India is short of skilled welders, and Terafac is teaching robots to fill that gap
The Chandigarh-based startup's AI vision system turns off-the-shelf industrial robots into welders that adapt to changing parts without needing reprogramming.
Industrial robots are precise but rigid. A robot arm on a factory floor repeats the same movement down to the millimetre, which works well when every part coming down the line is identical.
Indian manufacturing often does not work that way. Smaller factories handle short runs and frequent changes, and each change means reprogramming the robot. That takes a specialist, and it takes an amount of time the factory does not have the privilege to provide.
The result is that most small and mid-sized Indian manufacturers do not automate at all. They rely on skilled workers instead, in trades where those workers are increasingly hard to find.
, a Chandigarh-based startup, sells software that aims to remove the reprogramming step. It was founded in 2021 by Anubhi Khandelwal, its CEO, and later in 2023, Amrit Singh joined as the co-founder and CTO.
Building it from Chandigarh rather than a metro puts the company close to the manufacturing clusters of Punjab and Haryana, where the customers it is aiming at actually operate.
Software that watches the part
Terafac's platform uses computer vision to give an ordinary industrial robot a way to see what is in front of it. The robot scans the part, works out what needs doing, and adjusts its own movement. No one writes new instructions for each variation.
Its first product, WeldT, applies this to welding. Welding is a good starting point because it is a skilled, repetitive and physically demanding task, and India has a severe shortage of trained welders.
The company calls its approach ‘Software as a Skill’. Rather than selling a robot, it sells a capability that runs on machines a factory may already own. That keeps the cost of adopting it much lower than buying new hardware.
Terafac plans to extend the same platform to painting, glueing, inspection and assembly, which are variations of the same underlying problem.
Each of those tasks involves a robot doing skilled physical work on a part that is not always identical. If the vision layer can handle one, the argument goes, it can be adapted to the others without starting again.
Aimed at India's smaller factories
The company sells to original equipment manufacturers and to factories directly, with system integrators as partners. Its stated target market is India's small and mid-sized manufacturing base, which has largely been priced out of automation.
Working through integrators matters for reach. Those firms already install and service equipment for hundreds of factories, and they are the route through which a small software company can get in front of buyers it could never call on directly.
Terafac raised Rs 6.5 crore in a pre-seed round led by Inuka Capital, with DeVC, Bharat Founders Fund, Innovation Mission Punjab and several angel investors. The money is going into product development, hiring and market expansion.
Raj Sheth, founding partner at Inuka Capital, said the company was making automation accessible for factories of all sizes by upgrading machines they already have, starting with welding, where the shortage of skilled workers is most acute.
The company has said it aims for 80 deployments across more than 20 customers within a year of the round. Whether it reaches that would be the test of whether the model works outside a pilot.
Terafac is competing against established industrial automation suppliers, most of them selling hardware. Its position is that a factory should not have to replace its machines to make them adaptable, which is a different proposition from the one those suppliers offer.
(This story has been researched and compiled using publicly available information.)