In December 2025, AeroFarms — once called the poster child of vertical farming, a company that had already survived one bankruptcy — permanently shut down its Virginia facility and laid off 173 people. It joined a lengthening list of casualties: Bowery Farming, once valued at $2.3 billion, halted all operations in late 2024. Plenty, backed by Jeff Bezos and SoftBank, filed for Chapter 11 in March 2025 after its Compton, California facility saw energy prices jump 15 percent year-over-year, cutting output by 30 percent. AppHarvest liquidated entirely. According to one industry accounting, energy costs can consume 50 to 70 percent of a vertical farm’s operating budget — and it’s the single line item most responsible for turning a technologically impressive idea into a graveyard of billion-dollar bankruptcies. The industry’s survival now hinges substantially on a narrower, more technical question: can smarter light finally do what cheaper electricity hasn’t?
The Scientific Foundation
The core physics problem is unforgiving. Vertical farms replace free sunlight with electric light, and plants are remarkably picky about which wavelengths of that light they actually use. Chlorophyll absorbs strongly in the blue and red portions of the spectrum but reflects most of the green light that makes plants look green in the first place — meaning a substantial fraction of any “full spectrum” white light delivered to a crop is simply wasted, converted to heat rather than photosynthesis. LED lighting solved the first half of this problem by making it possible, for the first time, to emit narrow, specific wavelength bands rather than broad white light, and a 2026 industry report notes that lighting efficiency has improved by roughly 60 percent over the past decade, reaching around 3.5 micromoles of usable light per joule of electricity in the best current fixtures.
But the “easy gains,” as that same report puts it, are largely gone. The current research frontier is dynamic spectral tuning: rather than running one fixed light recipe for an entire growth cycle, modern systems shift the spectrum delivered at different growth stages, since plants respond to different wavelengths differently depending on their developmental phase. A 2025 study published in Scientific Reports on optimizing LED spectra for controlled-environment vertical farms confirmed that different LED wavelengths carry meaningfully different energy-to-photon conversion efficiencies, meaning the choice of spectral composition directly determines both electricity cost and biomass output, not just crop quality.
The Cross-Domain Connection
The genuinely cross-disciplinary work here pairs plant photobiology with electrical engineering and, increasingly, machine learning-driven control systems. Blue-dominant light during early vegetative growth promotes compact root development; far-red light triggers leaf expansion and flowering at the end of a crop cycle — a recipe-based approach that, according to a 2026 energy sector analysis, can reduce total photon usage by 15 to 20 percent simply by delivering only the wavelengths a plant needs at each specific moment rather than a constant broad spectrum throughout. This requires plant physiologists to characterize exactly which wavelength combinations trigger which developmental responses, and it requires lighting engineers to build hardware capable of actually shifting spectral output on a daily or even hourly basis rather than staying fixed.
A second cross-domain thread involves thermal and electrical systems engineering: because LEDs generate waste heat as a byproduct of light generation, some newer facility designs are integrating that heat directly into HVAC dehumidification processes rather than treating it as pure waste to be cooled away — turning what used to be two separate cost centers, lighting and climate control, into a single more efficient system. Separately, AI-driven control systems are being layered on top of spectral tuning to time electricity use itself: a 2025 study on shifting electricity demand in indoor vertical farms found that managing photoperiod and light intensity combinations while holding the total daily light integral constant could allow facilities to shift consumption toward cheaper off-peak electricity rates without sacrificing plant growth.
What Remains Undemonstrated
The uncomfortable honesty required here is that spectral tuning, while real and measurable, has not yet closed the fundamental cost gap that has bankrupted a string of well-funded companies. Industry accounts of Plenty’s failure noted that its Compton facility still succumbed to a 15 percent year-over-year energy price increase despite the company’s technological sophistication, and a 2025 analysis pointed out that a 2022 USDA report found greenhouses, which use sunlight supplemented by artificial light rather than fully artificial lighting, retain roughly 30 percent lower costs than fully enclosed vertical farms regardless of LED efficiency gains. No published spectral-tuning study has yet demonstrated energy savings large enough, at commercial scale, to have prevented any of the specific bankruptcies the industry experienced in 2023 through 2025 — the technology is improving the margins of a fundamentally difficult economic proposition, not eliminating the gap.
Why It Matters
The stakes extend beyond any individual company’s balance sheet. Vertical farming’s promise, when it works, is real: AeroFarms’ aeroponic approach reportedly uses roughly 90 percent less water than conventional field agriculture for the same crops, and facilities located near urban centers can cut transportation emissions and spoilage for perishable produce. Whether that promise survives as a viable industry, rather than a cautionary tale, likely hinges on exactly the incremental gains that spectral tuning, thermal integration, and demand-responsive lighting represent — combined with geography, since a 2026 energy analysis notes the technology is far more viable in regions with cheap, stable electricity or acute water scarcity than in markets with expensive, carbon-taxed grid power. Singapore’s push toward supplying 30 percent of its nutritional needs domestically by 2030, given how little arable land the country has, is a case where the economics may simply work differently than they did in Compton, California.
The Human Dimension
There’s a certain hard-won humility in an industry that promised to outcompete the sun and is now fighting, watt by watt, wavelength by wavelength, just to get closer to breaking even with it. The engineers optimizing spectral recipes today aren’t chasing a moonshot anymore — they’re doing the unglamorous, incremental work of squeezing a few more percentage points of efficiency out of a technology whose backers once genuinely believed it would remake agriculture outright. Whether that patient optimization is enough to save what’s left of the industry, or whether vertical farming ends up as a valuable niche rather than a revolution, is likely to be decided by exactly this kind of quiet, wavelength-by-wavelength engineering rather than any single breakthrough.
Sources:
1. “14 Vertical Farms Went Bankrupt in 2025. Here’s What Killed Them.,” Foodlore, 2026 — https://foodlore.blog/why-vertical-farms-go-bankrupt/
2. “Has Vertical Farming’s Dream Turned into a Nightmare for Investors,” Kavout Market Lens, 2026 — https://www.kavout.com/market-lens/has-vertical-farming-s-dream-turned-into-a-nightmare-for-investors
3. “Analysis: Five reasons why vertical farming is still the future, despite recent business failures,” UCL News, 2025 — https://www.ucl.ac.uk/news/2025/jan/analysis-five-reasons-why-vertical-farming-still-future-despite-recent-business-failures
4. “Optimizing LED lighting spectra for enhanced growth in controlled-environment vertical farms,” Scientific Reports, 2025 — https://www.nature.com/articles/s41598-025-15352-7
5. “Vertical Farming Energy 2026: The Race for Grid Parity,” Energy Solutions industry report, January 2026 — https://energy-solutions.co/articles/sub/vertical-farming-energy-optimization-market-intelligence-2027
6. “Lessons Learned Overview: Plenty’s Bankruptcy and Implications for Vertical Farming,” iGrow Marketplace, 2025 — https://www.igrowmarketplace.com/post/lessons-learned-overview-plenty-s-bankruptcy-and-implications-for-vertical-farming
7. “AeroFarms turned itself around after bankruptcy. Is it a sign that vertical farming can succeed?” 2025 — https://www.aerofarms.com/aerofarms-turned-itself-around-after-bankruptcy-is-it-a-sign-that-vertical-farming-can-succeed/
Idea originated at artificialideas.org. Article researched and written by Claude Sonnet 4.6. Published at artificialideas.org.
Artificial Ideas 