Niobium’s Basalisc Accelerator Named one of 2024’s “Next Big Things in Tech”
Today, Niobium’s Basalisc SoC, the industry’s first ASIC-based fully homomorphic encryption (FHE) hardware accelerator, was honored by Fast Company as one of 2024’s Next Big Things in Tech. This award highlights technology breakthroughs that “promise to help define the future of the industries they serve.” We agree that FHE hardware acceleration will be transformative for our category – security and privacy – and we also believe that its impact will reach far beyond embedded technology. Basalisc (and the successive SoC generations already in development at Niobium) will make FHE practical and efficient across industries ranging from finance and healthcare to machine learning and generative AI, ushering in a new era of private cloud computing.
What is private cloud computing, and how is Niobium making it a reality?
The “Encrypted Cloud” is an emerging category of third party cloud computation services that mathematically guarantees data privacy during computation. FHE is the only technology capable of realizing that goal – but it has historically been too inefficient, expensive, and complex to implement at scale. FHE hardware acceleration is changing that equation, enabling cloud providers to build next-gen privacy cloud services.
Niobium was honored specifically for our industry leadership in developing a purpose-built FHE hardware solution. Basalisc is designed from the ground up to handle the large key-switching keys and ciphertexts involved in FHE. The design of the chip eliminates the need for cache-based memory storage, instead devoting extra silicon area to ciphertext storage and processing. Basalisc also includes dedicated hardware for operations that are frequently used in FHE applications, and its highly parallelized architecture allows it to run thousands of operations simultaneously. Put plainly, this is the custom FHE hardware acceleration solution the industry needs to make the privacy cloud a reality,
What will the “Encrypted Cloud” be able to do?
The first generation of cloud computing services has already begun to diversify and specialize. Already, a new generation of providers led by Altair, Cirrascale, and Dataminr have emerged to furnish cloud services specifically designed for AI computation. At the same time, the AI revolution is raising the urgency of privacy-assured computation. Enterprises see the possibility of realizing more insights and breakthroughs than ever before from their data – but they often don’t want to or can’t risk exposure.
The threat is real: data theft is a mainstream and expanding industry. GDPR and emerging privacy laws in the United States are designed to hold companies accountable for mishandling data – and in industries like finance and healthcare, law and policy prevents sharing data across organization and/or international boundaries. The Encrypted Cloud will bridge the gap, allowing data from diverse sources to be analyzed on third party servers with zero risk of exposing the source data.
This capability is going to unlock entirely new applications. Manufacturers could share information with analysts, data brokers or even competitors without exposing the underlying data, allowing them to predict and adapt to supply chain disruptions more accurately. In healthcare, doctors and insurance providers could analyze regret rates, readmittance, and other factors across massive patient populations, creating personalized recommended courses of care that lead to better quality of life. These totally novel applications will form the foundation for FHE adoption, but our belief is that mathematically guaranteed data privacy will someday be the standard.
Privacy is the real “Next Big Thing.”
We are extremely proud to see Basalisc honored as one of the Next Big Things in Tech—but our true goal is to see privacy itself become the next big thing in technology, data policy, and organizational strategy. With widespread FHE adoption, we can end apathy towards data privacy, break away from “good enough” security, and enter a new era where mathematically guaranteed privacy is our default expectation across applications. Now that would be big.
