←
Super Protocol1,154 posts
Super Protocol avatar

Super Protocol

@super__protocol

Full-stack confidential AI for models, agents and data, across GPU clouds, on-prem and hybrid. Powered by #confidentialcomputing ❇️

◷ Joined March 2022▧ 476 Photos and videos
Super Protocol avatar
Super Protocol@super__protocol·Oct 29

A confidential GPU fails at 3 a.m. Who brings the service back, how do clients know the new node can be trusted, and can anyone log in to fix it? Hardware-based Confidential Computing protects data during execution. But it does not, by itself, address the operational challenges of running Confidential AI in production. A production service also needs attestation, key management, orchestration on attested nodes, trusted updates, and recovery without operator access. 2 𝐨𝐟 𝐭𝐡𝐞 11 𝐥𝐚𝐲𝐞𝐫𝐬 𝐧𝐞𝐞𝐝𝐞𝐝 𝐟𝐨𝐫 𝐚 𝐜𝐨𝐦𝐩𝐥𝐞𝐭𝐞 𝐂𝐨𝐧𝐟𝐢𝐝𝐞𝐧𝐭𝐢𝐚𝐥 𝐀𝐈 𝐨𝐟𝐟𝐞𝐫𝐢𝐧𝐠 𝐜𝐨𝐦𝐞 𝐰𝐢𝐭𝐡 𝐭𝐡𝐞 𝐡𝐚𝐫𝐝𝐰𝐚𝐫𝐞. 𝐓𝐡𝐞 𝐨𝐭𝐡𝐞𝐫 9 𝐚𝐫𝐞 𝐬𝐨𝐟𝐭𝐰𝐚𝐫𝐞 𝐭𝐡𝐚𝐭 𝐬𝐨𝐦𝐞𝐨𝐧𝐞 𝐡𝐚𝐬 𝐭𝐨 𝐛𝐮𝐢𝐥𝐝 𝐚𝐧𝐝 𝐨𝐩𝐞𝐫𝐚𝐭𝐞. Last week, we announced two UK projects using NVIDIA Confidential Computing. BE, a GPU cloud provider, added verifiable confidential LLM inference to its existing service. Reset Health validated AI on sensitive treatment data using externally hosted NVIDIA Blackwell systems. 𝐍𝐞𝐢𝐭𝐡𝐞𝐫 𝐡𝐚𝐝 𝐭𝐨 𝐛𝐮𝐢𝐥𝐝 𝐢𝐭𝐬 𝐨𝐰𝐧 𝐂𝐨𝐧𝐟𝐢𝐝𝐞𝐧𝐭𝐢𝐚𝐥 𝐀𝐈 𝐬𝐭𝐚𝐜𝐤. 𝐒𝐮𝐩𝐞𝐫 𝐏𝐫𝐨𝐭𝐨𝐜𝐨𝐥 𝐩𝐫𝐨𝐯𝐢𝐝𝐞𝐬 𝐚𝐥𝐥 11 𝐥𝐚𝐲𝐞𝐫𝐬 𝐚𝐬 𝐨𝐧𝐞 𝐩𝐥𝐚𝐭𝐟𝐨𝐫𝐦. Why a confidential server is not a confidential cloud: superprotocol.com/platform/c…

Post attachment
Super Protocol avatar
Super Protocol@super__protocol·Oct 23

What does Sovereign AI look like from a GPU cloud provider’s perspective? BE, Super Protocol and NVIDIA Blackwell are putting it into practice in the UK. Today, we are announcing that The BE Company, a UK-based AI infrastructure and services provider, has integrated Super Protocol into its existing inference service, enabling confidential LLM inference for enterprises handling sensitive business data, powered by NVIDIA Blackwell. Data residency addresses where information is processed, but not how it is protected while in use. The integration is designed to let BE’s enterprise customers protect sensitive information during processing, including from the provider and privileged administrators, and independently verify the environment and software handling it, without relying solely on the infrastructure operator’s assurances. Here’s how BE and @nvidia describe the value: Sath Ganesarajah, Founder and Chief Executive Officer, BE: “Our enterprise customers need to protect sensitive data while running AI at scale, with fast responses, reliable availability and clear service commitments. Super Protocol gives us the complete stack within our existing infrastructure, allowing us to deploy and scale protected inference without building a separate confidential computing platform. Customers can verify the service before sending sensitive information, giving them access to advanced AI with data privacy they can verify independently.” Anthony Hills, Regional Sales Leader for NVIDIA in the UK & Ireland: “As the UK expands its sovereign compute and inference capacity, enabling sensitive and regulated workloads is increasingly important. NVIDIA Confidential Computing provides hardware-backed protection during AI processing. Super Protocol supplies the complete software stack that makes this protection deployable and verifiable across inference services, enabling providers such as BE to add verifiable confidentiality to UK-based sovereign AI infrastructure.” 👉 Read more: superprotocol.com/resources/…

Post attachment
Super Protocol avatar
Super Protocol@super__protocol·Oct 23

What does Sovereign AI look like in healthcare? Reset Health and Super Protocol are putting it into practice with NVIDIA Blackwell. Together with Reset Health, we’ve announced a new Confidential AI capability for Reset AI within Reset Health’s clinician-led digital care platform for cardiometabolic conditions, including obesity and type 2 diabetes. Super Protocol enables Reset AI to analyse sensitive treatment data on externally hosted NVIDIA Blackwell systems, keeping patient data protected during processing and inaccessible to infrastructure providers and their administrators. Reset AI can turn this treatment data into clear clinical summaries, allowing care teams to personalise care based on each patient’s progress. Here’s how NVIDIA and Reset Health describe the value: Anthony Hills, Regional Sales Leader for NVIDIA in the UK & Ireland: “As the UK builds out its sovereign AI capabilities, healthcare and other regulated sectors need practical ways to apply advanced AI to sensitive data. NVIDIA Confidential Computing provides hardware-backed protection for data during processing. Super Protocol provides the complete, ready-to-use software stack for verifiable execution, with a Confidential AI layer that simplifies adoption. Reset Health demonstrates how a healthcare organisation can use AI on external NVIDIA GPUs while keeping treatment data inaccessible to infrastructure providers and retaining control over how it is used.” David Wong, Executive Chairman and Co-founder of Reset Health: “Our opportunity is to use AI to turn treatment data collected through Reset’s platform into more personalised care. Super Protocol enables us to analyse treatment data on external GPU systems while keeping it protected from infrastructure operators, without building confidential computing infrastructure in-house. Our AI team can therefore focus on helping care teams review treatment progress and developing therapeutic optimisation: identifying which combinations of medication and support work best for different patients.” 👉 Read more: superprotocol.com/resources/…

Post attachment
Super Protocol avatar
Super Protocol@super__protocol·Oct 20

How many clicks does it take to launch Sovereign AI Collaboration between 3 independent organizations? Sensitive data. Proprietary AI. No Confidential Computing expertise required. Watch the demo and count for yourself! 👉 youtu.be/x_M1CEU7Y48

Post attachment
Super Protocol avatar
Super Protocol@super__protocol·Oct 14

𝐒𝐨𝐯𝐞𝐫𝐞𝐢𝐠𝐧 𝐀𝐈 𝐡𝐚𝐬 𝐚 𝐠𝐚𝐩 𝐭𝐡𝐚𝐭 𝐫𝐚𝐫𝐞𝐥𝐲 𝐦𝐚𝐤𝐞𝐬 𝐭𝐡𝐞 𝐬𝐥𝐢𝐝𝐞. A sovereign cloud solves a real problem. Data stays in the region, the operator is local, and exposure to foreign jurisdictions is reduced. Now count who may still have access to a client's data, model weights and prompts: the infrastructure operator and its admins, the model provider, the agent vendor, the integrator, the MSP. Each of them is held back by a contract. 𝐀 𝐜𝐨𝐧𝐭𝐫𝐚𝐜𝐭 𝐥𝐢𝐦𝐢𝐭𝐬 𝐰𝐡𝐚𝐭 𝐭𝐡𝐞𝐲 𝐚𝐫𝐞 𝐚𝐥𝐥𝐨𝐰𝐞𝐝 𝐭𝐨 𝐝𝐨, 𝐧𝐨𝐭 𝐰𝐡𝐚𝐭 𝐭𝐡𝐞𝐲 𝐜𝐚𝐧 𝐬𝐞𝐞. What closes the gap is a confidential AI layer. That is what Super Protocol built. Models, agents and data run on CPUs and NVIDIA GPUs in confidential mode, together with the full stack around them: orchestration, storage and key management. Teams deploy their existing applications using standard tools, across GPU clouds, on-prem or hybrid environments. The infrastructure operator, the vendors and Super Protocol itself have no access to the protected data during processing, and the client's security team, partners and auditors can independently verify what's deployed. 𝐖𝐡𝐨 𝐢𝐬 𝐢𝐧𝐬𝐢𝐝𝐞 𝐭𝐡𝐞 𝐭𝐫𝐮𝐬𝐭 𝐛𝐨𝐮𝐧𝐝𝐚𝐫𝐲 𝐨𝐟 𝐲𝐨𝐮𝐫 𝐀𝐈 𝐬𝐭𝐚𝐜𝐤 𝐭𝐨𝐝𝐚𝐲, 𝐚𝐧𝐝 𝐰𝐡𝐨 𝐰𝐨𝐮𝐥𝐝 𝐲𝐨𝐮 𝐫𝐞𝐦𝐨𝐯𝐞?

Post attachment
Super Protocol avatar
Super Protocol@super__protocol·Jun 17

Zoom out on three announcements from the past month and a larger picture emerges. @HPE made NVIDIA Confidential Computing available across its entire AI Factory portfolio – from sovereign deployments to private cloud AI – marking a shift from optional add-on to standard enterprise offering. @Apple extended its Private Cloud Compute privacy guarantees to Google Cloud – the first time Apple trusted a public cloud with its privacy infrastructure, made possible by NVIDIA Confidential Computing. @WhatsApp launched Incognito Chat with Meta AI – completely private conversations processed inside TEEs where not even Meta can read the content. Consumer scale, default privacy. Enterprise & Sovereign AI. Private clouds & Hyperscalers. Consumer platforms. Different markets. Different requirements. Same direction. For organizations evaluating AI platforms, the question is increasingly shifting from "Do we need Confidential Computing?" to "How do we operationalize it across our existing infrastructure?" Super Protocol helps make that transition practical by providing a consistent confidential execution layer across public cloud, private cloud, on-premises, and hybrid infrastructure as a single operational environment – without requiring organizations to rebuild their AI stack or become Confidential Computing experts. The direction is clear. The question is whether your infrastructure is ready for it.

Post attachment
Super Protocol avatar
Super Protocol@super__protocol·Jun 9

Congratulations to @GoogleResearch Health AI Developer Foundations (HAI-DEF) team on the launch of the HAI-DEF Showcase. We're honored to see two confidential healthcare AI case studies developed by Yma Health and Super Protocol featured as the first entries in the "Technical Solutions and Tools" section of the showcase. The two case studies demonstrate how the open medical foundation model #MedGemma 27B can be deployed in real healthcare workloads through Confidential AI. By transforming #TEE-enabled infrastructure into a verifiable confidential execution layer, Super Protocol enables sensitive data to remain protected during processing regardless of where the workloads run. One implementation demonstrates confidential inference on @nvidia Blackwell B200 infrastructure hosted by @nebiusai AI Cloud, with patient EHR data remaining protected throughout processing. MedGemma anonymized patient records while preserving the clinical context and value of the data. Another demonstrates MedGemma fine-tuning on NVIDIA H200 infrastructure using real patient dialogues, achieving a 9.4/10 evaluation score while preserving data confidentiality. As the health tech ecosystem explores the potential of open foundation models and Confidential AI, sharing practical implementation examples on real healthcare data will be essential for accelerating adoption across research, clinical workflows, and precision medicine. 👉 HAI-DEF Showcase - link in comments

Super Protocol avatar
Super Protocol@super__protocol·Jun 2

This fall, Confidential Computing ships at rack scale. 72 GPUs. One TEE. NVIDIA announced Vera Rubin back in March. Now it's ramping to production. Currently, each server forms its own TEE boundary: up to 8 GPUs, 2.3 TB shared memory. Vera Rubin extends this to the entire rack – 72 GPUs, 20.7 TB of shared memory in one TEE. Imagine what model will fit in there! "Everything across this is secure because the AI model is so precious. This is the reason why this entire system obeys confidential computing." – Jensen Huang, NVIDIA CEO, GTC Taipei 2026 What can you do while waiting? We recently updated our GPU + CPU TEE requirements guide, covering all available GPU TEE-capable SKUs, from Hopper to Blackwell, with compatibility details for Intel TDX and AMD SEV-SNP – and what is not TEE-capable as well. Check them out and start deploying Confidential AI today. With Super Swarm – no TEE expertise required. Links in comments 👇

Post attachment
Super Protocol avatar
Super Protocol@super__protocol·May 29

More organizations think they have Confidential Computing than actually do. The CCC's new white paper "3 Degrees of Confidential Computing" makes this concrete: Level 1 migrating to Confidential VMs provides hardware isolation, but as the paper notes, “without integrating remote attestation, it does not meet the definition of Confidential Computing”. However, what is even more telling is the direction in which the paper points beyond Level 3 towards Confidential AI: multi-CVM interactions, AI agent sandboxes, CC-aware network protocols and CC-enforced software provenance. That future isn't theoretical for us. It's what Super Swarm is built on today – self-organizing, mutually attesting GPU clusters that form a single hardware-verified trust domain across cloud, on-prem, hybrid, and multi-cloud environments. Every interaction is independently verifiable. No custom builds. No TEE expertise required. The complexity of operationalizing Confidential AI shouldn't become a project of its own. That's exactly the problem the execution layer should solve. 👉 Link to CCC paper in comments

CC has transitioned from a niche security technology to a strategic imperative for protecting data in use, but its security benefits depend on how deeply it is integrated into your stack. Learn about the practical 3-level maturity model to help orgs roadmap their adoption. ⬇️

Post attachment
Super Protocol avatar
Super Protocol@super__protocol·May 25

The feedback loop healthcare AI never got A radiologist reviews a scan. The AI flags a suspicious mass. The patient is referred, biopsied, diagnosed. The physician closes the loop. The AI never does. Was the flag correct? Was it a false positive? The answer sits in a different EHR, a different department, sometimes a different institution – and arrives months later. Nobody systematically pipes that signal back to the model, because the infrastructure to do so was never built. It is how healthcare has always been organized: services separated, records siloed, pathways fragmented. Imperfect, but functional enough for clinical care – and invisible enough that nobody felt the cost. In most systems, the feedback loop is the first thing you set up. You ship, you measure, you iterate. The signal is fast, systematic, and the model improves. Healthcare AI never got that infrastructure. Rory Pilgrim, Product Manager at Google Research, made an observation in the "Confidentially Yours" episode worth sitting with: The slow feedback loop is not just a limitation. It is an opportunity. If closing the loop leads to better outcomes – fewer missed diagnoses, fewer unnecessary recalls, models that improve on real-world data – institutions have a concrete reason to build the outcome pipelines they never prioritized. AI creates the business case for data infrastructure healthcare never had sufficient reason to build. But acting on that immediately hits a structural wall. Outcome data is patient data – highly regulated and, in most architectures, legally immovable. Traditionally, that immovability is the barrier. The data that would close the loop cannot cross the compliance boundary, so the loop stays open. Super Swarm inverts the problem. Models can live anywhere – on-premise, in the cloud, across institutions. Instead of moving data to the model, computation runs inside a hardware-attested confidential computing environment – where even the operator cannot access what's being processed. Institution-specific outcomes never cross organizational or regulatory lines. The exposure risk is architecturally eliminated. The feedback loop healthcare AI never got is now within reach. 🎥 "Confidentially Yours" with Rory Pilgrim and host Mike Bursell (Advisor, Super Protocol). 👉 Scan to watch the full episode, or find the link in the comments

0:58