What challenges did Phagos face before AI implementation?

Developing highly specific bacteriophage treatments rapidly to address antibiotic resistance in animals and humans. Manual trial-and-error phage screening and huge genomic search space made matching impossible to scale.

What was the impact of the AI solution?

Reduced treatment development from 10+ years to 2 months. Cut wet-lab tests by 50% and achieved 99.5% screening time savings. Deployed in animal farming with more than half a million animals treated in France.

Source

ProductionEvidence: Medium65/100

Phagos uses Amazon SageMaker AI to create customized sustainable antibiotic alternatives

Use case typeDrug discoveryUpdated Jul 15, 2026

Phagos applies generative AI to match bacteriophages with target bacteria for personalized phage therapy. The company built AI models on AWS to simulate millions of phage-bacteria interactions and accelerate treatment discovery.

Organization: Phagos
Industry: Pharma
Location: France
Published: July 2026

Reported outcomes

500,000 count

animals treatedOther quantified impact

2 monthstreatment development time−50%wet-lab tests−99.5%screening time savings+290%screening speed

Strategic outcomes

New product / capabilityEnabled personalized veterinary medicines in FranceMarket & geographic expansionPlanned expansion from animal farming to human applications by 2030Innovation & cultureCreated a proprietary AI platform for phage-based treatments

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

Normalized claim

Treatment development time: 2 months decrease

AWS Customer StoryJul 15, 2026Customer storyExplicit claimMedium evidence strength

"develop a new cure in two months from scratch"

Normalized claim

Wet-lab tests: 50% decrease

AWS Customer StoryJul 15, 2026Customer storyExplicit claimMedium evidence strength

"Phagos needs 50% fewer tests in the wet lab"

Normalized claim

Screening time savings: 99.5% decrease

AWS Customer StoryJul 15, 2026Customer storyExplicit claimMedium evidence strength

"a 99.5% time savings when screening phage candidates"

Normalized claim

Screening speed: 290% increase

AWS Customer StoryJul 15, 2026Customer storyInferred claimMedium evidence strength

"10 minutes per bacteria versus 29 hours in the wet lab"

Normalized claim

Animals treated: 500,000 count increase

AWS Customer StoryJul 15, 2026Customer storyExplicit claimMedium evidence strength

"more than half a million animals have already been treated"

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

Customer: Phagos
Provider: AWS
Maturity: Production
Linked source: AWS Customer Story

Deployed in animal farming with more than half a million animals treated in France

Customer identity supportedSource describes one deploymentMaturity supported

Primary read

Use case focus

Showing 2 of 2

1Drug discovery
2Therapeutics research

Developing highly specific bacteriophage treatments rapidly to address antibiotic resistance in animals and humans.
Manual trial-and-error phage screening and huge genomic search space made matching impossible to scale.

Used Amazon SageMaker AI to train and fine-tune generative AI models on genomic data.
Ran live production simulations of millions of phage-bacteria interactions with AWS infrastructure backed by Amazon EC2, Amazon S3, and Amazon RDS.

Technologies

Amazon SageMaker AI Amazon EC2 Amazon S3 Amazon RDS

Reduced treatment development from 10+ years to 2 months.
Cut wet-lab tests by 50% and achieved 99.5% screening time savings.
Deployed in animal farming with more than half a million animals treated in France.

Architecture

Phagos trains and fine-tunes generative AI models on Amazon SageMaker AI using genomic data from public databases and its own lab-generated data. The models simulate millions of phage-bacteria interactions to identify optimal phage characteristics. The solution runs in production and is supported by Amazon EC2 for compute, Amazon S3 for the data lake, and Amazon RDS for structured queries and data management.

Sources & evidence1

Evidence: Medium65/100Evidence strength