Nothing hard gets done by one player. A new material needs a lab, a pilot plant, a first customer, a lender and a standard, and each has to hand off cleanly to the next. Miss one layer and the whole thing fails. We spend the family's capital, and our time, on building those layers and connecting them.
What we do.
- We bring capital, what we have learned and the people we know to the same problem, and we use equity, debt or grants, whichever it actually needs.
- We connect founders to people, lessons and relationships across the ecosystem: a lab, a pilot plant, a first customer, a lender.
- We try to get founders and their teams to the inflection point faster than they would get there alone.
- We go deeper than financing: more capital when it is actually needed, coaching, and the right call when it is not the easy one.
What we do not do.
- We do not take other people's capital. The capital is the family's own.
- We do not invest in other funds as an LP. We back companies and partner organizations ourselves.
- We do not decide for the founder. We bring an operator's mindset and say what we think is possible. The founder decides.
- We do not take grant applications. Our grants go to partners we find through our own work.
Hard problems are stacks.
Think of what it takes for one child in India to survive a serious illness: someone who notices, a flight that exists, a scanner, a trained surgeon. No single hero, a stack of things that all work. Climate technology is the same shape. Layers that must exist, and someone who connects them so value can flow.
We call that work orchestration, and in Indian deep tech it is rarely anyone's job.
The right capital for the right problem.
Profit is not a dirty word in impact. It is a tool, and so is its absence. Profit can solve a large share of problems. Some it cannot, and that is where grants belong.
So we use the whole spectrum, and we try not to confuse the tools.
-
Grants
Pays forWhat nobody can make a profit on: data, standards, policy, research, the first experiment.
ProofClimate data work with CEEW; the efficiency standards behind appliance star labels.
-
Equity
Pays forR&D, the team, the first plant.
ProofOur first cheque into Atomberg, about ten years ago.
-
Structured finance
Pays forFirst of a kind plants: philanthropic capital takes the first loss, equity sits on top, and most of the money can then be debt.
-
Debt
Pays forWorking capital and assets, once there is something to lend against.
ProofLoans to two young companies financing commercial electric three wheelers, before any bank would.
-
Bankable. Then we leave.
Debt should be accessible, not cheap.
Most impact lenders solve for affordability. We solve for access. Cheap debt ties a young company's business model to a price that no commercial lender will ever match, and the model breaks the day the impact money runs out. So we lend where nobody else will yet, at a price close to where the market will land, and the company can walk from us to a commercial lender to a bank.
A technology is solved when any bank will lend against it.
Lending is also how the data gets made. We lent to two young companies financing commercial electric three wheelers when nobody knew how long the batteries would last. The repayment record from those loans is what let banks and NBFCs come in after us. And debt changes the shape of a business: a company can stop selling equipment and start selling a service, so its customer never has to bet on a new technology up front. The same logic runs through BiofuelCircle and Ishitva Robotic Systems.
It is a relay race.
We learned this in chemistry.
A petrochemical company
supplied a raw material.
Aarti Industries
used it to make a specialty molecule in India for the first time.
An agrochemical company
used that molecule to make a product in India for the first time, and then exported it.
Three runners, and none of them could have won alone. In a new technology the runners might be a venture funded startup, a family business and a multinational.
The real question is whose job it is to get all three to the same race.
Testbeds are a public good.
To prove a new battery chemistry or a new membrane you need a pilot facility, and no startup can raise the money for a pilot facility to test one idea. So many good ideas are uninvestable for a reason that has nothing to do with the idea. Shared piloting and prototyping infrastructure is a public good: government, philanthropy and grants build it, and industry commits to use it. That is why our team helped start ITRI, India's Translational Research Initiative, and why the family funds the Gogri Hub for Membranes Research at IIT Bombay.
The same gap shows up one step later, at the first commercial plant, the one nobody wants to fund. Once it runs, everyone comes.
Someone has to take that first risk.
Generation before efficiency.
Efficiency is real, and it works at the edges. The large, ugly part of the problem is where the energy comes from, and where the calories come from. So we back always-on power that can sit under solar and batteries: thorium reactors at Copenhagen Atomics, deep geothermal drilling at Quaise Energy, fusion at HB11 Energy. And we back new ways to make protein, from New Harvest to Phyx44, because land released from the food system is one of the few levers big enough to bring the emissions curve back down.
Climate is a material and energy problem at the end.
Sector by sector.
Climate is not one problem. The limiting factor in solar has nothing to do with the limiting factor in steel, so we take it sector by sector.
Solar
The technology is there. What holds it back now is permitting and the speed of deployment on the ground. The next jump is perovskites, and those need testbeds.
Storage
Lithium wins vehicles. We doubt it wins the grid. Sodium ion is the most promising path.
Trucks
Batteries will get there. The operator chooses the truck, but the government chooses the charging, and charging needs one common standard. That standard is grant work.
Shipping
Someone has to pick the fuel standard, then every port has to build for it. Large chemical plants, not startups. Not everything in the transition is an innovation.
Steel and concrete
The two we are most cautious about. In concrete, part of the CO2 comes from the reaction itself.
Efficiency
Easy when the buyer does the arithmetic, hard when the buyer is a household. So we fund the work behind the star label.
Food
Match taste and texture exactly, at the same cost. Start with the hardest one, chicken.
Materials
They sit at the back of every value chain, where nothing is visible. That is why the field is less crowded, and why we look there: NoPo Nanotechnologies, Thermistance.
We start from the problem.
We keep a map of where the world's emissions come from, and under it, the solution spaces, and under those, the problems inside each solution. It lets us place a company from its first sentence. The map gives us red lines: we will not fund hydrogen in cars. It also gives us conviction early: nobody had to convince us about lightweight materials.
Then we read the papers and have the technical conversation, because these fields are not crowded and somebody has to go deep.
Beyond climate.
The first stack on this page was a health problem: someone who notices, a flight that exists, a scanner, a trained surgeon. Health, education, livelihoods and civic work are built the same way, in layers.
Profit can pay for some of those layers. The ones nobody can make a profit on, like data, standards, policy and research, are where grants belong.
How we work is from Spectrum Impact's own statement of who we are, 2026. The beliefs and the sector ledger are drawn from The Spectrum Impact Story, a profile by Priyadarshan Banjan for materials.club, July 2026.
Read the full profile (opens in a new tab)Essays by Mirik Gogri
Building a company?