India’s robotics revolution
From scavenging for parts in the markets of Mumbai to multimillion-dollar investment rounds, India’s robotics companies have come a long way.
Many people dream of a job that lets them travel the world but, after 450 flights within four years, young engineers Samay Kohli, 30, and Akash Gupta, 27, had had their fill of globe-trotting and wanted to put down roots. So six years ago, they co-founded a firm.
And not just any firm. They decided to take a stab at warehouse automation – arguably the most promising space for autonomous robots. The hunch worked. Their company, Grey Orange, is now one of India’s hottest ventures. Kohli and Gupta started small but now enjoy investments from Tiger Global Management, one of the “tiger cubs” or offshoots of Tiger Management, the hedge fund made famous by investment guru Julian Robertson.
Their babies: the Butlers, swarms of cute but smart cube-shaped bots that roam among the shelves to store and retrieve the goods, and the Sorters, which scan and sort all the packages coming in and out of a facility.
The duo bonded as students at the prestigious engineering college BITS Pilani in Rajasthan, over their shared love of robotics. In 2007, they built ACYUT – one of the first humanoid robots in India – and proceeded to tour the world, holding educational workshops and entering – and winning – robotics competitions.
Despite their success, including a gold medal at the 2009 ROBOlympics (now called the RoboGames), they realised that their technical knowhow was in greater need back home – and went into business in their native Delhi. “We had to do something that we liked,” says Kohli. “We liked robots – so we did robots!”
In recent years, India has developed a healthy start-up scene, fed by falling barriers to entry in software development, a flood of venture capital and a healthy supply of talent from the country’s well-developed IT industry. The focus is almost exclusively on software, and the blueprints for success are in fields such as e-commerce and financial technology.
In 2011, though, there weren’t too many start-ups yet, especially not hardware-oriented ones. The pair noticed that the warehousing industry was especially inefficient – or “ripe for disruption,” as Kohli puts it – and decided to improve it. But launching a hardware start-up back then, he adds, was a nightmare. “The ecosystem right from printed-circuit boards (PCBs) making to machine shops, all those concepts didn’t exist. Most of the components you needed, for the first two or three years we were shipping across the world,” he says.
Despite the initial difficulties, in little over five years Grey Orange has evolved into a leading warehouse automation company that serves Indian e-commerce and logistics giants such as Flipkart, Jabong, DTDC and Delhivery. With customers in Japan, Hong Kong, Singapore and Chile, and offices across the globe, it has gone multinational.
Its Butler bot is superficially similar to the one made by Kiva Systems, which was brought in-house by Amazon in 2012. Relying on machine learning, swarms of Butlers work together to determine the most efficient way to fetch products and bring them to workers at picking stations. The system simultaneously optimises storage locations based on current and past order data.
In a warehouse, on a typical day workers pick about 100 items per hour; by using Butlers, it’s possible to boost this number to 500-600 items an hour.
Then there are the Sorters. These machines use conveyors and scanners to read package barcodes, measure their dimensions and weight, automatically space them, and finally sort them for distribution. Warehouse automation is not new, but Kohli says what singles them out is their approach – uniting picking and sorting under a single overarching intelligent system.
Building from scratch
But Kohli and Gupta weren’t the very first ones on India’s burgeoning hardware start-up scene, though. Go back another few years, and the situation was even worse, according to two other engineers who studied together at the Indian Institute of Technology (IIT) in Bombay. When Gagan Goyal started his educational robotics firm ThinkLABS in 2005 (see box on page 30), even cobbling together a prototype was a challenge.
“We used to go to local markets to get second-hand material. We’d have to scrap things like motors from toy cars,” he says. “We’d speak to the people who made TVs in India because they do work on PCBs and we had to convince them to make just 1,000 units for us.”
His classmate Ankit Mehta founded ideaForge, and developed India’s first drone in 2004 while at IIT Bombay. Mehta claims he and his team stumbled over the now ubiquitous quad-rotor design independently. He says drones were barely on the radar in the West, let alone in the technological backwater India was at the time. They only realised others were working on similar designs when they attempted to file patents.
“It’s been a very arduous journey for us,” says Mehta. “In the early days, we literally had to build our own hardware ecosystem. We had to go and scavenge in the deepest, darkest lanes of the city to find components.”
Luckily, though, their work caught the eye of India’s Defence Research and Development Organization, and in 2009 they delivered what was at the time the world’s smallest and lightest UAV autopilot. This was followed swiftly in 2010 by the country’s first autonomous UAV – the NETRA quad-copter – and the company has now sold more than 400 drones to the country’s armed forces and internal security agencies.
The struggles these founders went through highlight the chicken-and-egg situation that India’s high-tech hardware companies have faced, says Jagannath Raju. A lack of high-profile success stories makes investors and component manufacturers unwilling to take a gamble, he says, which makes starting a hardware company all the more difficult.
Raju would know better than anyone, having founded his robotics firm Systemantics in 1995 after returning from the US. After the company started out consulting for various government agencies, designing everything from inspection robots to underwater vehicles, the Department of Science and Technology recognised the technical depth of Systemantics’ work and provided some soft financing to develop a robotic arm in 2000.
There was muted interest from industry, though, and the next decade saw the company surviving hand to mouth on often tardy government funding. Then in the late 2000s venture capitalists started paying attention to the Indian market, and in 2010 Systemantics received its first significant funding from Accel.
This allowed the firm to create a series of robotic arms, designed to target smaller businesses in emerging markets, unable to shell out for top-of-the-line industrial robots. “We thought the most appropriate way to disrupt things was on price point and not worrying about catering for someone looking for very precise and very high-speed solutions. That’s maybe 10% of the applications, but another 90% is looking for a much lower-cost and lower-spec kind of product,” says Raju.
While component prices have fallen dramatically in recent years, thanks mainly to cheap imports from China, most are general purpose. “You pay for 100% of the features but you only end up using 25%,” says Raju. To meet its ambitious price point, for the past five years the firm has been working with local suppliers to develop everything from motors to computer boards tailored to its needs.
These efforts may have a positive knock-on effect for start-ups following in Systemantics’ footsteps, because Raju has recognised that securing the firm’s supply chain relies on vendors having enough demand. “We very specifically say that you will supply to us and you will sell it to as big a market as possible,” he says. “No licensing, even if we help you with some money, some design ideas, you are open to do whatever you want as long as you supply our needs.” Raju has even pointed other robotics start-ups towards Systemantics’ vendors.
Ultimately, though, Raju thinks it will take some high-profile successes for Indian industry to get fully on board. Karthik Reddy, co-founder and managing partner of Blume Ventures, which has invested in both Systemantics and Grey Orange, agrees. He is hopeful that Grey Orange could be the breakthrough success that catalyses India’s robotics ecosystem.
“That’s the dream we have for that company,” he says. “It’s literally one of the best-performing companies we have, if not the best. There’s a long way to go, but in two or three years it can become an inspiration for a lot of youngsters to say we can do this.”
Another project that could inspire a generation of Indian engineering graduates is the country’s first privately funded space mission. Team Indus is one of five finalists in the Google-funded Lunar XPRIZE competition. It has challenged start-ups to land a robotic spacecraft on the Moon, travel 500m across its surface and send HD video to Earth before the end of 2017 – for a grand prize of $20 million.
Team Indus is booked onto an Indian Space Research Organisation (ISRO) rocket due to launch on 28 December and is entering the final stages of testing. Its workforce is conspicuously polarised between senior ex-ISRO scientists and engineering graduates fresh out of university. He’s only 24, but Karan Vaish is already heading the development of the team’s lunar rover.
“It’s great, it creates an image that the guy next door can do this,” he says, adding that opportunities like this, and the high-profile success of start-ups in the software space, is leading to a shift in attitudes in his generation.
“All my friends are breaking out of these hard-core clichéd jobs,” adds Vaish. “They’ve seen organisations grow from very small to a massive scale and they can just see it is actually possible, so they want to take the leap of faith and do it themselves.”
Confidence will only get you so far, though, as many of India’s first wave of start-ups are beginning to find out. The heady optimism of a few years ago has faded as heavily-backed companies struggle to achieve profitability and investors begin to exercise more caution. In hardware, the challenges are even greater.
“One of the difficulties is that you don’t have indigenous talent pools around in large numbers to say we’re suddenly going to build an industry. Developing that is a very slow process,” says Blume’s Reddy. While there is plenty of raw engineering ability among graduates, robotics is a complicated synthesis of multiple fields. And unlike the software industry, the hardware ecosystem hasn’t always got the resources to turn raw talent into productive employees.
Hardware and IoT
Reddy says that most of the Indian robotics start-ups are focused on areas such as retail, services and consumer tech, with little domestic demand. He doesn’t see any Indian robotics company becoming successful without first cracking its home market. “I don’t think there’s enough understanding of where the industry solutions are required,” he says. The real opportunities are in industrial applications, he adds, and will require a combination of hardware, sensors and smart analytics.
Avinash Kaushik, the founder of India’s first hardware accelerator Revvx, agrees. He says India’s IT industry has excelled in building end-to-end solutions for enterprise customers, and the hardware industry needs to follow suit.
That will mean focusing on the problem at hand, not whether you’re a hardware or a software start-up, and being willing to combine all kinds of technology from robotics to Internet of Things and machine learning.
“It’s a convergence of different technologies,” says Kaushik. “Just the hardware is not of much value in the long run. You need to build smart solutions, which can adapt, can monitor the surroundings, take action and be contextually aware.”