Why We Invested in Bitfount
Collaborating on sensitive data using AI across distributed sources, without moving it.
21/08/2025
Marius Swart
At Pace Ventures, we seek out pioneering teams leveraging cutting-edge technology to create transformational change. That’s why we’re excited to announce our investment in Bitfount, a world-class team of applied AI, federated learning, and privacy experts. They have built a platform for connecting any data, on any infrastructure, to any AI model, including LLMs.
The Opportunity
There is no denying that AI has a lot of potential. However, its full true potential will not be realized until the barriers to using sensitive data are overcome. Welcome Federated Learning (FL).
Fundamentally, instead of centralizing sensitive data, FL allows AI models to be trained, fine-tuned, evaluated, and deployed across multiple parties/devices while keeping raw data local or onsite. That makes it a perfect fit wherever data privacy, data sovereignty, or fragmented datasets are barriers to traditional AI.
The biggest opportunities for FL solutions right now are:
Healthcare and Lifesciences incl. Pharma
Finance and insurance
Consumer tech and IOT
Autonomous vehicles and
Government and other Public Sector
Bitfount’s initial focus area is on the healthcare and life sciences segment and more specifically, patient trial recruitment. Why is FL a good use case for this segment? Hospitals, labs, and clinics have sensitive patient data that can’t be shared due to HIPAA, GDPR, or institutional policies. The data is typically scattered across multiple sponsors and sites, resulting in up to 86% of clinical trials failing or being significantly delayed due to recruitment failures. Recruitment can account for up to 40% of a trial’s total budget and up to $8 million of lost revenue per day of delay for pharma companies.
Other areas within the health segment that also have large market sizes are:
Collaborative disease prediction (e.g., cancer, diabetes, cardiovascular risk).
Medical imaging AI (radiology, pathology) trained across hospitals without moving patient scans.
Genomics and rare disease studies (pooling data across institutions).
Endometriosis, Alzheimer’s, and other under-researched conditions where datasets are siloed.
The estimated market size for above mentioned healthcare use cases is> $150bn, and the clinical trial use case is ±$60bn.
Key Platform Features
Bitfount’s platform integrates multiple computational techniques and has several features that we felt provide a moat against potential competitors:
No code experience for data custodians: Connect data to projects and run analysis without the need for expensive data scientists.
Enterprise-grade platform: ISO27001 certified and HIPAA and GDPR compliant.
Ease of Deployment: Set up in minutes, not days. There is also a desktop app, perfectly suited for small clinics.
Support for all model types, including LLMs: Develop and deploy the next generation of AI applications.
The Team
Bitfount is led by 2 PhDs from Cambridge, Blaise Thomson (CEO & Co-Founder), a veteran in ML having sold his previous company to Apple, alongside Naaman Tammuz (CPO and Co-founder), previously CPTO at Duedil.
From our first interactions, we were impressed not only by their technical expertise but also by their deep commitment to execution. Their team has already demonstrated rapid progress, building a unique and scalable tech stack capable of screening and identifying potential patients for trials.
One example was a study with Moorfields Hospital in the UK, where they identified 600 potential patients for a trial vs 2 historically recruited, and with an estimated screen failure rate of 14% vs 60%. Another example is when they scanned 170K images across 7 private clinics in the US over 4 months to identify 8K patients that were potentially eligible for a trial.
Bitfount’s Impact
Right now, the world’s most valuable companies (Apple, Microsoft, Google, Amazon, Meta, NVIDIA) dominate because they control massive proprietary data pools (search queries, shopping histories, social graphs, sensor data, etc.). Their competitive edge comes from combining scale + data, + compute.
Our thesis is that in the future, the world’s largest data company won't hold any data (think Uberization of data). This is exactly what federated learning, data clean rooms, and decentralized AI ecosystems point toward, and we think Bitfount has the right team and execution focus to be a leader in this space.
The company has already achieved key commercial milestones since our investment in Q1 2024, and we are excited to be part of their growth story.
Our Co-Investors
We are thrilled to partner with Ahren Innovation Capital, Parkwalk Advisors, Foresight Group, and existing investor Speedinvest.
At Pace Ventures, we seek out pioneering teams leveraging cutting-edge technology to create transformational change. That’s why we’re excited to announce our investment in Bitfount, a world-class team of applied AI, federated learning, and privacy experts. They have built a platform for connecting any data, on any infrastructure, to any AI model, including LLMs.
The Opportunity
There is no denying that AI has a lot of potential. However, its full true potential will not be realized until the barriers to using sensitive data are overcome. Welcome Federated Learning (FL).
Fundamentally, instead of centralizing sensitive data, FL allows AI models to be trained, fine-tuned, evaluated, and deployed across multiple parties/devices while keeping raw data local or onsite. That makes it a perfect fit wherever data privacy, data sovereignty, or fragmented datasets are barriers to traditional AI.
The biggest opportunities for FL solutions right now are:
Healthcare and Lifesciences incl. Pharma
Finance and insurance
Consumer tech and IOT
Autonomous vehicles and
Government and other Public Sector
Bitfount’s initial focus area is on the healthcare and life sciences segment and more specifically, patient trial recruitment. Why is FL a good use case for this segment? Hospitals, labs, and clinics have sensitive patient data that can’t be shared due to HIPAA, GDPR, or institutional policies. The data is typically scattered across multiple sponsors and sites, resulting in up to 86% of clinical trials failing or being significantly delayed due to recruitment failures. Recruitment can account for up to 40% of a trial’s total budget and up to $8 million of lost revenue per day of delay for pharma companies.
Other areas within the health segment that also have large market sizes are:
Collaborative disease prediction (e.g., cancer, diabetes, cardiovascular risk).
Medical imaging AI (radiology, pathology) trained across hospitals without moving patient scans.
Genomics and rare disease studies (pooling data across institutions).
Endometriosis, Alzheimer’s, and other under-researched conditions where datasets are siloed.
The estimated market size for above mentioned healthcare use cases is> $150bn, and the clinical trial use case is ±$60bn.
Key Platform Features
Bitfount’s platform integrates multiple computational techniques and has several features that we felt provide a moat against potential competitors:
No code experience for data custodians: Connect data to projects and run analysis without the need for expensive data scientists.
Enterprise-grade platform: ISO27001 certified and HIPAA and GDPR compliant.
Ease of Deployment: Set up in minutes, not days. There is also a desktop app, perfectly suited for small clinics.
Support for all model types, including LLMs: Develop and deploy the next generation of AI applications.
The Team
Bitfount is led by 2 PhDs from Cambridge, Blaise Thomson (CEO & Co-Founder), a veteran in ML having sold his previous company to Apple, alongside Naaman Tammuz (CPO and Co-founder), previously CPTO at Duedil.
From our first interactions, we were impressed not only by their technical expertise but also by their deep commitment to execution. Their team has already demonstrated rapid progress, building a unique and scalable tech stack capable of screening and identifying potential patients for trials.
One example was a study with Moorfields Hospital in the UK, where they identified 600 potential patients for a trial vs 2 historically recruited, and with an estimated screen failure rate of 14% vs 60%. Another example is when they scanned 170K images across 7 private clinics in the US over 4 months to identify 8K patients that were potentially eligible for a trial.
Bitfount’s Impact
Right now, the world’s most valuable companies (Apple, Microsoft, Google, Amazon, Meta, NVIDIA) dominate because they control massive proprietary data pools (search queries, shopping histories, social graphs, sensor data, etc.). Their competitive edge comes from combining scale + data, + compute.
Our thesis is that in the future, the world’s largest data company won't hold any data (think Uberization of data). This is exactly what federated learning, data clean rooms, and decentralized AI ecosystems point toward, and we think Bitfount has the right team and execution focus to be a leader in this space.
The company has already achieved key commercial milestones since our investment in Q1 2024, and we are excited to be part of their growth story.
Our Co-Investors
We are thrilled to partner with Ahren Innovation Capital, Parkwalk Advisors, Foresight Group, and existing investor Speedinvest.
At Pace Ventures, we seek out pioneering teams leveraging cutting-edge technology to create transformational change. That’s why we’re excited to announce our investment in Bitfount, a world-class team of applied AI, federated learning, and privacy experts. They have built a platform for connecting any data, on any infrastructure, to any AI model, including LLMs.
The Opportunity
There is no denying that AI has a lot of potential. However, its full true potential will not be realized until the barriers to using sensitive data are overcome. Welcome Federated Learning (FL).
Fundamentally, instead of centralizing sensitive data, FL allows AI models to be trained, fine-tuned, evaluated, and deployed across multiple parties/devices while keeping raw data local or onsite. That makes it a perfect fit wherever data privacy, data sovereignty, or fragmented datasets are barriers to traditional AI.
The biggest opportunities for FL solutions right now are:
Healthcare and Lifesciences incl. Pharma
Finance and insurance
Consumer tech and IOT
Autonomous vehicles and
Government and other Public Sector
Bitfount’s initial focus area is on the healthcare and life sciences segment and more specifically, patient trial recruitment. Why is FL a good use case for this segment? Hospitals, labs, and clinics have sensitive patient data that can’t be shared due to HIPAA, GDPR, or institutional policies. The data is typically scattered across multiple sponsors and sites, resulting in up to 86% of clinical trials failing or being significantly delayed due to recruitment failures. Recruitment can account for up to 40% of a trial’s total budget and up to $8 million of lost revenue per day of delay for pharma companies.
Other areas within the health segment that also have large market sizes are:
Collaborative disease prediction (e.g., cancer, diabetes, cardiovascular risk).
Medical imaging AI (radiology, pathology) trained across hospitals without moving patient scans.
Genomics and rare disease studies (pooling data across institutions).
Endometriosis, Alzheimer’s, and other under-researched conditions where datasets are siloed.
The estimated market size for above mentioned healthcare use cases is> $150bn, and the clinical trial use case is ±$60bn.
Key Platform Features
Bitfount’s platform integrates multiple computational techniques and has several features that we felt provide a moat against potential competitors:
No code experience for data custodians: Connect data to projects and run analysis without the need for expensive data scientists.
Enterprise-grade platform: ISO27001 certified and HIPAA and GDPR compliant.
Ease of Deployment: Set up in minutes, not days. There is also a desktop app, perfectly suited for small clinics.
Support for all model types, including LLMs: Develop and deploy the next generation of AI applications.
The Team
Bitfount is led by 2 PhDs from Cambridge, Blaise Thomson (CEO & Co-Founder), a veteran in ML having sold his previous company to Apple, alongside Naaman Tammuz (CPO and Co-founder), previously CPTO at Duedil.
From our first interactions, we were impressed not only by their technical expertise but also by their deep commitment to execution. Their team has already demonstrated rapid progress, building a unique and scalable tech stack capable of screening and identifying potential patients for trials.
One example was a study with Moorfields Hospital in the UK, where they identified 600 potential patients for a trial vs 2 historically recruited, and with an estimated screen failure rate of 14% vs 60%. Another example is when they scanned 170K images across 7 private clinics in the US over 4 months to identify 8K patients that were potentially eligible for a trial.
Bitfount’s Impact
Right now, the world’s most valuable companies (Apple, Microsoft, Google, Amazon, Meta, NVIDIA) dominate because they control massive proprietary data pools (search queries, shopping histories, social graphs, sensor data, etc.). Their competitive edge comes from combining scale + data, + compute.
Our thesis is that in the future, the world’s largest data company won't hold any data (think Uberization of data). This is exactly what federated learning, data clean rooms, and decentralized AI ecosystems point toward, and we think Bitfount has the right team and execution focus to be a leader in this space.
The company has already achieved key commercial milestones since our investment in Q1 2024, and we are excited to be part of their growth story.
Our Co-Investors
We are thrilled to partner with Ahren Innovation Capital, Parkwalk Advisors, Foresight Group, and existing investor Speedinvest.