THESIS

Early partners to founders building at the frontier of AI

We invest at pre-seed and seed in visionary founders across Europe who are building solutions that enable the development of AI models and applications, and solutions that apply AI technologies to mission-critical workflows.

Two Fronts

AI is still in its early innings. Most organisations are only beginning to embed real intelligence into their stacks and workflows.

This opens two greenfield fronts: building the infrastructure and tooling that make AI practical at scale, and crafting domain-specific products that wield it inside core operational processes.

We partner with founders defining either frontier.

Enablers
VIEW ENABLERS
Applications
VIEW APPLICATIONS
ENABLERS

We invest in the foundational software powering AI, spanning data infrastructure, developer tools, the full MLOps lifecycle, and next-generation compute efficiency. These solutions,
‍
• Enable enterprises to progress their technological maturity by addressing the challenges that organisations face in handling their data, applying AI, and scaling its deployment
‍
• Empower the individual builders of models and applications by addressing the challenges that developers, AI engineers, and data scientists face within their workflows

Data Infrastructure
Developer Tools
MLOPS
COMPUTE

Data Infrastructure

The platforms and tooling providing the underlying infrastructure for collecting, managing, and working with data at scale. Ensuring that data is trusted, accessible, and scalable, serving as the foundation for building models, powering applications, and driving analytics.
RELATED PORTFOLIO
Acquisition, Labelling and Ingestion
Storage and Lakehousing
Transformation and Processing
Catalogue, Quality, Governance and Security
Serving, Analytics and Activation

Developer Tools

Solutions that help developers to automate and streamline the process of building, testing, and deploying software. Designed to enhance reproducibility, quality and iteration speed.
RELATED PORTFOLIO
Planning and Collaboration
Coding and Code Intelligence
Frameworks and Agent Toolchains
Build, Test and Security
CI/CD, Deploy and Reliability Engineering

MLOps

Solutions that manage the end-to-end machine learning lifecycle, from model training and deployment to monitoring, observability, optimisation, and compliance in production. Built to move models from research to production at scale.
RELATED PORTFOLIO
Experiment Orchestration and Training Pipelines
Feature, Experiment and Model Management
Deployment and Inference Serving
Monitoring, Observability and Performance Optimisation
Governance, Compliance and Safety

Compute

Solutions that optimise compute performance, cost efficiency, and sustainability across demanding AI workloads, from inference to large-scale model training.
RELATED PORTFOLIO
Provisioning and Cluster Orchestration
Distributed Training and Execution Frameworks
Compilers, Runtimes and Graph Optimisers
Model Compression and Precision Optimisation
Cost and Carbon-Aware Optimisers
Applications

We invest across industries in teams that combine deep domain insight with advanced AI technologies—predictive machine learning, computer vision, speech and multimodal models, generative models, and agentic and decision systems—to reinvent mission-critical B2B workflows. We look for solutions that are,

‍• Domain specific and built on behavioral insights, data moats, and rich operational context
‍
• Built by teams that iterate fast, work closely with their users, and hold strong convictions about product and interaction design

Domain Specificity and Insight

Focus on clearly defined, high-value workflows within targeted industries, delivering a quantifiable ROI and grounded in deep knowledge of user behaviour, processes, and integration touch-points.
RELATED PORTFOLIO
Predictive Machine Learning
Computer Vision
Speech and Multimodal
Generative Models
Agentic and Decision Systems

Legacy Bound Markets

Tackle sectors where incumbents rely on outdated tech and move slowly on AI—fertile ground for step-change products.
RELATED PORTFOLIO
Predictive Machine Learning
Computer Vision
Speech and Multimodal
Generative Models
Agentic and Decision Systems

Defensible Data Moats

Secure hard-to-access, domain-specific data and build proprietary flywheels that improve the product—and its advantage—with every use.
RELATED PORTFOLIO
Predictive Machine Learning
Computer Vision
Speech and Multimodal
Generative Models
Agentic and Decision Systems

Persistent Memory and Context

Capture and retain rich, longitudinal user and workflow data so models can deliver context-aware predictions and anticipate needs.
RELATED PORTFOLIO
Predictive Machine Learning
Computer Vision
Speech and Multimodal
Generative Models
Agentic and Decision Systems

Speed, Adaptability and GTM Agility

Ship quickly, iterate relentlessly with tight customer feedback loops, and translate learnings into agile go-to-market experiments that unlock scalable distribution.
RELATED PORTFOLIO
Predictive Machine Learning
Computer Vision
Speech and Multimodal
Generative Models
Agentic and Decision Systems

Opinionated UX and Interaction Design

Craft a clear product point of view that defines how humans and intelligent systems collaborate—prioritising usability, trust, and delight.
RELATED PORTFOLIO
Predictive Machine Learning
Computer Vision
Speech and Multimodal
Generative Models
Agentic and Decision Systems
View Enablers
Enablers
Applications
ENABLERS

We back the core software powering AI — spanning data infrastructure, developer tools, the full MLOps lifecycle, and next-generation compute efficiency. These solutions address the challenges that enterprises face in handling their data, applying AI and scaling its deployment. They enable the individual builders of models and applications by addressing the challenges that developers, AI engineers, and data scientists face within their workflows.

Data Infrastructure
Developer Tools
MLOPS
COMPUTE

Data Infrastructure

The platforms and tooling providing the underlying infrastructure for ingesting, storing, transforming, and working with data at scale. Designed to ensure data is trusted, governed, and scalable, serving as the foundation for building models, powering applications, and driving analytics.
RELATED PORTFOLIO
Acquisition, Labelling and Ingestion
Storage and Lakehousing
Transformation and Processing
Catalogue, Quality, Governance and Security
Serving, Analytics and Activation

Developer Tools

Solutions that help developers to automate and streamline the process of building, testing, and deploying software. Designed to enhance reproducibility, quality and iteration speed.
RELATED PORTFOLIO
Planning and Collaboration
Coding and Code Intelligence
Frameworks and Agent Toolchains
Build, Test and Security
CI/CD, Deploy and Reliability Engineering

MLOps

Solutions that manage the end-to-end machine learning lifecycle, from model training and deployment to monitoring, observability, optimisation, and compliance in production. Built to move models from research to production at scale
RELATED PORTFOLIO
Experiment Orchestration and Training Pipelines
Feature, Experiment and Model Management
Deployment and Inference Serving
Monitoring, Observability and Performance Optimisation
Governance, Compliance and Safety

Compute

Solutions optimising compute for performance, cost efficiency, and sustainability across demanding AI workload, from inference to large-scale model training.
RELATED PORTFOLIO
Provisioning and Cluster Orchestration
Distributed Training and Execution Frameworks
Compilers, Runtimes and Graph Optimisers
Model Compression and Precision Optimisation
Cost and Carbon-Aware Optimisers
View Applications
Applications

We invest in teams that fuse deep domain insight with cutting-edge AI technologies, predictive machine learning, computer vision, speech and multimodal models, generative models, and agentic and decision systems, to reinvent mission critical B2B workflows.

Across industries and workflows we look for sharp domain focus, defensible data moats, deep understanding of user behaviour and processes, and opinionated product and interaction design.

Domain Specificity and Insight
LEGACY DOMINATED MARKETS
DATA MOATS
MEMORY AND CONTEXT
Speed and Adaptability
Opinionated Product and Interaction Design

Domain Specificity and Insight

Targeting specific critical workflows within specific industries, with a deep understanding of user behaviours, processes, and the landscape of integrations
RELATED PORTFOLIO
Predictive Machine Learning
Computer Vision
Speech and Multimodal
Generative Models
Agentic and Decision Systems

Legacy Dominated Markets

Markets where incumbents are characterised by outdated technology and slow innovation
RELATED PORTFOLIO
Predictive Machine Learning
Computer Vision
Speech and Multimodal
Generative Models
Agentic and Decision Systems

Data Moats

Leveraging difficult-to-access, domain-specific data, and creating proprietary data flywheels through product usage
RELATED PORTFOLIO
Predictive Machine Learning
Computer Vision
Speech and Multimodal
Generative Models
Agentic and Decision Systems

Memory and Context

Product-generated data that persists in a model’s memory, the ability to provide deep context to models and predict what users want before they ask
RELATED PORTFOLIO
Predictive Machine Learning
Computer Vision
Speech and Multimodal
Generative Models
Agentic and Decision Systems

Speed and Adaptability

Shipping fast and often, rapid prototyping, and tight feedback loops
RELATED PORTFOLIO
Predictive Machine Learning
Computer Vision
Speech and Multimodal
Generative Models
Agentic and Decision Systems

Opinionated Product and Interaction Design

Designing from a clear point of view, defining how humans and intelligent systems relate to each other
RELATED PORTFOLIO
Predictive Machine Learning
Computer Vision
Speech and Multimodal
Generative Models
Agentic and Decision Systems
View Enablers