Niobium raises $5.5M to develop fully homomorphic encryption accelerator chip

Niobium has secured $5.5 million in venture financing to develop a specialized accelerator chip for fully homomorphic encryption (FHE). The company has focused on advancing FHE hardware for commercial applications since its inception and has successfully raised significant non-dilutive capital through federal programs over the past two years.

Homomorphic encryption is utilized to achieve unlinkability, irreversibility, and renewability for biometric template protection schemes. This encryption method allows computations to be performed in the encrypted domain.

The funding round was led by Fusion Fund, with participation from Morgan Creek Capital, Rev1 Ventures, Ohio Innovation Fund, and Hale Capital. Niobium also welcomed a new board member, Shane Wall, Fusion Fund partner and president of CXO Network.

“Fusion Fund is proud to support Niobium’s journey towards commercializing their technology, which we believe will play a crucial role in shaping the future of secure computing,” says Shane Wall, Niobium’s Chairman and Fusion Fund partner.

Niobium has developed an advanced customized silicon solution, a PCIe card based on the company’s FHE system-on-chip, that can integrate with any cloud server to significantly enhance the performance of FHE software solutions, the company says.

According to the company, this solution can improve performance by up to four orders of magnitude. Niobium intends to showcase this technology to its customers and initiate pilot programs in Q4 of 2024.

“Securing this financing validates the groundbreaking work we’re doing at Niobium and propels us into the next critical phase of bringing our FHE accelerator chip to market,” says Kevin Yoder, president and CEO of Niobium.

Ensuring data security and privacy are important factors for biometric systems. Any compromise in protection could lead to severe problems, such as identity theft and unauthorized cross-matching. To address these challenges, biometric template protection has been a field of active research. The IEEE 2410-2021 standard calls for homomorphic encryption to protect the security of biometric templates, for instance.

Biometrics providers working with homomorphic encryption include Infinity Optics, Humanode and NEC.

Deloitte predicted at the end of 2021 that homomorphic encryption and federated learning would become more widely adopted as they enable data to be processed without decryption, thereby maintaining its security.

Article Topics

AI chips  |  biometric template protection  |  data privacy  |  data protection  |  funding  |  homomorphic encryption  |  investment  |  Niobium