Key Takeaways

The Most Expensive Mistake in Indoor Farming

If you could distill the past five years of indoor farming failures into a single lesson—the one thing that the bankruptcies, the shutdowns, and the billion-dollar write-downs have in common—it would be this: the companies that treated farming as a technology problem failed, and the companies that treated technology as a farming tool survived.

That distinction sounds simple. It is not. It represents a fundamental difference in organizational culture, hiring priorities, capital allocation, and decision-making that shaped everything from facility design to daily operations. And it is the single most reliable predictor of which indoor farming companies are still operating today and which are not. Why Vertical Farms Keep Failing — And What the Survivors Are Doing Differently

The Tech-First Trap

The tech-first approach to indoor farming followed a recognizable pattern. Companies were founded and staffed primarily by software engineers, data scientists, and hardware designers who viewed agriculture as an engineering optimization problem. The assumption was reasonable on the surface: plants need specific inputs (light, water, nutrients, CO2, temperature), and if you build a sophisticated enough system to deliver those inputs precisely, the plants will grow optimally. Therefore, the core challenge is engineering, and the core talent needed is engineers.

The assumption was wrong in a way that cost the industry billions of dollars. Plants are biological systems, not machines. They respond to their environment in complex, sometimes unpredictable ways that cannot be fully captured in sensor data or optimized by algorithms. A tomato plant that receives theoretically optimal light, temperature, and nutrients may still underperform because of a root zone oxygen issue that no sensor detected, a humidity micro-climate that the environmental model did not account for, or a pathogen pressure that required the kind of pattern recognition that experienced growers develop over years but that AI models have not yet learned.

The organizational symptoms of tech-first thinking were visible in hindsight. IT personnel were paid significantly more than agronomists—a signal about what the company valued. Technology budgets dwarfed crop science investment. Engineering teams made facility design decisions that agronomists would have flagged as problematic if they had been in the room. The resulting operations were technically sophisticated and agronomically inadequate: impressive systems that could not consistently grow profitable produce.

What the Survivors Did Differently

The companies that navigated the shakeout share a fundamentally different starting point. They began with agronomy—what does the plant actually need?—and then asked how technology could deliver those conditions reliably and cost-effectively. Technology was a tool in service of growing, not the other way around.

AeroFarms’ turnaround under new leadership is one of the clearest examples. When the company emerged from bankruptcy, the new CEO’s first priority was hiring people with deep expertise in food production. Not more software engineers. Not more data scientists. People who understood how to grow food at commercial scale and get it to market profitably. The shift was not anti-technology—AeroFarms remains one of the most technologically sophisticated operations in the industry. It was a reordering of priorities: agronomy leads, technology supports. From AeroFarms to Profitability: The Turnaround Story That Could Redefine Vertical Farming

80 Acres Farms offers a different version of the same lesson. Co-founder Mike Zelkind has been direct about their approach: they focused on unit economics from the beginning because they came from the food business, not the technology business. That background meant they evaluated every technology investment through a profitability lens rather than a capability lens. The question was never “can we build this?” It was always “will this make the operation more profitable?” That discipline is now reflected in one of the industry’s most comprehensively automated and operationally sound facilities.

The consensus that emerged from Indoor Ag-Con 2025 reinforced this pattern. The companies presenting as successful were consistently described as cautious spenders focused on doing one thing consistently well, rather than ambitious technologists trying to solve every problem simultaneously. The shift in tone from the same conference three or four years earlier was striking.

What Farmer-First Technology Actually Looks Like

The distinction between farmer-first and tech-first is not about the sophistication of the technology. It is about who the technology is designed to serve and how it integrates into actual farming workflows.

Farmer-first technology provides decision support that enhances grower expertise rather than attempting to replace it. The best crop management tools in indoor farming present data, surface anomalies, and suggest actions—but leave the final decision to the grower who understands the context that no sensor can fully capture. The experienced grower who notices that a crop section looks slightly off—before any sensor registers an anomaly—is exercising a form of pattern recognition built over years. Technology that complements that expertise produces better outcomes than technology that overrides it.

Farmer-first interfaces are designed for the people who actually use them in daily operations. This sounds obvious, but it is the exception rather than the rule. Too many agtech platforms are designed by software engineers for software engineers—featuring complex dashboards, extensive configuration requirements, and terminology drawn from data science rather than agriculture. The grower standing in the facility at 6 AM should be able to see what matters, understand what is happening, and take action without a training manual. If the interface requires an IT team to configure or interpret, it is not farmer-first technology.

Farmer-first systems are modular and adoptable incrementally. A farm that is not ready for AI-driven predictive optimization should be able to start with environmental monitoring and data collection, build a foundation of historical data, and add predictive capabilities when the data and the team are ready. Systems that require all-or-nothing adoption—replacing every existing tool and workflow in a single implementation—fail at the adoption stage because they ask operators to take on too much risk and disruption at once.

Farmer-first platforms integrate with existing equipment rather than requiring wholesale replacement. Most indoor farms run a mix of systems from different vendors—climate controllers, irrigation systems, lighting controllers, sensors—and any new platform that demands replacing all of that infrastructure is creating a barrier that most operators will not cross. Open platforms that work with what is already in place earn adoption; closed ecosystems that demand full commitment lose it. The Automation Playbook: How Robotics Are Making Indoor Farms Profitable

The IP Trap

One of the most candid observations at Indoor Ag-Con 2025 came from Tisha Livingston of 80 Acres and Infinite Acres, who reflected on the industry’s early years with a statement that resonated across the room: if they had been more open and less afraid of sharing IP, the industry could have progressed much faster.

The IP protectionism that characterized indoor farming’s first wave had understandable origins. Companies invested heavily in proprietary systems and felt—reasonably—that those systems represented competitive advantages worth guarding. But the result was an industry where every company was independently solving the same problems, often making the same mistakes, and accumulating hard-won knowledge that died inside the organization when the company failed.

The irony is significant. Billions of dollars in collective learning—about crop recipes, environmental control strategies, equipment reliability, cost structures, and operational workflows—was generated and then largely lost during the bankruptcy wave because it was locked inside proprietary systems and cultures that treated operational knowledge as trade secrets. The industry essentially paid for the same education multiple times, at enormous cost, because each company insisted on learning its own lessons in isolation.

The shift toward greater openness is not altruism. It is pragmatism born from painful experience. The operators who survived understand that a growing, healthy industry benefits everyone more than marginal competitive advantages benefit any single company. The rising tide argument was always theoretically compelling; now it has the weight of empirical evidence behind it.

The Path Forward

The farmer-first versus tech-first distinction is not a binary. Every successful indoor farming operation uses sophisticated technology. The difference is in the hierarchy: does agronomy drive technology decisions, or does technology drive agronomy decisions? In the companies that survived, the answer is consistently the former.

This has implications for how the industry evaluates technology vendors, hires teams, and allocates capital going forward. Technology platforms should be evaluated on whether they make experienced growers more effective—not on whether they can theoretically replace growers with algorithms. Hiring should prioritize agricultural expertise and add technical capability, not the reverse. Capital should flow first to the agronomic fundamentals—climate control reliability, nutrient delivery precision, crop science research—and then to the technology that enhances those fundamentals.

The companies that will define indoor farming’s next chapter are the ones building technology that a grower with twenty years of experience would look at and say: this makes my job easier, this helps me grow better, this gives me information I could not get on my own. That is the standard. Technology that meets it will earn adoption. Technology that does not—no matter how technically impressive—will end up as another expensive lesson in the industry’s growing catalog of them.