We are excited to bring Transform 2022 back in-person July 19 and virtually July 20 – 28. Join AI and data leaders for insightful talks and exciting networking opportunities. Register today!
San Francisco-based Opaque Systems, a company enabling collaborative analytics and AI for confidential computing, today announced it has raised $22 million in a series A round of funding.
Confidential computing has been a game-changer for enterprises. It encrypts sensitive data in a protected CPU enclave or trusted execution environment (TEE), giving companies a way to move beyond policy-based privacy and security to safeguard their information in the cloud. However, with this level of encryption, which can only be unlocked with keys held by the client, multiple parties struggle to access, share, analyze and run AI/ML on the data in question. Imagine data scientists and analysts from different teams looking to access patient data to improve different aspects of care.
Opaque Systems’ platform
To solve this challenge, Opaque Systems provides a platform that performs scalable analytics and machine learning directly on encrypted data using familiar tools such as Apache Spark and notebooks.
“What’s unique is the innovation we’ve done in Spark that enables the analytics and AI to run directly on encrypted data, so whether data is at rest, in transit or being processed, there is absolutely no exposure or risk of the data being exposed to entities that should not see it, own it or malicious actors. Opaque integrates seamlessly with TEEs, including enclaves and confidential VMs, with the ability to securely scale clusters,” Rishabh Poddar, cofounder and CEO of Opaque Systems, told VentureBeat.
The enterprise-focused platform builds on top of the open-source MC2 initiative, which was started at UC Berkeley to enable collaborative analytics and AI on confidential and sensitive encrypted data. It allows companies to share the encrypted or blended datasets across workspaces and teams (with set policies) for analysis –– while keeping the encrypted results specific to each party. This way, multiple teams could build a distributed model that informs each party on what they are looking for without ever revealing any specific data that the entity is not authorized to see.
Since its launch, Opaque Systems has seen demand from across sectors for use cases such as money laundering, collaborative drug research, loan stacking prevention and supply chain tracking.
“Our customer base is primarily Global 2000s, including several of the largest banks, financial institutions and healthcare providers in North America. Customers also include consortiums, as many of our use cases are multiparty, so that means one customer could, in turn, represent 3-4 separate entities or discrete organizations,” Poddar said.
Many enterprises rely on homomorphic encryption, where data is converted into ciphertext, and multiparty computation to perform analysis on encrypted data without compromising the encryption. The methods, Poddar says, do work but are also accompanied by high resource consumption and performance overhead.
“Through extensive research, we’ve seen that these technologies are far from being practical for scalable, highly secure data analytics and machine learning that is required to execute critical business cases. Some of these solutions can sustain simple computations, but it soon becomes prohibitive in performance for scalable data analytics and ML training,” he added.
With this round of funding, led by Walden Catalyst Ventures, Opaque Systems will focus on expanding its team and building out its offering to serve the accelerating market demand for collaborative analytics and AI in confidential computing. According to Gartner, over 50% of organizations will adopt privacy-enhancing computation to process sensitive data and conduct multiparty analytics by 2025.