Gamaya: Making sustainable global farming possible with groundbreaking imaging technology
About Gamaya
Gamaya provides intelligence on farmland health via AI, hyperspectral imaging, and other remote-sensing technologies. Its vision: sustainable farming for a world that will have 10 billion people by 2050.
Tell us your challenge. We're here to help.
Contact us
About Wabion
Google Cloud Premier Partner Wabion helps businesses digitally transform, with 50+ certified specialists for Google Cloud, data and analytics, AI/ML, Google Workspace, and Google Maps Platform.
Gamaya reduced cloud computing costs by up to 70%, achieved more than tenfold gains in operational efficiency, and unleashed transformative GPU solutions with Google Cloud.
Google Cloud results
- Enables Gamaya to invest more deeply in its groundbreaking imaging solution, by reducing cloud infrastructure costs by 70%
- Lightweight design and fast Google Cloud tooling reduces development time by 50%, with 10x faster VM deployment
- Extensive Google Cloud GPU profile library enables 30%+ cost reduction for creating stitched orthoimage from drone images
- Accelerates access to data-driven insights that help farmers to boost yield, avoid blights, and cultivate crops more efficiently
Enables 70% cost reduction for virtual machines (VMs)
For Swiss agritech startup Gamaya, there's more to sustainable farming than meets the eye.
The company captures light waves that are invisible to human sight, by deploying hyperspectral imaging, a technology originally developed by NASA. This enables Gamaya to paint a hyper-realistic canvas of farmland in the full range of wavelengths, from ultraviolet to infrared, beyond the tiny palette that is available to our vision.
The insight behind it all? Healthy plants reflect light in different ways than diseased plants. So, by capturing every last nuance of light in fields from Brazil to India, Gamaya can create granular crop diagnostic heat maps to boost yield, avoid blights, and deploy only the water and nutrients needed for a bountiful harvest. It’s a groundbreaking approach that’s helping emerging nations to meet their rising food needs in a sustainable manner.
But the method calls for crunching staggering amounts of imaging data. Gamaya leverages drones to fly over industrial farms that cover the size of 25,000 football pitches, capturing tens of thousands of images with a proprietary hyperspectral camera and other imaging cameras (RGB).
"We started with a different leading cloud provider, and from our perspective, we find Google Cloud to be the most professional, simply the smoothest in terms of dynamics, documentation, and user interface. For our developers and engineers, Google Cloud just works better."
—Evgeny Bogdanov, CTO, GamayaThese images, which can easily produce several gigabytes of data after one flight, must be uploaded, stored, and "stitched together" into one enormous precision map of the land's health. It's a challenge that requires the most powerful cloud computing, storage, and monitoring capabilities available. To make it possible, Gamaya turned to Google Cloud.
"We started with a different leading cloud provider, and from our perspective, we find Google Cloud to be the most professional, simply the smoothest in terms of dynamics, documentation, and user interface," says Evgeny Bogdanov, CTO, Gamaya. "For our developers and engineers, Google Cloud just works better."
Power and flexibility to meet any big data imaging needs
Gamaya's task of collating a gigantic image payload and running it through AI analysis is complicated enough. The job is made vastly more complex by natural contingencies such as wind, rain, fog, and cloud that make each drone mission unique, with shifting parameters. Once everything's been stitched together, moreover, the team may discover the picture isn't properly aligned to the planet, and need to perform a rotation.
Such issues are all in a day's work for Gamaya. That means its cloud provider must be powerful and flexible enough to handle an unpredictable array of natural and positioning factors that can impede accurate hyperspectral farmland mapping.
Under pressure to deliver timely results, Gamaya ran into frustrations with its original cloud provider. These, says Evgeny, included issues with speed, versatility, and complicated interfaces. Gamaya decided to give Google Cloud a try when it won Google funding for a kickstart program that included partnership with cloud service provider Wabion.
“Testing Google Cloud gave us confidence. In days, we proved that with Compute Engine, we could dynamically create resources as needed, and do in weeks what would otherwise take a year. We were emboldened to make major changes, and that meant a full transition to Google Cloud.”
—Evgeny Bogdanov, CTO, GamayaEvgeny, who had only just joined Gamaya as CTO, launched a month-long test to run the two solutions side by side. The results quickly overcame any reservations the Gamaya team may have felt about leaving their cloud comfort zone.
Development efforts that Gamaya could expect to take a year were accomplished in just 30 days with Google Cloud. The key was moving from a "slow and complicated" process of firing up virtual machines to dynamic resource provisioning that allowed Gamaya to spin "hundreds of VMs" simultaneously. It was a game changer for a company with enormous image processing requirements.
“Testing Google Cloud gave us confidence,” recalls Evgeny. “In days, we proved that with Compute Engine, we could dynamically create resources as needed, and do in weeks what would otherwise take a year. We were emboldened to make major changes, and that meant a full transition to Google Cloud.”
“If a data payload that used to take 10 minutes to process only takes a minute on Google Cloud, I don’t need to pay for 10 minutes on a virtual machine. It’s not about hardware profile differences; Google Cloud simply has a clearly better infrastructure than our previous provider.”
—Evgeny Bogdanov, CTO, GamayaReal (and significant) cost savings to put back into the business
Evgeny says the speed of spinning up, and closing down, multiple virtual machines has yielded huge cost savings: “If a data payload that used to take 10 minutes to process only takes a minute on Google Cloud, I don’t need to pay for 10 minutes on a virtual machine,” explains Evgeny. “It’s not about hardware profile differences; Google Cloud simply has a clearly better infrastructure than our previous provider.”
Evgeny says that in addition to the ease of dynamic resource provisioning, Google Cloud offers about 10 other advantages that have delivered up to 70% in cost savings. One key example? The robust library of Cloud GPUs for stitching together drone images into one massive heat map.
"We'll spend days exploring profiles for different image processing needs," Evgeny says. "With Google Cloud, we found a profile that was much more cost-effective, and that's exactly what we needed due to our bulk data needs. This was only available on Google Cloud. And it seemed a perfect fit for our specific purpose. For us, it was proof of the versatility of Google Cloud, which simply allows us to find the solutions we need."
Imaginative collaboration to sow the seeds of farming's future
For Gamaya, one of the drivers of building a next-generation agritech platform has been Google Cloud's engagement in problem-solving: "Google Cloud is very open, very responsive," says Evgeny. "There's this dynamic that we never found elsewhere, that we're all in the same boat, working together as a team."
Evgeny recalls a specific instance that drove home the ethos: "Google Cloud is the only cloud solution that we've ever encountered in which we create a VM and get a message from Google telling us that 'the way you utilize this resource seems to tell us that you're not using it fully, so you might as well turn the spec down. And save money.' For us, this was unbelievable. Google Cloud was the only one who told us we should be spending less money, not more."
A critical aspect of the team spirit: Google Cloud introducing Wabion as a cloud services partner. Google Cloud made resources available for Gamaya to use Wabion as part of its kickstart program. It's a partnership, says Evgeny, that led to transformation "in just one phone call."
At the time, Gamaya was looking for yet more ways to rein in the costs. In a call that also included Google Cloud, Wabion suggested switching to Preemptible VMs. In cloud computing, preemption allows users to run a VM at a much lower cost, with the tradeoff that Compute Engine can terminate (or preempt) these instances if the resources are needed elsewhere.
"Many digital solutions serve real-time users, and in those cases preemption is very bad. If you're cut off in the middle of a job, that could cost you your client," Evgeny says. "But in our case it's different. Because we launch hundreds of VMs simultaneously, we're perfectly OK with Google Cloud cutting us off in favor of someone with a higher priority. We can simply try again a few minutes later. That's the kind of balance and flexibility that Google Cloud provides us."
The result of heeding Wabion's advice on preemptive instances? A further 80% reduction in Gamaya's costs of deploying VMs for its big data processing needs.
"We probably wouldn't have thought of this possibility without Wabion's input," says Evgeny. "It's all about choosing the right team members. And in our industry, a team is made up of choosing partners and products and infrastructure and people. It's all together. We win as a team."
“From the very beginning, we were struck not only by Gamaya’s revolutionary technology, but also its determination to harness the full potential of Google Cloud,” says Samuel Pasquier, Managing Director of Wabion in Switzerland. “Thanks to this mindset, we were able to leverage our expertise for Gamaya in the most efficient way. We look forward to continuing our productive collaboration.”
Solving for sustainable farming with AI and machine learning
As Gamaya seeks to reinvent sustainable farming for a world that will need to feed nearly 10 billion people by 2050, it's exploring new avenues of partnership with Google Cloud using AI and machine learning.
Gamaya is currently carrying out a revamp of its AI and Deep Learning processes, testing new ideas with Google Cloud. In particular, it plans to explore how the AutoML suite of machine learning products can help it combine automated algorithms with human curation to continuously improve its platform.
"It's a vision of human-machine evolution," says Evgeny. "We look forward to diving deeper into Google Cloud Auto ML as we build the next steps of AI-driven agriculture for a sustainable world."
Tell us your challenge. We're here to help.
Contact us
About Gamaya
Gamaya provides intelligence on farmland health via AI, hyperspectral imaging, and other remote-sensing technologies. Its vision: sustainable farming for a world that will have 10 billion people by 2050.
About Wabion
Google Cloud Premier Partner Wabion helps businesses digitally transform, with 50+ certified specialists for Google Cloud, data and analytics, AI/ML, Google Workspace, and Google Maps Platform.