1. Yes, agtech can help with getting relevant outcomes but this should not be an assumed outcome
2. Individual actors may not have a drive toward the societal objective
   • Founders may already be too invested in the tech. Funders have two objectives return on fund and raising next fund.
3. There may be a role for regulation (better if it happens before the tech is in the market than after it is deployed) — discussed clean water act and Silent spring / Rachel Carson
4. There may be a role for insurance to drive certain types of outcomes if their risk models are at risk
5. People matter too not just tech
6. This conversation matters most if there is a window of opportunity that we can exploit for impact on farms (per Pete)
7. Ecosystems in support for founders matter
8. Regulations protect incumbents

• Greg's thought — where do we invest in engineering research as a point of policy… can we ease the translation of technology.
• It is not reasonable to expect to get people to act outside of their perceived economic interest, at least in B2B interactions, as we have obligations to

1. Equity – equity (dual meaning) – question on what equity (in investing sense) means and defining terms
2. Debt did a number on small businesses – when did equity for small biz come on scene?
3. Social question – do people want to farm or not?
   1. The barrier isn’t desire – barrier to land, ownership of aging pop
   2. Tech problem – 1960s tractors, anything more recent you can’t repair yourself
      1. Right to repair, like a car
   3. Succession planning – land trust to help with aging farmers
      1. Conservation easements 
      2. Non technical innovation
4. Ag is too complicated for simple VCs / Silicon Valley model
   1. Researchers agree with VC
   2. Narrow margins
   3. Lack of understanding  
5. Omnipresence of big ag – what are the strategic decisions that can make this different?
   1. Employee buyback

What do we know about agtech’s ability to produce positive social impact? A few main reflections.

1. The first is that it is idiosyncratic, highly contextual, we cannot answer this question in a general way. I think we saw this with our first panel this morning – just how much context matters when you’re thinking about agtech and social impact.
2. The other piece is that as far as VCs are claiming to make social impact, for the most part, the sentiment amongst our group was that VCs aren’t really meeting their own standards. So, beyond returns to shareholders, they are not achieving their goals. However, the point was made as it has been a few times today, that VC at its core is not well placed to achieve social impact goals nor is it in a position to address market failures or unintended consequences. The question was raised if social impact is even an appropriate goal for VC? Again this comes back to the point that capital is not coded for social problem solving.
3. The trend is toward investments at the later stages in the life cycle of an innovation. Less appetite to come in early and take the risk required. There seemed to be consensus that start ups do not currently have the funding and support vehicles that they need throughout their lifescycles– there simply aren’t many programs that connect farmers, innovators, government etc.

What we need to know/do to progress?

1. On this topic, our conversation centred around models that emphasize an ecosystem based approach where VC is just one tiny part of the puzzle. As an example – we talked about the importance of networks or collectives. Paul brought up catchment area groups in the UK, which facilitate knowledge exchange between farmers and serve an important role in terms of identifying farmer needs and connecting them to other innovators and capital. Picking up on this idea of the power of the collective – in India farmer producer organizations provide farmers with group buying power, access to best practices as well as investments for pilot programs from NGOs, government and at times, private capital. These are also seen as highly valuable organizational forms in the innovation ecosystem for achieving social impact.
2. Our main takeaway from this conversation was that successful agtech for social impact requires sophisticated, elaborate structures – you have to leverage the philanthropic, corporate and public sectors. As Manoj said “magic happens at the intersections”. And I believe this emphasis on intersections ties into Steven’s point about the importance of intermediation in the innovation ecosystem.
3. And then we kind of ran out of time, but we basically concluded by saying there’s a role for regulation, we need more regulation, but that’s as far we got!

• In our conversation, we kept coming back to data governance, data ownership, definitions of data itself…
• Data governance is relevant to everything we were talking about in the morning of Day 1
• In relation to the perspectives from innovation intermediaries: How does data play a role in the decisions that investors and intermediaries make? How are ag-tech innovators valuing it and how is it transformed into an asset it and of itself?
• In relation to the conversation that the academics were having: they were discussing the pitfalls of ‘open data’ strategies, they were also discussing questions around how the generation, analysis and assetization of data about agricultural systems are re-enforcing existing inequitable relationships of power in the agri-food system? How can we change that?
• What does good data governance look like in the Global South? Where, as Ed says, interoperability is one of the biggest challenges? So, farmers would like apps that share data – but the reality is many different platforms that would like to keep their data proprietary. The reality really isn’t all that helpful for farmers because there are many different apps on the market that do one thing, but they don’t talk to one another. 
• Maaz told us about the serious games he’s developing in his lab to help answer questions around how we govern big data? Asking questions like what kinds of prices are farmers willing to get for their data?
• As Alexa said, the nut that we haven’t cracked is is the consumer going to pay for more sustainable food products? Is more data going to shift anything? Which comes back to that central question – do we need more information, or do we need more impact?
• Jen was speaking about the ways that large and powerful funders can dictate data strategies, so smaller start-ups that have ambitious data governance strategies that include openness and democratic governance, those ambitions can be quashed by larger funders. Which goes to show how powerful those funders are in the political economy of agricultural data.
• And then we were discussing the different data politics in different institutions – private sector vs. academia – the ways that we govern data are different, and that’s because we value data differently – to private institutions, it’s valuable economically; to academics, it’s a means to an end in some way.
• From myself, someone who’s really interested in data justice, I’m wondering how we can build an agricultural data economy that addresses issues of power and that builds long-term capacity of historically marginalized groups, like small-scale farmers, BIPOC farmers, Indigenous communities and farm workers.
• Another really interesting issue Alexa put forward was that we’re speaking different languages to one another – which is why conversations in this room are so important, we’re actively learning to speak over disciplines, from different perspectives, from different industries.
• The definitions of ‘scale’, of ‘data ownership’ are different to different folks so we end up speaking across one another and not fully comprehending each other.
• One of my takeaways here is that the definition of ‘data ownership’ is really blurry. Start-ups and larger ag-tech firms like to tell farmers that they ‘own their data’, but what even is ownership when you’re dealing with an intangible, non-rivalrous good? Why are we applying binary property relationships to something like data that is valuable in so, so many variable ways to so many different actors. Would it be impossible to, instead of understanding agricultural data through a model of ownership, understand through a model of equitable benefit sharing?
• One of the major themes of the day was scale.
• In the context of the fact that we are all speaking from different perspectives and different industries, what does scale mean to us all? And how is it defined and conceptualized differently by all of us in this room? And then how is it rationalized and operationalized by people with different levels of power in the agri-food space?
• And then coming back to Emily Reisman’s discussion about the ways that stakeholders and funders in Silicon Valley have an affinity for placelessness, how could scaling up work while also being place-based? Is that possible? We’re talking about agriculture – it’s inherently place-based, material, and relative to the land upon which it happens. The pressure to scale seeminly causes so much friction when it comes to real technological solutions that are on offer from the ag-tech industry.
• How are ag-tech innovators actually constrained by the incentive to scale, economically? And what technological possibilities could we dream up if those pressures were different?

• We discussed lots of about open access data systems
  • Effective open access systems depend on farmer’s providing data
  • Farmers have a reluctance to provide data for very good reasons…Colombia rice anecdote…don’t’ want to seem too rich and get kidnapped. Commercial reasons…stealing of good honey-bee sites…

• Solving the challenges:
  • Cooperative structure of data …Dairy One is a farmer owned cooperative…Farmer own the data…Valley Ag created some challenges bringing in a “silicon valley” approach
  • FOMO…is creating resistance…we want to collect all data,,,we don’t know for what but we want it. Do we need to be more specific of what data we need and what we will use it for?
  • Direct benefits to farmers…herd management and herd improvement that allowed collection of data

• Data systems: What data do we collect?
  • Maybe better for quantitative data
  • Not interested in transcripts of interviews
  • Are there challenges to creating a standard data reporting structure (ie do we loose nuance of various varieties of tomatoes)?
  • What kind of data resistance? Ie goats find their ways around digital fencing

• Continuing questions :
  • Is open access good, we can’t predict this summer’s weather…what data specifically would be effective at driving innovation?
  • How do we pay for it?
  • How to generate commercial value from open access?

• All of innovation has passed…
• Changes come on waves…some of the patents are first to file..rush to PTO when
• Open access issues
• Relaxing structure barriers to open access
• Architecture….
• Two sets of rules based on regulatory vs case studies
• Total disconnect between innovation and engagement
• US regulations…set up protocols…a few years later hear an answer…Japanese systems…more flexible.
• Farmers unwillingness to share data…
• If we have underdigitization…more about historical context…
• Trust factor for farmers to give data
• Institutional way …level of farmers…surprising…
• Farmer story…interviewing a farmer with national initiatives rice …lots of data got out…how is cultivation growing and changing overtime…farmers were concerned about hostages and kidnappings.
• Lots of debate about data cooperative…benefits all..joint decision making
• Already exists…farmer led cooperatives…farmers are involved
• EU lead the charge on traceability…
• Smallholder concerns
• Risk of open data…more data is fixation…everyone wants transparency…
• Open science…not interested in reading your interview transcripts
• Hard sciences is easier
• Social scientists
• More infrastructure, data overload, what is appropriate data
• All new technologies…unleashed …shifting…natural process…full digital unleashing…
• For what purpose…FOMO of data
• Farmers not comfortable with given open access
• How to farmers determine valuable
• Indigenous data …
• Inter-operability…data vocabulary
• What is lost from a standard vocabulary…
• I don’t want ot know factor…writing up a bad review…biologicals conference…people who are willing
• Sales man of distributor …how do they know product…farmer trust..don’t believe a word you say
• More control…we can go where we ant…
• Models can’t provide
• Other agendas who respond to
• Super weeds
• Data resistance…become data resentence …cybort goats…goats play with virtual fencing

• The theme from today’s discussions that struck me most was about the push and pull of innovation or the supply and demand of agtech. This topic comes up in my own work as I try to articulate the need for investing in and supporting specific verticals (livestock feed, cobots for farm labor, manure management, tools that are adapted for small holders’ needs). These verticals are do not have a huge ‘supply’ of innovations and they are not the trendiest technologies. BUT, I believe, and my research on investment dollars and also farmers’ sentiments suggests, that these are innovations that have a strong ‘demand.’ I think that the future of agtech, if its to be farmer-centric and advancing social impact in addition to environmental impact (and if it has any hopes of having widespread adoption) should be more focused on demand.
• A question that I want to continue to explore is ‘what else beyond M&A (or IPO) offers a viable exit strategy for early funders but enables a company to continue operating and executing on its impact values without much corruption.’

• What today’s sessions again pointed out for me was the idiosyncrasy and heterogeneity of agtech, agtech intermediaries, and agtech investors.
• For me that raised the question what we are actually referring to when we are talking about agtech innovation? What are the defining features of the agtech innovation process that distinguish it from other possible forms to organize innovation for agriculture and food? Where are its conceptual boundaries?
• For example, does the Silicon Valley model of innovation necessitates the involvement of venture capital?

• This was a tremendously rich and open conversation with critically minded people across an array of positions in relation to ag tech
• The immensity of the problems agriculture is facing seemed so self-evident, perhaps, that we didn’t really spend time on which problems are most important right now, or whether we would suggest different problems for scholars and practitioners alike to be foregrounding in their work. It would be interesting to spend a short session drilling down to the fundamentals of what the core problems are from everyone’s perspectives.
• We had an incredibly frank and productive conversation about impact – how it gets used and misused in the ag tech and related sectors, how it can be a discursive construction more than actual measured outputs in practice, and the ambiguity about what even counts. I especially appreciated the frank insights from various people working in industry about how impact is often self-selected or not even reported on in practice. This leaves me wondering: what is the point of claiming or trying to achieve impact through these channels? Has it entirely lost meaning and coherence, and if so, how can a more robust framework be developed?
• I’d like to hear more from everyone about what further venues for these kinds of discussions would be helpful. This has been incredibly illuminating, and I wonder what future applications or processes could come out of this. I’ve never been in a conversation quite like this, and appreciate the selection of individuals who arrived ready to dig in.
• I appreciated the real life case studies people brought in from their work experience.

Significant Takeaways:

• I really appreciated the detail to which we discussed the ways that ag-tech innovators are incentivized away from measuring climate impact – it’s good to really understand the ways that certain start-ups may want to measure impact but they’re simply constrained in ways that don’t let them
• Really enjoyed Kenney’s provocation that “we are at the end of the era of neoliberalism”
• How do we build the world we want in its wake?

Exclusions:

• I’m worried about the level to which the impacts of climate change are being under-estimated by stakeholders in the ag-tech industry.
• The conversations we had today didn’t really focus on the ways that ag-tech can facilitate climate adaptation – I’m apprehensive that those will be the technologies that we need much more than digitized carbon markets in the future.
• I would love more conversation around data governance, data politics, etc!

• Questions about time came up repeatedly: some have advocated slowness, that VC as a system, and technoscience generally, are aiming to speed up, when these time scales do not mesh with community ones, local needs, ecological realities. Similarly, many in the group encouraged patience for startups, and spoke in terms of many years to decades to see the change reach ‘scale:’ demanding speed for sustainability innovations in agtech is seen as unfair and potentially counterproductive.  Yet at the same time, climate crisis is upon us.  We must act to address and redress these issues and have no more time to spare.  Should we be patient with capital, even patient capital?  Is this the right temporal mechanism?
• Disruption / optimization: these are two of the great promises of digital technoculture, VCs, startups. They can align or be opposed.  But the question I found repeatedly emerging is are one/either appropriate for agrifood systems?  Some suggest disruption is the entire purpose of this entry, and without it, there’s no reason to engage.  Yet others insist disruption is counterproductive, and point to farmer adoption as the mechanism for change
• Scale vs. locality – this point has been made repeatedly, but no one seems to be engaging from either camp with the other directly.

• Innovation-specific theories of change with supporting evidence as a qualitative diligence tactic
• “Anti-goals” as documentation and strategy component
• Structuring LLCs in order to create direct investment opportunities among farming community who would not otherwise be SEC-compliant investors
• New indices to better align capital and innovation, moving away from commodity index which mainly rewards old capital
• I found it interesting that there is broad and vocal alignment on the needs for 1-patient capital, to address innovations unrewarded by 5-yr P and Ls and 2-novel incentives schemes whereby revenue can be wrangled into impact delivery. It’s motivating that there are some emergent models discussed here that can deliver progress, even if imperfect, despite some incompatibilities between traditional financial vehicles and impact outcome provision.

Really good conversation today that highlighted most of the big issues as follows:

• The ag innovation/agtech sector is operating in self imposed restraints (influence of big corporates, uniformed VC expectations, etc.)
• There are different ways to think about capitalizing the startups with blending capital sources, etc.
• The certification/standardization/metrics game is the wild west, but there is a general consensus that the future
• Data ownership/use was a good discussion I hope we get to pickup on more tomorrow. There was more alignment that I expected between pragmatism and idealism, so that was refreshing and the conversations where honest without people feeling defensive that I could detect.

Things that would be good to move forward from from the intermediaries’ perspective is:

• Best practices that can be adopted to new realities of investment, including a deeper understanding of financing, sustainability measurements, etc.
• More clarity on ownership potential for stakeholders and diverse founders and how to get there.

• Surprising alignment over the problem being one of incentives, not information
• Curiosity but limited examples of non-VC mechanisms that can center impact
• Entrepreneurs with high integrity around impact (Dewayne Dill) refuse to participate in green washing, yet are disincentivized by collaborators to validate/verify sustainability metrics
• How to support modest ‘life-style’ business that do not interest VC but might bring real value to rural communities
• Widening the scope of stakeholders beyond farmers or farmworkers to rural communities and even global flows of people
• Making an anti-goal list. Being able to identify when a product fails to achieve its intended outcome. Adoption rates x hypothetical efficiencies are not enough, needs to be place based.
• To what extent are technologies like automation and efficiency/precision short-term bandaids for fundamental social and political problems: dignified livelihoods for all, land tenure, regulation of environmental outcomes

Key impressions:

• Scalability, and refusal to scale, as a major theme – interest in options that are more local and tailored
• Degrowth – as also in some sense already the aim of US ag policy (trying to keep farmers from growing too much, to keep price up)
• Question of "unintended consequences" – not necessarily the same as "unanticipated consequences" – how to set up systems, tactics, and institutional structures so that those imagining solutions need also to account for the things besides what they are aiming for [one option: "anti-goals" (Maria)]
• VC understood as maybe not the greatest fit for innovation in ag for sustainability, in part because of the pressure to scale, also because the time horizon it emphasizes doesn't meet the pace of agricultural technology (and because margins are low, etc.)
• VC as not necessary for innovation – there are other avenues also (incl. public investment)
• Challenge of operationalizing impact, knowing whether you are achieving it, lack of baked-in incentives to be disciplined about asking & answering this queestion
• Rebekah: "There is intense friction" between achieving environmental goals and producing enough revenue to survive
• The level of thoughtful, critical reflection and self-reflection about limitations and strictures in current systems in this room is very refreshing

What we didn't talk about:

• Models for AgTech that don't require large-scale data collection at all – it is possible to design useful digital technologies where data simply stays local, but the vision of Big Data continues to dominate imagination in this space and then produces the question "how to collect it all equitably" when we also have the option to not collect it

• VC model is looking for a “killer app”, but the diversity of agricultural needs by sector, geography, economy means that incremental solutions are more likely to have measurable impacts – many small improvements, some of which are not innovative or are mundane
• Need for nuanced structures to help ag tech address sustainability, spanning regulatory, public, philanthropic
• Where there can be no outsized profit, there will be no meaningful VC participation, and so solutions that are simple, open, built on e.g. traditional ecological knowledge, are inherently disadvantaged when VC is the one path to mass market
• Place-based approach

• People are a key asset, the human capital
• There is need for exploring the various sources of funding & other resources beyond “VC”
• There is not much understanding of what “VC” is, how it works, where the funds are from
• There is need to understand the different needs of innovations at various stages of its business life
• I talked too much but felt I had to represent early stage innovation
• How to draw out the other experts in the room who are quiet (?)
• There is a lot of misunderstanding of “metrics” and the expectations of early stage companies
• There are very many topics about which much more study is needed
• There are identified good places for interface & synergy

• We have painted with a broad brush in our references to VC. My interest lies is in agrifood innovation and contemporary infrastructures/imaginaries focused around private capital. VC may be an unhelpful reference or term. We have heard many people – Manoj, Martin, Pete, Wolf and many others argue that there are capital pools outside of VC that must be created/tapped to realize innovation for impact. A focus on VC may be unhelpful.
• That said, it is possible that some will suggest that patient capital or public/philanthropic capital can derisk innovations, thereby making them appealing to VC. In this case, it is a sequential model, whereby VC remains quite central at a later stage. Yet, Rebekah Moses’ experience suggests that VC may not be capable/interested in producing social impact.
• We learned a lot today about questions of scale. Possible that agr innovation must be place based and geographic, sector specific. If this is valid (and CEA takes us to placeless innovation), VC may again be a poor horse to ride.

• What are the burdens and hurdles for startups in tracking and reporting?
• What should early stage companies be thinking about re: metrics and reporting?
  • Companies should be planning and thinking about metrics from the beginning. This helps them bake it into the DNA of the company and orient the business’ growth appropriately
  • Reporting should be commensurate with the stage of the company (seed stage reporting will look different from Series B)
• Social impact = adoption X impact
  • Is an incremental impact acceptable?
  • Or does it need to be a step change?
  • Very few agtech innovations will be truly transformative, many more will be incremental.
  • Many enabling ag technologies will pop up to support those few transformative innovations
    • Ex: no-till and conservation tillage as a transformative practice change. Then new tools and chemistries come onto the market to support this practice change.
• Impact achievements in agtech will be slower because it takes longer for tools to get to market
• Is all agtech climatetech? Probably not.
• Interrogating the following thesis:
  • As agtech adoption increases, efficiency will improve, inputs will decrease, which will boost social and environmental outcomes.
  • ^ this is too simple a thesis. Agtech is so different depending on the context, farm, geography, crop. An impact thesis for agtech needs to be more nuanced and specific.
    • ^ which in turn reminds us of Social Alpha’s ‘problem driven’ investment process.
    • And when looking at a ‘problem driven’ investment, you must then ask yourself ‘can this problem be solved with innovation?’

Jenn Smith

Does the context in which intermediary drives conditions/is embedded (university, state, cooperative) matter? Which institutional form is most effective at addressing impact goals?

• Centrality of org’s mission orientation –silicon valley model includes corporate headhunting, problem solving – reason for being, mission is to address corporate needs – including CSR goals

• Dedicated resources to staff accelerator – work to meet these needs outsourced by companies like General Mills, Mondelez

• SV model is innovation as homeruns rather than base hits – incremental model isn’t sexy

• Org disconnected from incumbents, without expectation of quick returns, freedom can lead change – new definitions, new markets, new economies – a long leash gives freedom to be revolutionary

• Revolutionary vs evolutionary startups
  • Tight context – resource and ecosystem benefits, doomed to path dependency.
  • Incremental change
  • Either way it leads to normalization

• Mission driven acceleration at national level – Netherlands national commitment to circular ag – how to enact – private consortia provide $, incumbents/firms, making policy available

• Public policy programs – requires orchestration – how do you reconcile ambition to swing for the fences with public resistance

• Benefits and costs of national innovation system, serving specific sectors, through economic development lens
  • Raising bar with standards, change inputs/conditions, regulatory pressures to push move

• National champions – are there models where they align public and corporate
  • Province led initiatives in Canada (we didin’t discuss, but I’m fascinated by Alberta’s blended protein strategy)
  • NZ – publicly funded ag tech accelerator

• Private capital to leverage public money, prove effectiveness, remove barriers to adoption
  • Are there other customers that are relevant in the conversation about adoption and impact
  • Process, product, system innovation – optimizing current system – infrastructure

• Hattie not seeing environmental change, seeing social change – community development financial institutions getting capital out into areas that have been underinvested
  • Constant fundraising means don’t have resources to focus on IMM practice – government money towards that would be welcome.
  • Debunk the myth of startup heroism – never go alone

• SV is largely a finance ecosystem – how do networks shape – looked at SV example and found out it was just about the money, not about talent pool, resources, culture. Simply a finance ecosystem

University embeddedness – does university ecosystem facilitate more wild card, futurist research, who is doing a good job of translating research to practice?

Maaz Gardezi

Reporting to the group:

• What makes relationships work and trustworthy with farmers?

• Technology and data is a commodity.

• Transformational change is: all of this has to be studied (by academics) what we are building with the blockchain, and data commons. In this way, we could rethink how this can work and be more impactful.

• Building safe and fair playground for everyone. In the investor ecosystem, simple homogenous ecosystem exist (rich older guys). Who else has to be at the table? More forums (like this) especially for the ones at the highest level ($4B funds) that are hard to get into these forums.

• If you care about the impact, you will have to start with the assumption that there will always be positive and negative impacts. They’re pushing all the positive stuff, imagery of the world where it is only a positive world. Can you change those structural incentives so that there may be alternative futures that are possible?

• Incentives are aligned a bit differently as you move from the VC to the academic world.

• What are the failures of the VC fund? Shaking out the myth that everyone is missing. Half of them fail, few percent get to 1.5X, 10% get to the 100X.