What was the impact of the AI solution?

Trains machine learning models more than 20x faster than before. Saved 30% in monthly data operations costs by removing unused resources. Improved operational speed and ability to run a higher volume of experiments more reliably.

Source

GCPProductionEvidence: Medium65/100

Ordaos case study - Google Cloud

Use case typeCloud migrationUpdated Jul 18, 2026

Ordaos is a human-enabled, machine-driven drug design company that uses generative AI to design mini-proteins and analyze millions of protein structures. The company moved its cloud computing and storage to Google Cloud to better support larger-scale datasets, improve scalability, and reduce outages that had been affecting its AI projects. Ordaos also uses Google Cloud technologies to optimize SQL performance on Kubernetes, ingest third-party datasets such as AlphaFold, and scale experiments more efficiently.

Organization: Ordaos
Industry: Pharma
Location: United States
Published: July 2026

Reported outcomes

−30%

monthly data operations costsCost savings

20xmodel training speed

Strategic outcomes

Scale & capacityImproved ability to run larger-scale experiments reliablyRisk & complianceReduced outages and improved uptimeNew product / capabilityEnabled generative AI mini-protein design at scale

Catalog median for cost savings deployments: −40% across 177 reported metrics. Compare benchmarks →

Why do we believe this?Outcome claims, sources, and evidence checks

Normalized claim

Model training speed: 20 x increase

Google Cloud Customer StoriesJul 18, 2026Customer storyExplicit claimMedium evidence strength

train its machine learning models more than 20x faster than what was possible with its previous provider

Normalized claim

Monthly data operations costs: 30% decrease

Google Cloud Customer StoriesJul 18, 2026Customer storyExplicit claimMedium evidence strength

saved the Ordaos team 30% in monthly data operations costs

Why do we believe this deployment?Customer identity, provider attribution, maturity, and source checks

Customer: Ordaos
Provider: GCP
Maturity: Production
Linked source: Google Cloud Customer Stories

Improved operational speed and ability to run a higher volume of experiments more reliably

Customer identity supportedSource describes one deploymentMaturity supported

Primary read

Use case focus

Showing 3 of 3

1Cloud migration
2Training infrastructure modernization
3Data platform modernization

Needed reliable cloud to handle growing, large-scale protein structure datasets and improve performance and uptime for training and experimentation; prior provider caused recurring outages impacting AI projects.

Migrated cloud computing and storage to Google Cloud Storage and Google Kubernetes Engine.
Optimized SQL performance on Kubernetes using Google Kubernetes Engine and Cloud SQL.
Used Google Kubernetes Engine to ingest and experiment with third-party datasets such as AlphaFold.
Connected Google Kubernetes Engine clusters to Memorystore for Redis and used KEDA for event-driven scaling.

Technologies

Google Cloud Storage Google Kubernetes Engine Cloud SQL Memorystore for Redis KEDA Google Cloud

Trains machine learning models more than 20x faster than before.
Saved 30% in monthly data operations costs by removing unused resources.
Improved operational speed and ability to run a higher volume of experiments more reliably.

Architecture

Ordaos migrated cloud computing and storage workloads to Google Cloud, using Google Cloud Storage and Google Kubernetes Engine as the core infrastructure. It optimized SQL workloads with Cloud SQL on Kubernetes, integrated Memorystore for Redis for latency-sensitive scaling, and used KEDA for event-driven scaling. The team also ingested and experimented with large third-party datasets such as AlphaFold on GKE.

Sources & evidence1

Evidence: Medium65/100Evidence strength