Startup Deep Dive : Promethean Power Systems — the MIT solar project that became a diesel-free milk chiller

Promethean Power Systems set out to sell solar power to rural India and ended up abandoning solar entirely. The pivot paid off: its thermal-battery milk chillers now cool raw milk for roughly 18 paise a litre against about 50 paise on a diesel generator, and the company says its equipment has touched the incomes of more than 150,000 dairy farmers.

That reversal is the whole story. A hardware startup founded by two MIT men in 2007 spent its first four years building the wrong product, ran its idea into the ground, and only found a business when it threw the sun overboard and kept the battery. Nearly two decades on, the India arm still turns over less than the price of a single Mumbai penthouse, its balance sheet leans on impact-debt lenders rather than venture capital, and its revenue actually fell in its most recent filed year. This is a deep dive into how a scrapped clean-energy science project became a working cold chain, and why staying small has been both its discipline and its ceiling.

Quick facts

What they do

Promethean makes refrigeration for places the grid forgets. Its core product chills raw milk at the village collection point within roughly two hours of the animals being milked, dropping it from around 35°C to about 4°C, without a running diesel generator and without needing continuous mains power. The trick is a thermal battery: a pack of phase-change material that is frozen while power is available and then releases that stored cold to chill milk during the early-morning and evening collection windows, when the grid is often down. From that base the company has widened its range:

• Rapid milk chillers and bulk milk coolers for dairy collection centres, sized up to about 1,000 litres a day (MIT News, September 2015).
• Cold rooms and refrigerated transport for meat, fish and fruit and vegetables, added from 2018 (Greentown Labs, October 2019).
• Software for milk procurement and remote equipment monitoring (company website, 2026).

The customer is the dairy processor or the farmer co-operative, not the individual farmer. Named users include Amul, Hatsun, Mother Dairy, Nestlé, Parag, Heritage, ITC and Bangladesh’s BRAC (Forbes India, April 2019; company website).

The origin

The company began as an MIT student project, not a dairy plan. Sorin Grama, an electrical engineer from Romania with a master’s in engineering and management from MIT, joined a team building a solar microgenerator for the developing world: an array of mirrors that would track the sun, focus its heat and boil a refrigerant to make power. The idea won prizes. It took third place at MIT’s 2006 Ignite Clean Energy competition and second in the energy track of the 2007 MIT $100K contest (MIT News, September 2015). Through that circuit Grama met his co-founder, Sam White, an American with a commercial bent, and the two set out to find a market.

They went to India in 2007 to sell the generator and came back convinced they had the wrong product. On the ground they kept running into the same problem: milk. India collects enormous volumes of milk from tiny farmers, but the liquid sits warm in cans for hours before it reaches a chilling centre, and up to 30% of the value can be lost to spoilage (Acumen). A solar generator did not fix that. So Grama and White founded Promethean late in 2007 and set out to build a milk chiller from scratch. As Grama has put it, the thermal battery was what “rose from the ashes of all those mistakes” (MIT News, September 2015).

The struggle years

The founding insight was right and the first execution was wrong. Promethean’s early chillers were solar-powered, which sounds obvious for rural India until you notice when milk is actually chilled. Collection happens in the early morning and the early evening, largely outside the hours when a solar panel produces useful power. The product could not do its main job at the moments it was needed, and solar hardware was expensive for a price-sensitive market.

The pivots stacked up over years, not months:

• 2007: the original solar-thermal generator for schools and clinics is judged a poor fit for India and shelved (Business Standard, January 2017).
• Late 2007 onward: the team rebuilds around milk chilling, still trying to keep solar in the loop.
• 2011: Promethean starts developing a thermal battery to store energy so chilling can happen off the solar window (MIT News, September 2015).
• Eventually the solar component is dropped entirely as “too expensive and impractical,” and the whole system is rebuilt around the battery (MIT News, September 2015).

Those were expensive lessons for a hardware startup with little capital. The founders relocated within India to chase sales, and the co-founders’ own account is blunt about a “succession of failures and setbacks” before anything worked (Business Standard, January 2017). The company that survived was, in effect, the opposite of the one it had pitched: a battery business that had quietly deleted the word solar from its core product.

The turning point

The break came in early 2013, when Promethean landed its first large commercial order: 50 milk chillers for Hatsun Agro, India’s largest privately owned dairy processor (World Economic Forum; MIT Technology Review, January 2013). A single reference customer of that size did two things. It validated the thermal battery in the field, and it gave a small importer of an unproven idea a name that other processors recognised.

The numbers on either side of that order show the shift. Before it, Promethean was a prototype shop with pilots. After it, deployment compounded:

• About 100 chillers installed by September 2015 (MIT News).
• Around 330 units installed and roughly $3 million in annual turnover by April 2019, with exports to Bangladesh, Sri Lanka and Tanzania (Forbes India).
• About 60,000 dairy farmers served and roughly 3 million litres of diesel displaced by October 2019 (Greentown Labs).

The Hatsun deal did not make Promethean big, but it made it real. It converted a science story into a supplier relationship, and every later claim about farmers reached and diesel avoided traces back to that first fleet going to work.

The money behind it

Promethean has never raised like a typical venture-backed startup, and its funding shape tells you what kind of company it is. Equity has been modest and impact-led; the heavier lifting has been done with debt from social-finance lenders.

• Equity, total: about $1.87 million across seven rounds from a reported 12 investors (Tracxn, 2026).
• Acumen: $1 million of preferred equity in 2019, structured to fund the capital expenditure behind a subscription model and to hire a services and technology team (VCCircle, October 2019; Acumen). Greentown Labs reported the same month that the company had “raised most of its Series B” and was leaving about $1 million of the round open (Greentown Labs, October 2019).
• Tech Emerge and others: listed among backers alongside Acumen (Tracxn).

The debt side is where the real money moved, and the lender list reads like a who’s who of Indian impact finance. Charges filed against the India entity include (Tracxn; TheCompanyCheck, MCA filings):

• Caspian Impact Investments — several charges between September 2018 and December 2023, from ₹53 lakh to ₹4 crore.
• UC Inclusive Credit — ₹5 crore, September 2019.
• Grameen Impact Investments India — ₹1 crore, December 2020.
• Samunnati Agri Value Chain Solutions — ₹1.5 crore (December 2021) and ₹2 crore (January 2024).
• Akshayakalpa Farms and Foods — ₹8 crore, March 2025.

As of the latest data, the India entity carries about ₹20.5 crore of open charges against ₹8.53 crore satisfied (TheCompanyCheck). That is a lot of borrowing relative to a paid-up capital of just ₹13.4 lakh and revenue under ₹30 crore, and it is the clearest signal that Promethean funds its growth with lender balance sheets rather than equity dilution.

How it makes money

Promethean earns in two overlapping ways, and the mix has shifted over time from selling boxes to selling cold as a service.

• Equipment sales and leases: the traditional model, where a dairy buys or leases chillers, coolers and cold rooms. Revenue is lumpy and tied to processors’ capital-spending cycles.
• Subscription / pay-as-you-go cooling: introduced in 2018, where Promethean owns the hardware and the customer pays for use. This is what the 2019 Acumen equity was raised to fund, because the company must lay out the capex up front and recover it over years (Greentown Labs; VCCircle, 2019).

The economics that make either model work are the running cost. Promethean’s chilling costs roughly 18 paise per litre, against about 50 paise per litre for a diesel-run chiller (Forbes India, April 2019); Acumen puts the saving on remote refrigeration at about two-thirds. That gap is the product. The thing people get wrong is assuming this is a climate story first — the diesel it avoids is real, but the buyer signs because the milk that used to spoil now survives the journey, and because the fuel bill falls. The margin sits in the hardware and the service contract; the moat sits in the patented thermal battery and the field network that keeps it running.

The numbers

Only one recent year of the India entity’s revenue is available for free from filings, and profit and loss are locked behind paid MCA reports, so they are not stated here rather than guessed. Some databases show apparent multi-year figures that are internally impossible (net profit larger than revenue, three-digit margins); those are paywall placeholders and have been ignored. What can be stood behind, in ₹ crore, is a revenue line that grew for a decade and then dipped:

Supporting figures: the India entity employed 44 people as of April 2025 (Tracxn); the US parent reported about 53 employees in 2023 (Tracxn); the auditor of record is S.M.Suratwala & Co. The headline to hold on to is that FY2025 revenue of ₹28.65 crore was down 7.0% on the prior year — a company that had spent a decade going up went slightly backwards.

Where the money comes from

Promethean’s revenue concentrates around dairy, large processors and a handful of geographies, with a spreading edge into other cold-chain uses.

• By product: milk chilling remains the core; meat, fish, fruit and vegetable cold rooms and refrigerated transport are the newer, smaller lines added from 2018 (Greentown Labs).
• By customer: a short list of big buyers — Amul, Hatsun, Mother Dairy, Nestlé, Parag, Heritage, ITC — drives volume, which concentrates risk in a few relationships (Forbes India; company website).
• By geography: mostly rural India, with exports to Bangladesh, Sri Lanka and Tanzania (Forbes India; Times of India, January 2017).

The surprise is the impact base underneath a small revenue line. The company states its equipment has affected the income of more than 150,000 farmers, empowered more than 3,000 women and eliminated more than 3 million litres of diesel (company website, 2026). Acumen’s older count was “over 40,000 dairy farmers.” A firm with revenue under ₹30 crore claiming six-figure farmer reach tells you the unit economics are built on many tiny users sharing one machine — about 20 to 30 farming families per chiller (MIT News, 2015) — not on a few big-ticket sales.

The risks

• Revenue is going the wrong way. FY2025 revenue fell 7.0% to ₹28.65 crore (TheCompanyCheck). Hardware demand is tied to dairy processors’ capex cycles, so a single slow year of orders shows up immediately in the top line. A subscription model smooths this only after years of upfront capex.
• The balance sheet leans on debt. About ₹20.5 crore of open charges sit against paid-up capital of ₹13.4 lakh and revenue under ₹30 crore (TheCompanyCheck; Tracxn). Servicing impact-debt from lenders such as Caspian, Samunnati, Grameen Impact, UC Inclusive Credit and Akshayakalpa is a fixed cost that a lumpy revenue line has to cover regardless of the order book.
• Working-capital stress signals. Tracxn records 42 delayed MSME supplier payments for the India entity in the April to September 2025 window (amounts not disclosed). For a capex-heavy hardware firm, delayed supplier payments are a classic sign of cash being tight, and they compound with the debt load above.
• Customer concentration. A short roster of large dairies drives volume; losing one big processor, or a pause in a co-operative’s expansion, would hit revenue hard.

The takeaway

The transferable lesson is not “pivot.” It is that the founding insight and the founding product are different things, and confusing them can cost you years. Promethean was right about India’s problem from its first trip in 2007 — milk spoils before it is chilled — and wrong about the answer for the next four years, because it was in love with solar. The company only worked once it kept the insight and killed the technology it had started with. The battery it now sells is the residue of admitting that. For anyone building hardware for a hard market, that is the discipline worth copying: hold the problem tightly, hold your first solution loosely, and be honest about the day the thing you were proud of is the thing holding you back.

Frequently asked questions

What does Promethean Power Systems make?

Off-grid refrigeration built around a thermal battery: rapid milk chillers and bulk milk coolers for dairy collection points, plus cold rooms and refrigerated transport for other perishables. The battery stores cold when power is available and releases it during collection hours, so chilling works without a running diesel generator.

Who founded it and when?

Sorin Grama and Sam White founded the US parent in 2007 after meeting through MIT’s entrepreneurship competitions. India operations run through Promethean Spenta Technologies Private Limited in Pune, which now trades as Coolectrica.

How much money has Promethean raised?

Equity is modest — about $1.87 million across seven rounds per Tracxn, including $1 million of preferred equity from Acumen in 2019. Most of its growth capital has come as debt from impact lenders, with about ₹20.5 crore of open charges on record.

What is its latest revenue?

The India entity reported revenue of ₹28.65 crore for the year ended 31 March 2025, down 7.0% year on year, per MCA filings summarised by TheCompanyCheck. Audited profit or loss is not available in free filings.

Why did it drop solar power?

Milk is chilled in the early morning and evening, largely outside the hours when solar panels generate useful power, and solar hardware was too costly for the market. Promethean kept the idea of storing energy but moved it into a thermal battery and dropped the panels.

Sources

Figures are as of September 2026. Currency converted at $1 ≈ ₹96.0 as of 18 September 2026 (Trading Economics).

• TheCompanyCheck — Promethean Spenta Technologies Private Limited profile and FY2025 financials (data updated February 2026)
• Tracxn — Promethean Spenta Technologies legal-entity and Promethean Power Systems company profiles (2026)
• MIT News — “Saving rural India’s ‘liquid cash'” (September 2015)
• MIT Technology Review — “Novel Energy Storage to Chill Milk in Rural India” (January 2013)
• Forbes India — “Promethean Power: A chill pill for dairy farmers” (April 2019)
• Business Standard — “Eventual sweet taste of success” (January 2017)
• VCCircle — “Impact investor Acumen bets on refrigerated solutions firm Promethean Power” (October 2019)
• Acumen — Promethean Power Systems company and case-study pages
• Greentown Labs — “Promethean Power Systems brings sustainable dairy chilling to over 60K dairy farmers” (October 2019)
• World Economic Forum — “How milk-chillers could revolutionise India’s dairy industry”
• Times of India — “Promethean Power Systems to expand into Sri Lanka and Bangladesh” (January 2017)
• Promethean Power Systems / Coolectrica — company website (2026)

Found an error? Write to us and we’ll correct it in the open, dated, on the piece.