This page is also available in Deutsch, Español, Français, Italiano, 日本語, and Português.

Taranis Cuts Cost per Photo 10x

Google Cloud migration with V100 GPUs, Kubernetes, and TensorFlow lets Taranis process 30TB of drone imagery across 20M+ acres.

The Challenge

Taranis needed to upload large volumes of high-resolution drone images from remote locations worldwide and scale infrastructure to train complex machine learning models. Each drone flight collects 10,000 images of 10-20MB each. Up to 40% of crops are routinely lost due to insects, disease, weeds, and nutrient deficiencies. The company needed better connectivity, speed, and scalable processing power without massive infrastructure investment.

The Solution

Taranis migrated to Google Cloud Platform, leveraging global data centers for fast connectivity and V100 GPUs on Compute Engine for image processing. The solution includes automatic scaling from 1,000 to 4,000 V100s, Kubernetes Engine for satellite image processing, Cloud SQL for data storage, and TensorFlow for machine learning model training. The platform processes 100 million distinct features across 700,000 images.

Results

Reduced image upload time from a full day to just a few hours - 3-4x faster
Achieved 10x lower cost per photo taken after migration
Enabled continuous feature releases with parallel version deployment
Scaled infrastructure to handle millions more acres without operational concerns

Agriculture is a seasonal business, so we have certain months of peak activity followed by quiet months, and we also have peaks throughout the day. During quiet times, we can scale back all our high-level Compute Engine GPU resources automatically so we don't have to prepare our system in advance.
Eli Bukchin, Co-founder and CTO

Meeting the Global Food Challenge

According to a United Nations report, the world's population will reach 9.8 billion by 2050, requiring significant increases in food production. Meanwhile, urbanization and unpredictable weather patterns are reducing agricultural output. Taranis addresses this challenge using drone technology and AI to help farmers reduce crop loss, increase yields, and lower costs. Founded in 2014, the company now manages over 20 million acres worldwide through its intelligence platform.

The Data Upload and Processing Challenge

Taranis collects vast amounts of data from remote locations including Russia, Eastern Europe, and South America. Each drone flight captures around 10,000 images, with each image between 10-20MB. The company needed to upload these large volumes quickly while maintaining the processing power to train complex machine learning models. Up to 40% of crops are routinely lost due to insects, disease, weeds, and nutrient deficiencies, making early detection critical.

Google Cloud Infrastructure Solution

Taranis migrated to Google Cloud to leverage global data centers offering fast connectivity for their 30TB throughput. The solution uses V100 GPUs on Compute Engine with automatic scaling from 1,000 to 4,000 units based on demand. The architecture includes Kubernetes Engine for satellite image processing, Cloud SQL for data storage, Cloud Functions, and Cloud Pub/Sub. This flexibility allows the company to scale back resources during quiet agricultural seasons automatically.

Machine Learning with TensorFlow

Taranis uses TensorFlow for machine learning model training, processing tens of millions of photographs collected over the past year and a half. Each photo contains up to a thousand items of interest, such as insect damage or leaf discoloration. In total, the company has processed around 100 million distinct features across 700,000 images. The open source TensorFlow community provides extensive support for rapid model development.

Operational Improvements and Cost Reduction

The migration to Google Cloud reduced upload times from a full day to just a few hours - three to four times faster than before. The company now releases features almost continuously using Kubernetes parallel deployments, reducing downtime and eliminating scheduled updates. This allows faster product improvement and quicker feedback cycles. Most significantly, the cost per photo taken is now ten times lower than before.

Future Expansion and Analytics

Taranis is exploring additional Google Cloud tools including Cloud Bigtable, BigQuery, and Cloud Dataflow for better data analytics and business intelligence insights. The company plans further geographical expansion and continuous improvement of machine learning models for detecting new disease categories. Thanks to the scalable infrastructure, the sales team can onboard new customers without worrying about capacity constraints.

See how DoiT helps cloud teams control spend

Explore how DoiT Cloud Intelligence helps teams improve visibility, governance, and unit economics across cloud environments.

Visit DoiT Talk to our team

What they say

What I really like about DoiT's approach is that you're very hands-on and proactive. Satyam would ping me a few times a sprint, letting me know about the most current features, checking in on how things are going. When we are going through a peak time, that proactiveness makes a real difference. Satyam always comes through whenever we need support and helps us leverage the right experts to get us where we need to be.

Chiamaka Ibeme, Engineering Manager, Platform

SELECT has made important cost data readily accessible. I will often pull it up during engineering design reviews so we can quickly evaluate cost impact and projections and factor that into our design decisions.

Douglas Zickuhr, Senior Data Platform Engineer at Personio

I love clicking through SELECT to understand how our environment and workloads are evolving. I probably check it every day. It's coffee and SELECT for me every morning.

Ian Fahey, Senior Analytics Engineer at Loop

You guys have the best UI experience that I've had of any software. It's like you just read my mind where, like, oh, I wish I could click there. Oh, I can

Diana Koshy, Sr. Director of Data Engineering at Kargo

SELECT feels like exactly what Paul and I would have built if we had locked ourselves in a room for 18 months to create our ideal monitoring solution.

Devin McGee, Data Engineering Lead at Home Chef

SELECT dramatically lowers the cognitive load to understanding Snowflake costs. I'm able to sit there and easily understand what's driving the cost. Not to blow smoke up your ass, but it's just so easy to do in your platform

Blake Baggett, Head of Data Operations at Entain

One of the most helpful cost rituals we've setup from SELECT is the weekly spend digest sent to Slack. I can start high level and ensure things are in check. If not, I can very quickly drill down into specific workloads which may have driven the cost spike and remediate them before they become a bigger issue.

Michael Revelo, Manager of data and analytics engineering at ClickUp

Through SELECT's automated savings feature and deep cost visibility, we were able to instantly lower our Snowflake spend by over 40% and achieve a 20X ROI on our SELECT investment.

Skyler Chi, SVP, GTM Productivity & Excellence at Exiger

Our costs had jumped up 3X as we scaled, so we're talking about 60% savings in Snowflake spend after adopting SELECT.

Edward Mancey, GTM Lead at Synthesia

DoiT was a true partner, not a vendor. They helped us understand the problem, refine the vision, and build something production-ready far faster than we could have on our own. Their expertise, responsiveness, and commitment made all the difference.

Dr. Anish Kapur, Founder & CEO, Promptly

DoiT's Customer Success and Forward Deployed Engineering teams work very closely with us. The regular sessions with our CSM keep us focused on the right priorities, and the FDEs provide the deep technical guidance we need to validate decisions and optimize our environment. That combination has been genuinely valuable for us.

Alexander Lundberg Santos, Platform Engineer at Extenda Retail

When we started working with DoiT, we deployed Flexsave to save time and reduce complexity. We still use it today. But what really stands out is the expert support. Having someone to collaborate with on deep cloud cost topics, someone who really understands the nuances, is incredibly valuable.

Jesper Terkelsen, CTO at Monta

DoiT's Cloud Accelerator turned our AI idea into a shipped product, saving at least three months of development and delivering reliable, explainable insights our customers trust.

Scott Desgrosseilliers, CEO and co-founder, Wicked Reports

Your cloud bill shouldn't be a mystery

Let us show you what ships this week.

Book a demo