Radical Semiconductor's upcoming smartcard chip uses a novel cryptographic engine to run any lightweight cryptographic algorithm, with only a NFC signal.
• All-in-one secure element.
• Lower area and power than the secure element in your smartcard today.
• Capable of adding new kernels without redesign.
• Access to PQC, BLS, and other algorithms without ASICs today.
Processing-in-memory is an emerging computer architecture technique used to improve the efficiency of computation involving parallel processing. Radical Semiconductor is among the first teams to use this technique to accelerate cryptographic computation.
Our architecture enables us to skip the traditional process of adding algorithm-specific ASICs to extend cryptographic agility, and instead add new algorithms by loading on new kernels to a single circuit block.
This means our architecture can bring any lightweight cryptographic algorithm to an NFC tap.
Radical Semiconductor taped out its first processing-in-memory test chip in March 2022, called Bookchin.
We wrote and tested kernels for multiple cryptographic algorithms on our architecture, including the post-quantum key encapsulation mechanism Kyber.
We found that Bookchin is able to compute these algorithms, and many other algorithms, incredibly power-efficiently, making our architecture well-suited for low-power devices like smartcards.