Super Protocol first to bring Confidential AI to Healthcare – powered by NVIDIA Blackwell, Nebius AI Cloud, and Google Research
New verifiable infrastructure enables healthcare providers to run AI on real medical data under full confidentiality and compliance
Introduction
Super Protocol, a pioneer in decentralized confidential computing and verifiable AI infrastructure with a team bringing over a decade of expertise in confidential computing and cloud security, today announced that Yma Health, a developer of personalized medicine solutions, has successfully deployed its privacy-preserving anonymization platform on Super Protocol's secure cloud powered by NVIDIA Confidential Computing technology running on NVIDIA Blackwell AI infrastructure on Nebius AI Cloud. The integration enables healthcare providers to use AI on real patient data without exposing identities or violating regulations, dramatically improving the efficiency and compliance of medical AI applications.
The Challenge: Bridging the Gap Between Data Privacy and AI Innovation
The explosion of AI in healthcare has exposed a critical gap between innovation and privacy. Medical institutions possess vast amounts of patient data essential for accurate AI models, yet strict data protection laws and compliance frameworks prevent this data from being shared or processed externally. Traditional de-identification is no longer sufficient - combinations of age, dates, and rare conditions can still reveal individual identities. As a result, hospitals hesitate to adopt AI, EHR providers cannot expand functionality, and developers lack access to real-world medical data. This growing "data wall" blocks progress in personalized medicine and underlines the urgent need for confidential computing infrastructure - a technology that allows data to be used without ever being exposed.
The Solution: Verifiable Confidential AI Through Collaborative Innovation
Super Protocol, together with Yma Health, has introduced a groundbreaking approach that reconciles privacy and innovation through Confidential Computing. By running all anonymization, processing, and AI inference inside Trusted Execution Environments (TEEs) powered by NVIDIA AI running on Nebius AI Cloud, the system guarantees that sensitive data remains encrypted and verifiable at every step. Yma's anonymization middleware acts as a "VPN for Healthcare," automatically retrieving and anonymizing patient data before AI processing – without requiring any infrastructure changes on the clinic side. The collaboration combines Yma's domain expertise in medical data with Super Protocol's verifiable decentralized architecture, delivering a fully auditable, zero-trust solution that enables compliant AI use on real medical data for the first time.
Results and Impact: Proving That Confidential AI Works in the Real World
The collaboration between Yma Health and Super Protocol has delivered measurable outcomes that redefine what is possible in secure medical AI.
"By combining decentralized confidential computing with verifiable AI infrastructure, we're proving that privacy and innovation can finally coexist," said Nukri Basharuli, Founder and CEO of Super Protocol. "Our collaboration with NVIDIA and Yma Health demonstrates how hardware-rooted trust and GPU-accelerated confidential computing can protect data end-to-end – enabling compliant, high-performance AI across the entire healthcare ecosystem."
"Super Protocol's TEE technology helped us solve a fundamental challenge in medical data transfer. Previously, we had to write extensive custom anonymization code. Now, Google's medical AI model MedGemma handles this within the secure environment. I'm confident this approach will soon become standard across the health tech industry." Savvov Sergei, Chief Technology Officer, YMA Health.
"We wanted to partner with AI-native companies like YMA, but secure data transfer was always the blocker. The anonymization solution from YMA and Super Protocol finally solves this - we can now confidently share medical data knowing it's fully protected in their secure environment." Giri Rajan, CTO and Managing Director at Simplex EHR.
"As industries from healthcare to finance accelerate their adoption of AI, securing data during processing has become essential – not just for compliance, but as a foundation for innovation and trust," said John Fanelli, Vice President, Enterprise AI, NVIDIA. "With NVIDIA Confidential Computing on NVIDIA Blackwell GPUs, Super Protocol extends protection to every layer of AI workflows, ensuring verifiable security and performance in trusted execution environments from the infrastructure core all the way to frontline clinics."
By combining over a decade of Super Protocol's expertise in confidential computing with Yma's domain knowledge in healthcare data, the project has proven that real-world patient information can be safely used for AI-driven innovation without compromising privacy or compliance.
• End-to-End Data Protection:
All operations – from EHR extraction to AI inference – run entirely inside Trusted Execution Environments, ensuring that sensitive data is never exposed, even to infrastructure providers. Only anonymized, clinically valid outputs leave the enclave.
• Zero-Trust Transparency:
Through Super Protocol's Certification System and remote attestation, every workload and environment is verified cryptographically. Immutable execution reports provide public proof of compliance, creating unprecedented transparency and auditability for medical AI.
• Regulatory Readiness:
The architecture directly supports emerging global frameworks such as the EU AI Act, HIPAA, and GDPR, aligning with regulators' shift toward data-in-use protection rather than just data-at-rest or in-transit security.
• Operational Efficiency:
Clinics and EHR providers can integrate AI modules instantly – without any infrastructure changes or data-sharing agreements – while maintaining full compliance with DHA and international policies. This drastically shortens time-to-deployment and reduces legal and integration overhead.
• Resilience Against Modern Threats:
The use of verifiable confidential GPUs and hardware-based encryption provides protection even against advanced memory or model-extraction attacks, positioning the system at the frontier of modern cybersecurity.
• Trust and Adoption:
Early implementations at Simplex Himes and JointSpace demonstrate tangible improvements in diagnostic workflows, treatment personalization, and patient trust – all achieved without compromising confidentiality.
In the Yma Health and Super Protocol system, MedGemma 27B Multimodal functions as the intelligent anonymization engine. Running inside Trusted Execution Environments on NVIDIA Blackwell AI infrastructure, it analyzes both medical text and images to identify and remove Protected Health Information (PHI) while preserving all clinically relevant context.
This enables a new level of context-aware anonymization, where privacy protection and diagnostic value coexist. Integrated with Super Protocol's confidential cloud infrastructure, MedGemma operates securely, verifiably, and at GPU-level performance – forming the cognitive core of the "VPN for Healthcare."
We extend our gratitude to the MedGemma team at Google Research for their open collaboration, innovation, and support in enabling secure, multimodal healthcare AI under the principles of transparency and responsible technology.
Ultimately, this collaboration showcases that Confidential AI is no longer theoretical – it is a deployable, auditable, and regulatory-ready foundation for the next generation of healthcare innovation.
By enabling computation on sensitive data without exposure, Super Protocol and Yma Health have set a new global benchmark for secure, verifiable, and ethical AI in medicine.
Nebius AI Cloud provides the GPU backbone for large-scale model inference within securely attested environments. Supporting multiple generations of NVIDIA accelerated computing – including NVIDIA Blackwell and Hopper GPUs – Nebius delivers the performance, scalability, and reliability required for confidential AI workloads at production scale.
Its deep control over both hardware and orchestration layers enables seamless integration with Trusted Execution Environments and remote attestation workflows, making Nebius a perfect infrastructure partner for Super Protocol's verifiable confidential computing architecture.
We extend our appreciation to the Nebius AI Cloud team for their technical collaboration and continued support in advancing secure, high-performance AI infrastructure for regulated and data-sensitive industries worldwide.
Experience NVIDIA Confidential Computing in action at GTC Washington, D.C., on the show floor Oct 27-29, 2025 or learn more about this solution at https://superprotocol.com/case-studies/
About Super Protocol
Super Protocol is a decentralized cloud platform built on Confidential Computing and AI infrastructure, enabling secure and verifiable data processing in Trusted Execution Environments.
Founded by Nukri Basharuli, the team has over 10 years of experience in managing intellectual property, personal data, and AI security, with successful collaborations with Global Brands.
In collaboration with NVIDIA, Nebius AI Cloud, and Canonical, Super Protocol helps industries like healthcare and finance deploy confidential, compliant AI solutions at scale.
