I was honored to speak last week at the Confidential Computing Summit 2024, where we discussed how confidential compute can be used to help us scale generative AI effectively and with confidence. By enabling multi-party, privacy-preserving data to enhance and activate gen AI models, confidential compute is unlocking untapped opportunities for companies and entire industries. Thank you to all who joined me to dive into the three critical areas where generative AI amplifies the value of Confidential Computing: data-sharing, IP protection, and secure inferencing. And a big shout out to my colleagues Teresa Tung and Giuseppe Giordano who also took the stage to explore this rich and evolving landscape. Data is a valuable and powerful asset for companies in the age of gen AI. We're excited to be part of any conversation around how to use it—safely and effectively—to accelerate ongoing reinvention. Mauro Capo | Aaron Fulkerson | Raluca Ada Popa #ConfidentialComputing #GenAI #DataPrivacy
Lan Guan
3w
This will be a huge differentiator for Gen AI adoption!
Raju Ivaturi
3w
At this time as inference is consumable as a service, data plays a key differentiating role for any enterprise. Multiparty collaboration becomes a key challenge due to trust, this is a huge opportunity for confidential compute to help bridge the gap in a safe and secure way.
Confidential computing really helps bring to life shared data, shared insights and shared analytics and quick action. It is a true service model that can really help the community at large across multiple use cases where a single entity cannot do the job effectively or falls short and relies on true collaboration of data and knowledge sharing (from other entities) without worrying about compromising on trade secret, confidential breach, and other aspects that are stopping organizations from truly collaborating and leveraging cloud as the basic compute platform.
Congratulations Karthik Narain on the wonderful session! The potential for multi-party, privacy-preserving data to enhance gen AI models is indeed a game-changer for many industries. Confidential computing plays a crucial role in enabling secure inferencing of foundational models pre-trained on IP data from competing enterprises. This technology ensures that sensitive information remains protected while allowing collaborative innovation. Would you recommend, for example, in life sciences, forming a consortium where enterprises can pool their data to train foundational models for Drug Discovery and Development. This approach not only secures data but also shares the benefits and value across all members, fostering a collaborative environment to bring drugs to market faster and improve patient lives.
Absolutely agree—confidential computing indeed enhances GenAI scaling, particularly in areas like data-sharing, IP protection, and secure inferencing. It’s also an exciting opportunity to refine and optimize areas such as performance overhead and integration challenges. Looking forward to seeing how these developments will continue to unfold!
Market Analyst at Everest Group | Strategic Advisory & Research
3wI am happy to see confidential computing receiving the recognition and visibility it deserves. Enterprise awareness of confidential computing has significantly increased in recent years, leading to unique use cases that extend beyond confidential AI, including generative AI. Some interesting use-cases I have seen include effective ad targeting and customer insights in advertising industry, proprietary algo-trading by banks, and RWD sharing for confidential computing based MPC in HLS vertical. Additionally, emerging use cases in edge devices are noteworthy too, with Microsoft CoPilot+ and Apple's recent announcements on privacy-preserving technologies. It's an interesting space and the hype has just started to build.