Apheris AI Raises $8.3M for Federated AI Platform

Berlin-based startup Apheris AI has raised $8.25 million in Series A funding to advance its groundbreaking federated computing platform, designed to unlock the vast potential of artificial intelligence in the life sciences industry while addressing critical data privacy and security concerns. This innovative approach allows AI models to be trained on decentralized data sources, eliminating the need for data centralization and paving the way for accelerated drug discovery and improved patient outcomes.

The core issue hindering the use of AI in healthcare is the inaccessibility of valuable health data. While AI is fundamentally dependent on large amounts of data, concerns surrounding patient privacy, regulatory hurdles, and intellectual property (IP) protection often prevent data sharing. This is where Apheris AI steps in, aiming to bridge this gap through a technology known as federated computing. This allows AI models to be trained across multiple, decentralized data sources without needing to move or share the raw data. According to a company press release, Apheris is the sole company addressing governance and IP protection in such collaborative networks, a market exceeding $30 billion.

Federated Learning: A Game Changer for Life Sciences

Federated learning is a type of machine learning that allows AI models to learn from data spread across different locations, without the data ever leaving its original source. This is especially important in the life sciences field, where data privacy is paramount due to strict regulations like HIPAA and GDPR. Federated learning offers a powerful solution by enabling collaborative research without compromising sensitive patient information.

Here are some examples of how federated learning is being used in medicine:

• Distinguishing between healthy and cancerous brain tissue: Using MRI images to pinpoint cancerous areas accurately.
• Leukemia diagnosis: Analyzing blood samples to detect and diagnose leukemia.
• Tuberculosis and lung pathology diagnosis: Using X-ray images to identify tuberculosis and other lung conditions.
• Predicting clinical outcomes in COVID-19 patients: Using medical records and X-rays to predict how COVID-19 patients will fare.

These applications often rely on deep neural networks, particularly convolutional neural networks, to achieve precise results. Apheris AI has chosen NVIDIA FLARE as the foundation for its platform, incorporating features that ensure security, privacy, and governance in high-stakes scenarios. The company also developed the Compute Gateway to meet regulatory needs, giving data providers control over how their sensitive data is processed at the algorithmic level.

Addressing the Data Bottleneck in AI for Life Sciences

The life sciences industry is under immense pressure to speed up drug development and find new therapies. AI holds great promise, but a major obstacle is effectively accessing and using the necessary data. As noted in a recent article on Applied Clinical Trials, the traditional drug discovery process is slow and expensive, making AI an attractive alternative. However, as highlighted by a Forbes article, data quality remains a significant hurdle.

The rapid advancement of AI, fueled by improvements in machine learning, specialized hardware, and data processing, underscores the need for solutions like Apheris to address data bottlenecks. These advancements have enabled AI systems to handle massive datasets and infer patterns with greater accuracy, but they also highlight the urgency of unlocking access to more high-quality data. In drug development, the shift from target discovery to target validation and hit identification is particularly challenging, requiring extensive lab experiments that consume significant time and resources, as explained in a recent Technology Networks article.

Apheris AI’s platform directly addresses these challenges by facilitating secure data collaboration without compromising privacy. The company’s innovative Compute Gateway allows computations to be executed locally where the data resides, and only essential model parameters are aggregated centrally. This approach effectively addresses the challenge of data protection in sensitive areas like healthcare.

Robin Röhm, co-founder and CEO of Apheris AI, stated in a recent interview with TechCrunch, “Without addressing the data owners’ concerns in providing data to AI, we don’t think that the impact of AI can really be unlocked, and that’s ultimately the core mission of what we’re building.” This emphasizes the company’s commitment to resolving the data bottleneck issue.

A Competitive Landscape and Apheris AI’s Unique Position

Several companies are working on similar solutions, focusing on data privacy and federated learning in healthcare. According to CB Insights, Apheris AI has a few competitors, including Rhino Health and Sherpa.ai. Other notable players in the federated learning ecosystem include NVIDIA, with its FLARE platform, and Owkin, with its Substra platform. However, Apheris AI has managed to find a niche with its focus on the data owner side and the life sciences. Its initial goal was to compete with open-source federated learning frameworks, but a pivot in 2023 to focus on the needs of data owners in the pharma and life sciences sectors has proven successful. According to Röhm, the company found product-market fit with a new product launched in late 2023, and has since quadrupled its revenue.

Marcin Hejka, co-founder and managing partner at OTB Ventures, which co-led the Series A funding round, believes Apheris AI could become a crucial part of the emerging federated data networks. He told TechCrunch, “We see a maturing ecosystem of third-party software tools (open source federation engines, data quality tools, and security products). Apheris also enables seamless integration with complementary privacy-enhancing technologies (homomorphic encryption, differential privacy, synthetic data).”

The Road Ahead: Apheris AI’s Future Impact

The $8.25 million in Series A funding, bringing Apheris AI’s total funding to $20.8 million, will be used to expand the team and further develop the platform. The company plans to hire senior talent with life sciences experience, particularly on the commercial side. Apheris AI is already being used by the AI Structural Biology (AISB) Consortium, a group that includes major pharmaceutical companies like AbbVie, Boehringer Ingelheim, Johnson & Johnson, and Sanofi. The company plans to focus further on protein complex prediction, an area where there is limited public data but vast amounts of valuable, sensitive data within life sciences companies.

Apheris AI’s solution has the potential to significantly impact the life sciences industry by accelerating drug discovery, improving AI model accuracy, enhancing diagnostics, and even improving manufacturing processes, as highlighted in a recent article on The Impact of AI on Life Sciences. By fostering secure data collaboration, Apheris AI is positioned to unlock the full potential of AI in life sciences, ultimately leading to better patient outcomes and a more efficient healthcare system. As explained in a report by Deloitte, the use of AI in pharma and life sciences is growing rapidly, and companies like Apheris are at the forefront of this transformation.

In conclusion, Apheris AI is pioneering a new era of data collaboration in the life sciences industry. By leveraging federated computing and its unique Compute Gateway, the company is enabling secure and efficient data analysis, paving the way for faster drug discovery, more accurate diagnostics, and personalized medicine. With a strong team, substantial funding, and cutting-edge technology, Apheris AI is poised to revolutionize how life sciences companies utilize data and AI to drive innovation and improve patient outcomes, shaping the future of healthcare.