Why India's robotics firms are integrators, and how they get out of the middle
SRF Capital Studio Research DeskFunding Intelligence, SRF Capital StudioMost Indian robotics companies are system integrators, which puts them in the least profitable layer of the chain. That is not a verdict. It is a starting position with three ways out.
Summary
- A factory robot passes through five layers: components, robot makers, system integration, software and service. Integration, where most Indian firms operate, typically earns the lowest margin.
- India sits there for sound reasons: the precision parts are controlled by a few Japanese suppliers, and integration needs engineering talent, which India has, rather than capital, which it lacks.
- The founders who escape the middle productise their cells, attach service contracts to every installation, and build software on their installed base. Building robot arms is rarely the answer.
Walk around any automotive cluster in India and you will find dozens of small automation firms with a few engineers, a workshop and a list of cells they have commissioned. Most describe themselves as robotics companies. In value-chain terms, nearly all of them are system integrators, and that position shapes their economics more than anything they do day to day.
The five layers
It helps to follow a single robot from factory to factory floor.
Where the margin sits in the industrial robotics chain
| Layer | Who does it | What earns the margin | Indicative gross margin |
|---|---|---|---|
| Precision components | Nabtesco, Harmonic Drive Systems, Sumitomo; FANUC and Yaskawa in-house | Scarce manufacturing know-how, few credible suppliers | 60 to 70% |
| Robot makers | Fanuc, abb, Yaskawa, KUKA, Kawasaki, Epson | Brand, controller software, global service | 25 to 35% |
| System integration | Regional firms, OEM partner networks | Engineering hours on each project | 15 to 25% |
| Software | Platform vendors, OEMs, specialists | Code that is written once and sold many times | 60 to 80% |
| Service and spares | OEMs, integrators, service specialists | A captive installed base | 45 to 60% |
The pattern is clear even with ranges this wide. The layers that sell something made once and used many times, components, software and service, keep the most of each rupee. The layer that sells engineering hours project by project keeps the least.
Why India ended up in the middle
India's robot fleet reached 52,570 units in 2024, according to the International Federation of Robotics, and nearly every one of those arms was imported. That is not an accident of policy. It follows from where the hard parts are made.
The gearboxes in a robot's joints, the precision reducers, are made by very few companies. Nabtesco estimates it supplies about 60% of the heavy RV reducers used in medium and large arms, and Harmonic Drive Systems leads the smaller wrist reducers. An Indian firm that wants to build arms either buys these parts at prices set by suppliers who sell much larger volumes to its competitors, or spends many years trying to make its own.
Integration faces no such barrier. It needs people who can design a cell, program a path, specify a gripper and commission a line, and India trains such engineers in large numbers at modest cost. So Indian automation firms grew in the layer where their advantage was real. That was the right call. The problem is what happens next.
Indian automation firms grew in the layer where their advantage was real. The problem is what happens next.
The squeeze on integrators
An integrator's economics tend to worsen as it grows unless something changes. Each project is partly custom, so engineering time scales with revenue. The robot maker controls the price of the largest component and can reprice it. Customers pay in milestones that often arrive months after the integrator has paid for hardware. And good engineers, once trained, are exactly what larger competitors want to hire.
The result is a familiar shape: a firm at ₹30 to 80 crore revenue with decent margins on paper, rising receivables, a founder still personally closing most deals, and no asset a buyer would pay a premium for. We have seen the same ceiling in other Indian project businesses, and wrote about it in where the founder's gut stops working.
Three ways out
Productise the cell. Pick the two or three jobs you do most often and turn each into a standard product: a defined machine-tending cell, a welding station for one family of parts, a palletiser for one carton range. Fix the scope, the price and the delivery time. A standard cell can be sold by a salesperson who is not an engineer and installed from a manual by a technician. It is the single most effective step out of founder dependence.
Attach service to every installation. Each commissioned cell is a customer who will need preventive maintenance, spares, reprogramming for new parts and operator training. An annual contract turns that into recurring revenue at far better margins than the project. Measure the share of installed customers on contract and push it up before chasing the next new logo.
Build software on the installed base. Once you have cells in dozens of plants, simple monitoring and analytics, uptime, cycle time, weld quality, reject rates, become a product in their own right. Addverb, now majority-owned by Reliance, shows the logic at larger scale: it built its own warehouse robots and the software that runs them, rather than reselling imported systems.
What not to do
- Don't try to build a general-purpose robot arm to escape integration margins. The component economics are against you and the incumbents are decades ahead.
- Don't compete with Chinese hardware on price. Choose the customers for whom downtime is expensive and sell them uptime.
- Don't spread across every vertical. Application depth in one industry, pharma packaging or two-wheeler welding, is what lets you charge for expertise.
- Don't let milestone terms drift. An integrator that grows fast on 150-day receivables can run out of cash while profitable.
What investors should look for
For anyone backing an Indian integrator, the useful questions follow from the above. What share of revenue comes from standard offers rather than bespoke projects? How many installed customers are on service contracts? Is there software revenue, and is it actually paid for rather than bundled free? How long is the cash conversion cycle? A firm improving on all four is climbing out of the middle. One growing only on project revenue is buying a bigger version of the same problem.
For the full chain, including the FANUC and ABB numbers behind these margin layers, see our report on robotics in manufacturing in India and our piece on robotics industry margins. Founders planning this transition can talk to us about strategy consulting for automation businesses.
How useful was this article?
One tap. It tells us what to write more of.
About the author
SRF Capital Studio Research Desk
Funding Intelligence, SRF Capital Studio
The next one
Get what we publish next, by email.
Working notes on raising, borrowing, protecting, growing and structuring capital in India. One email a week at most, and you can leave any time.
We use your address only to send this. See our privacy policy.
We store your address to send you these emails and nothing else. See our privacy policy.
Related reading