likely a moving target for Bayer and Microsoft. It might be companies in the ag value chain that don’t have big IT departments. It might be startups. It might be all of the above as Bayer presented it in their release.
Response #1: Startups are already doing this, but how are we defining infrastructure anyway?
A lot of the discussion that came out of this announcement focused on defining digital infrastructure, leading to larger questions around what types of platforms could be defined that way, versus being a software application.
Sachi Desai, a Climate Corporation corporate strategy manager, said on Twitter: “It depends on what we define as digital ag; if it’s about providing the rails for information to move then most definitely it would be… like all things it’s nuanced.”
Jake Joraanstad, CEO of Bushel, added “we need to clarify who gets to use the word digital infrastructure in ag… Hint: it’s not these two.” Obviously, the technical idea is much more akin to Azure and AWS offerings.”
There were also references to startups already doing a lot of this work such as Leaf, which is specifically working on data infrastructure, Regrow, in which Microsoft is an investor, and Bushel. There are potentially a whole generation of new ag companies working on the software tools, data feeds, and communication as a new ag-focused technology stack
In Mark’s opinion, here’s what happened: Bayer realized that it had created a bunch of technology, data feeds, and analysis… and they saw that many, many other agtech companies were creating the same thing (e.g., access to satellite imagery with NDVI and field boundaries) and that the smart thing to do is to share that infrastructure across the industry and compete on a higher-level value. With ready-made tools, the best product with the highest value to the end-user is what wins, not the underlying technology. Any actual technology investments should be something unique to that particular app, not what many others have built.
Niall Haughey, the “agfintech guy” was more agnostic about who gets to create the infrastructure but reflected his keenness to see more action. “What I would love to see is more independence in the data innovation space in ag,” he wrote in a LinkedIn comment. “If this partnership does that, great! If not, then it would be fantastic to see other independents such as Leaf Agriculture or Bushel to name two examples crack on with it. Either way, I think the data infrastructure space is about to get very interesting and the best governance will win.”
Response #2: How does this relate to Microsoft’s existing footprint in agriculture? Is it a trojan horse for something else, and where do other cloud providers fit?
And we don’t necessarily mean FarmBeats, which by and large did not receive great feedback from respondents familiar with the platform in terms of scale or functionality, although FarmBeats’ ambition to become a marketplace of apps to customers and data infrastructure layer was alluded to by some.
But more Microsoft’s entrenched nature at many of the large agribusinesses; ADM’s and Cargill’s IT departments still don’t allow Google Meet as a collaboration app, it’s all Teams. All laptops are Windows. Google is seen as a consumer company; and Amazon is always potential competition because they are deep in the supply chain.
Seana Day from Culterra Capital had some interesting thoughts: “What I can say with more conviction is that Azure, GP, Dynamics, Power BI are being used by a lot of operators and the footprint is sizable and valuable in Ag. Maybe they are trying to figure out what role to play in the field by learning from as many partnerships as possible, and someday in the future, they will close the gap with the productivity and BI apps where they can dominate.”
“Maybe MSFT is trying to elevate their profile in the space, in anticipation for the long game. It just seems like their current agtech capabilities, relative to marketing fanfare, are not quite aligned. But if you are a strong horizontal software player and you are lacking vertical-specific datasets, you partner until you build up enough institutional ag expertise (and incremental cash flow to satisfy The Street).”
But this level of trust among the large agribusinesses could also be key to making this Bayer deal work, and give them a leg up on AWS and GCP because they are Microsoft.
Another way to frame this response is: what’s will other cloud vendors do?”
Response #3: This deal is not significant because Bayer and Microsoft don’t have the right structure and are too big and cumbersome to make it work.
Coming partly back to Niall Haughey’s comment, another question that arose was whether building out the infrastructure for agtech was actually better suited to a startup than two big giants.
Of course, the question of trust comes up and Louisa had some initial remarks on that in her initial article regarding this becoming a ‘Meta of agriculture.’ Can Bayer+Microsoft be independent enough? Microsoft is an agnostic provider of technology but Bayer most certainly is not, because of their seed/inputs businesses. To be successful in the long term, will they need to spin out? That is, will other companies trust using services from Bayer? There is some precedent here with MyJohnDeere and even FieldView itself integrating across the industry.
Other respondents also thought the duo did not have the chops to make this work.
Sivam Krish, CEO of GoMicro, expressed doubts on LinkedIn that the two corporates are nimble enough to make this into anything significant. “Can Microsoft build the windows for AgData? I doubt it very much. Committees often don’t have the right mental infrastructure to do really cool stuff – even though both companies have everything in place, except the appetite for risk, which is the necessary ingredient for transformative outcomes.”
Response #4: Time will tell; it’s hard to say
This is a frustrating response, particularly for a journalist, but many offered great reasons for sitting on the fence, not least that the deal was littered with vagaries. Several respondents wanted to see some initial use cases before they could opine on the significance of this partnership. Agri consultant Richard Ferguson responded to AgFunder’s Twitter thread about the deal:
“I’m 50/50 on the relevance of this. There are plenty of examples of firms (existing and start-ups) engaged in this type of work. However, few of them attract the capital (not unlike the multiple competitors of @Meta who didn’t become THE platform for social media),” he wrote.
“Farmers are fragmented and conservative. Do they embrace this? Consider some of the agtech ventures you’ve covered in the past. Did Agriconomie ever become “The Amazon of agricultural supplies” as sold to investors? So, who knows. If it works – you spotted it early. If it doesn’t – it just joins an already long list of input providers, traders, processors and CPGs trying to figure it out.”
What do you think? Email [email protected]