Awais Ahmed was just 19 when he was part of the first Indian team that built a pod to participate in the second edition of the SpaceX Hyperloop Pod Competition (2017). The team came within the top ten and Ahmed got to take a selfie with Elon Musk. That excitement was enhanced when he got to tour SpaceX’s factory in Los Angeles. When he got a ‘touch and feel’ of the rocket engines, Ahmed decided his career lay in space.
After completing his master’s in mathematics at BITS, Pilani, Ahmed turned to the space sector and found there were already a few companies making rockets. He further learned that there was a premium market for high-resolution satellite images. He joined hands with classmate Kshitij Khandelwal to set up Pixxel Space, a satellite company, in 2019.
When it launched its first satellite in 2021, Pixxel was the first ‘hyperspectral’ imaging company in the private sector.
Ahmed spoke with Quantum about Pixxel and its plans. Excerpts:
You have raised $71 million so far. What are your plans?
We have so far launched three satellites. They are demo satellites for us. But a couple of hyperspectral companies have come up in the last three years — our demo satellites were better than their commercial satellites. But we called them ‘demo satellites’ because they do not do daily revisits anywhere and they give 10-m resolution rather than 5-m. But we were able to prove the concept of hyperspectral, and also sell data from them.
Next year we will launch six satellites — we have named them ‘Firefly’. They will be our flagship commercial satellites. These six will give us daily revisits anywhere on earth. Next we will launch 18 more satellites in 2026 and 2027.
To whom do you sell the data? Do you have anchor customers?
We already have about 60 customers globally — oil and gas, mining companies — they are buying data from our demos and will buy data from our Firefly satellites. We have some big government customers also. NASA is a customer, the Indian Air Force is another.
What are the features of these satellites?
The first three will be 50 kg each and the rest 200 kg. They will be both optical and infra-red.
Not SAR (synthetic aperture radar)?
No, not SAR. There are problems in putting both optical and SAR in the same satellite. Radar is always looking at an angle of 30-40 degrees, whereas optical is looking straight down. If the optical looks at 30 degrees, the image will not be clear. You can do optical and SAR in different satellites.
We have a contract with the IAF to build for them a multi-payload, modular satellite. The 150 kg satellite will have four payloads — an electro-optical camera, a hyperspectral camera, a thermal camera and an SAR. But here we are not co-collecting images. The idea is to use optical when there are no clouds and SAR when there is.
So, we can build these (SARs), but right now we are focusing on hyperspectral because globally there is a gap — there are companies doing well in SAR and in optical but none in hyperspectral. We want to be the global leader in hyperspectral. When we launched our first hyperspectral satellite in 2021, we were the world’s first commercial (non-government) hyperspectral satellite company.
Why hyperspectral?
Think of hyperspectral as the next evolution of multi-spectral. Multi-spectral cameras can do imaging in 4-5 wavelengths, but the definition of hyperspectral is minimum 40 wavelengths.
So, the world is moving towards hyperspectral imaging?
There are, maybe, three other companies doing hyperspectral in the world — none in India. There is a Canadian company called Wyvern. But they are doing in 32 wavelengths, which is not quite hyperspectral because the definition is 40 wavelengths — but you can call it ‘superspectral’. Then there is a company called Orbital Sidekick. We believe the resolution of our demo satellites is the same as theirs. And Planet Lab has just launched its first hyperspectral satellite.
Your satellites can produce images of 5 m resolution. But the definition of a high-resolution image is 30 cm resolution.
At 30 cm it would be state-of-the-art. Anything below 80 cm will be very high-resolution. High resolution is below 2 m, anything above will be medium resolution. When I am talking about hyperspectral, the best that was achieved by NASA or anyone else before us was 30 m. When we launched our first demo, we made it 10 m — three times better. Now we are making it 5 m.
It is not possible to go below 5 m, because we are also capturing 150 wavelengths. It is a trade-off. I can reduce 150 to, say, 80, then I can get a resolution of 2.5 m. That is our plan — we intend to get down to 1 m and below. But 5 m is more than enough for the use cases we cater to. Our customers come to us because of 150 wavelengths, they don’t come to us for 30 cm resolution.