01
Signal admission
Hardware-rooted evidence for live-source, provenance, freshness, session, device, and policy state before consequential compute accepts an observation.
Fabless Semiconductor · Trusted Biological Compute · Trust Hardware
Living Cipher builds trusted biological compute.
What deserves to be computed—and what deserves to become hardware?
Living Cipher is an architecture and research company working at the boundary between biological evidence, compute architecture, hardware trust, and silicon.
We determine what computation deserves to become hardware, what evidence deserves admission to it, and how the resulting system can remain trustworthy from signal acquisition through execution.
“
Trust begins before the chip and does not end when computation finishes.
The Living Cipher thesis
What We Build
Trust boundaries most systems leave implicit.
A secure processor can execute authorized computation perfectly on an observation that should never have been trusted. Living Cipher moves evidence upstream and keeps it bound to the computation that follows.
02
Post-quantum trust in fabric
Inspectable cryptographic datapaths and hardware-rooted evidence designed for protected FPGA, SoC, and future ASIC paths.
03
Multi-die trust
Admission, identity, containment, provenance, revocation, and lifecycle boundaries across heterogeneous dies and chiplets.
04
Architecture-to-silicon
Partitioning, RTL evidence, implementation closure, PPA, interface strategy, substrate choice, and the boundary between programmable compute and fixed silicon.
The Trust Trajectory
Evidence before admission. Evidence after execution.
Trust is not a single cryptographic event. It is a chain of decisions about what may enter, what may execute, what may influence the system, and what can later be proved.
Observe
Qualify
Admit
Compute
Prove
A trusted channel cannot repair an untrusted source. A trusted chip cannot retroactively authenticate a false observation. The evidence boundary has to be architected where meaning still exists.
Biological Compute & Silicon Strategy
What biological computation has earned the right to become hardware?
Biological compute systems often reach hardware decisions while the signal model, algorithm, or workload is still evolving. A successful prototype establishes possibility; it does not automatically establish what should be fixed in silicon.
Living Cipher evaluates the path from biological signal to programmable compute, heterogeneous SoC, dedicated acceleration, and custom silicon—identifying what should remain flexible, what is sufficiently recurrent to harden, and where capital becomes difficult to reverse.
Experimental programs can establish the evidence for those decisions while client-specific protocols, datasets, implementation details, and architectures remain protected.
Architecture & Capital
Architecture determines where capital becomes irreversible.
The move from software to FPGA, module, custom acceleration, chiplet, or ASIC is not only a technical decision. It converts assumptions into NRE, supplier dependencies, qualification burdens, rights commitments, and increasingly expensive reversibility.
Living Cipher traces the technical premise, evidence state, dependent commitment, and realistic recoverability before capital is released.
Living Cipher Lab
Public evidence of the architecture, without publishing the private machinery.
The Lab turns the trust thesis into observable behavior across signal admission, compute admission, and the hardware-rooted execution path.
LIVE SOFTWARE PROTOTYPE
Signal Witness
Live-source input → nonlinear witness state → admission evidence → verifier-facing receipt. Compare subsequent observations and see receipt-level deviation without exposing the private witness machinery.
PUBLIC SIMULATION
Chiplet Trust & Enrollment
A heterogeneous component presents identity and state evidence, is enrolled into role and policy, and is admitted, constrained, quarantined, or revoked before participation.
Decision Cases
The hardware can work and the system can still say no.
Decision Cases show how trust, infrastructure, rights, deployment, and capital can overturn a technically attractive implementation.
CASE 01
A trusted chip receives an untrusted signal
Strong device trust does not establish source truth. Preserve evidence before context disappears.
CASE 02
The round funds the prototype
The architecture may nevertheless require a hardware company: boards, qualification, packaging, supply, lifecycle, and silicon.
CASE 03
Integration becomes transfer
API access, bounded IP, reusable implementation rights, and strategic ownership are different assets with different option value.
Evidence, Not Assertion
Architecture grounded in what actually closes.
Public evidence is deliberately sanitized. Protected equations, scoring machinery, implementation details, and client-confidential architecture remain private.
53.4 MHz
Post-route ML-DSA datapath maximum frequency at a slow, low-voltage, high-temperature corner on PolarFire SoC.
Zero
Vendor cryptographic cores in the custom trust datapath; security-critical behavior remains inspectable at RTL.
End to end
Requirements, architecture, RTL, simulation, place-and-route, timing closure, target integration, and pre-fabrication evidence.
Living Cipher Analysis
Public arguments. Private decision machinery.
Living Cipher Analysis publishes the concepts and evidence needed for serious technical debate: irreversibility, signal veracity, compute sovereignty, capital allocation, rights, and infrastructure control.
A sovereign fabless semiconductor and research company.
Living Cipher LLC holds its intellectual property directly. Semiconductor IP, research, and advisory work are separated through explicit ownership, confidentiality, and background-IP boundaries.
Founder
Dana Xiadani
Founder & Chief Architect
Structure
Living Cipher LLC
Massachusetts · U.S.
Work
Trust hardware · architecture
Research, semiconductor IP, technical diligence
Contact
dana@livingcipher.com
Confidential architecture inquiries
Confidential Inquiry
What decision is approaching?
Start with the decision, not a generic request for services. What becomes difficult to reverse, and what evidence exists today?