A look at the AI technology involved in vertical farming

As vertical farming captures imagination with its promise of transforming agriculture, the industry stands at a critical crossroads. Through 2023-2024, several heavily funded players have shuttered operations such as AeroFarms and Infarm. Yet amid this dramatic industry shakeout, well-positioned players are finding success, particularly in Europe where venture funding has begun flowing again. The pattern suggests a crucial industry maturation: while billions in venture capital initially fuelled rapid expansion, only those companies with robust business models and technological advantages are emerging from this period of consolidation.

For a nascent industry, this is not unprecedented. The dot-com bubble of the late 1990s saw over 450 tech companies go public in 1999 alone. When the bubble burst, nearly half of them failed and the industry lost $5 trillion in market value. Yet from those ashes emerged today's tech giants, companies that combined innovative technology with sustainable business models. The question now is: which vertical farming companies will survive and thrive through this market correction? For a glimpse of what that might look like, we need only visit New Jersey, where an unlikely success story is unfolding.

From factory to farm
In a former Budweiser factory transformed into a high-tech agricultural facility, strawberries grow in perfect vertical rows under LED lights, tended by a choreographed dance of robots and bees. This is Oishii's new Amatelas Farm – a 237,400-square-foot testament to defying industry conventions.

Courtesy Oishii

The facility's name, honoring the Japanese Goddess of the Sun, hints at its ambitions. While other vertical farms focused on quick wins with simple crops like lettuce and herbs, Oishii took aim at one of agriculture's most challenging products: strawberries.

A new approach
Navigating the challenging landscape of vertical farming requires more than just technological innovation – it demands a visionary approach. For Hiroki Koga, Oishii's CEO and co-founder, this contrarian approach was informed by lessons from his native Japan, where he witnessed a similar boom-and-bust cycle in vertical farming a decade earlier. "What we are seeing in the broader industry is the impact of when companies scale too quickly without a differentiated product," he reflects, seated in the facility's control room where screens display real-time data from thousands of sensors.

"When it comes to produce like lettuce, kale, spinach, etc., it's hard for consumers to differentiate between traditionally farmed vs. vertically farmed—which has hindered scalability and profitability."

What sets Amatelas Farm apart isn't its choice of crop – it's the integration of artificial and biological intelligence that makes its success possible. While most vertical farms resemble static warehouses, Oishii's facility is in constant motion. Two hundred and fifty automated racks move in carefully orchestrated patterns, creating an environment where robots, humans, and bees work in harmony.

Courtesy Oishii

This mobile architecture reflects a fundamental rethinking of indoor agriculture. Inside the facility's individual farming units, each rack is stacked with eight growing levels that move through a choreographed 24-hour cycle, mimicking the natural rhythm of day and night. The result is what Koga's team calls their "environmental recipe" – a precise calibration of conditions that seemed impossible just a few years ago.

Big data
Central to this achievement is an artificial intelligence system that processes an astounding 60 billion data points annually. These insights come from approximately 50 proprietary robots equipped with machine vision, monitoring everything from flower development to bee activity.

To put this in context, each "data point" represents a specific piece of information extracted from images – identifying everything from a flower's stage of development to a bee's location, which the AI uses to make real-time decisions about growing conditions.

This level of sophisticated data processing echoes broader trends in agricultural AI, where companies like Microsoft and Google are deploying advanced machine learning to transform farming from an art into a data-driven science.

Reimagining pollination
The most remarkable aspect of this system might be its success with pollination – a challenge that has long bedeviled indoor farming attempts. "Our bees act as they would in the wild," Koga explains, gesturing toward a section where bees buzz between flowering plants. "Flying to find the nectar, pollinating each strawberry flower, and then returning back to their hive. Inside our farms, our bees experience the perfect day, every day."

The numbers tell a compelling story. While traditional farming typically achieves pollination rates of 60-70%, Oishii reports rates above 95%. This dramatic improvement isn't just about numbers – a 2023 study in Agronomy found that effective bee pollination can increase fruit set by more than 50%, directly improving both quality and marketability.

"Combined with other environmental data, our AI calculates how much bee activity is needed that specific day to help with the pollination," notes Koga. "This enables us to achieve higher efficiency than conventional farming, which ultimately allows us to operate a profitable farm with a high initial capital investment." The marriage of AI and agriculture at Oishii isn't just a technological triumph – it's proving to be a viable business model in an industry desperate for profitability.

Breaking cost barriers
When Oishii first launched, eyebrows raised at their $50 price point for a single package of strawberries. These weren't just any strawberries, though. The company had chosen to cultivate particular varieties from Japan, where fruit often commands premium prices and is deeply embedded in gift-giving culture. In Japan, luxury fruit is such a significant part of the culture that companies like Sembikiya – considered the "Tiffany of fruits" – operate more than 10 stores in Tokyo, and premium fruits are traditionally offered to close friends and family during the gifting seasons of Ochūgen and Oseibo. The cultivation process is painstaking; farmers often monitor each fruit constantly, sometimes even hand-pollinating or individually shading each piece to protect it from too much sun.

Today, Oishii's premium berries sell for $10 – a price trajectory that defies typical patterns in vertical farming, where many companies have struggled to reduce costs at scale. The achievement stems from what Koga describes as a relentless focus on operational efficiency driven by AI insights.

"Our goal at Oishii has always been to make high-quality produce an everyday enjoyment," he shares. The price reduction has been "primarily driven by continuous operational improvement, cultivar development, automation, and scaling."

Beyond strawberries
This success with strawberries has led to expansion into other challenging crops. The company recently introduced the Rubī Tomato, leveraging its AI systems' ability to maintain precise growing conditions for different varieties.

"Expanding to other fruits and vegetables has always been on the Oishii roadmap since day one," Koga explains. "We have already been able to apply our proprietary technology to other types of produce." The company plans to continue launching new varietals of berries at different price points, while also developing other flowering produce. "Ultimately, the decision point is determined by whether we think there is a clear roadmap to cost parity with conventional products," he notes, planning to share more details at the AGENTIAL AI AgBio conference this coming May.

Indeed, the prospects for tomatoes in vertical farming appear particularly promising: recent trials with innovative horizontal growing systems show yields could reach 110 kg per square meter compared to traditional greenhouse yields of 40kg. Some companies are achieving these improvements through novel approaches – from gene-edited varieties promising up to 180-400% yield increases to multi-layer cultivation of dwarf varieties that can produce more tomatoes with less effort and fewer resources.

For Oishii, their success with strawberries provides a blueprint for growing these more complex flowering crops at scale. For a growing number of agricultural researchers, this approach represents more than just a business strategy – it's part of a broader technological revolution. As emerging research is noting, artificial intelligence is fundamentally rewriting how we understand and manipulate crop genetics, with companies like Oishii at the forefront of translating these technological advances into practical agricultural innovations.

For investors who have watched multiple vertical farming ventures collapse under the weight of high operating costs, Oishii's trajectory offers a different narrative. Rather than racing to scale with simple crops, the company focused on perfecting complex processes that create high barriers to entry. As Koga puts it, "In order to stay competitive in the market, you need a strong tech moat to keep competitors at bay. We wanted to tackle something that people thought was impossible, and we knew if we could crack the code on growing flowering fruits at scale indoors, the possibilities were endless."

The implications of Oishii's success extend far beyond premium produce. As global agriculture faces mounting challenges, their AI-driven approach offers potential solutions to fundamental industry problems.

Inside the Amatelas Farm's monitoring center, Koga contemplates these broader implications. "I am very worried about our current failing agriculture system," he reflects, his expression serious. "It is no longer sustainable – the rate of production cost is at an all-time high, resources are becoming harder to source, we are dealing with extremely unpredictable and unprecedented weather conditions, and the global population is exploding."

Sustainable farming
Yet his concern is balanced with optimism about technology's potential to address these challenges. The Amatelas facility, powered by an adjacent 50-acre solar field, demonstrates how vertical farming could evolve beyond its current limitations.

The industry is increasingly turning to renewable energy to solve one of vertical farming's biggest challenges – energy costs. Some pioneers are showing what's possible: South Australia's Sundrop Farms produces 13,500 tonnes of tomatoes annually using just sunlight and seawater, while others are integrating solar directly into their operations, like Jones Food Company's 710kWp solar system at what they call the world's largest vertical farm. For Oishii, their solar-powered facility represents not just environmental sustainability but economic viability – a crucial step toward making indoor agriculture both profitable and scalable.

Oishii's sophisticated water recycling system offers another glimpse of agriculture's potential future. "Our bees, robots, farmers, and engineers synergistically work together to grow quality produce all year-around," Koga explains, "by blending nature and ancient Japanese farming techniques with cutting-edge technology and innovation. We are the only company that has been able to create the perfect indoor ecosystem for bees and naturally pollinated produce at this scale – once widely thought to be impossible in the agriculture industry."

The future of farming
This fusion of artificial and natural intelligence could represent a new chapter in agriculture's long history. While most vertical farms have focused on replacing natural systems, Oishii's approach suggests a future where technology enhances rather than supplants nature's complexity.

As climate change threatens traditional growing regions and population growth strains food systems, the ability to create perfect growing conditions anywhere becomes increasingly valuable. "Indoor vertical farming allows us to grow great-tasting produce in a way that's better for people and for the planet," Koga notes. "At Oishii, our vision for the future is to continue to deliver on our mission to make delicious food more sustainable and accessible."

Indeed, as noted by experts like Graeme Smith, Chair of the AI Reference Group at the International Society of Horticultural Science, "AI-powered controlled environment agriculture (CEA) systems like Oishii's are at the forefront of creating more climate-resilient and efficient farming methods. This trend towards emerging 'deep-tech' for reducing labor and increasing profitability is a much-needed incentive for farmers to invest in sustainable practices and will play a pivotal role in our long-term food security. These transformative technologies and their practical implementation will be a key focus at the conference, where industry leaders will share insights on scaling successful AI adoption in agriculture."

Oishii's journey from premium strawberries to scalable AI-driven agriculture represents just the beginning of technology's transformation of farming. At the upcoming AGENTIAL AI AgBio conference in Bangkok this May, Koga will join world-leading experts exploring this new frontier of deep tech in CEA and crop science.