Technology thesis · Computing Infrastructure
low conviction conceptHomomorphic encryption
FHE stays niche behind confidential-computing TEEs through 2027: inference runs 4–6 orders slower than plaintext, viable only in banking, government and blockchain until accelerators close the gap.
Position maintained continuously · last reviewed Jun 24, 2026
The thesis
Hardware FHE accelerators (Cornami, Optalysys, Intel HEXL) could close performance gap 2027-2028
Cornami chiplet + Optalysys photonic + Intel HEXL targeting 100-1000x speedup vs CPU FHE. If 10-100x of plaintext achievable by 2027-2028, FHE becomes commercially viable alternative to Confidential Computing for stronger cryptographic guarantee + no TEE trust. Pre-commercial 2026.
State of the art (2026)
Homomorphic encryption splits cleanly in 2026. The commercial momentum has moved to blockchain: Zama, founded by Pascal Paillier and Rand Hindi, became the first FHE unicorn in June 2025 on a $57M Series B led by Pantera and Blockchange (over $150M raised, $1B+ valuation), and is shipping its FHEVM confidential-smart-contract protocol on Ethereum. Apple's open-source swift-homomorphic-encryption (BFV scheme) runs in production behind iOS Live Caller ID Lookup. Duality, Inpher and Microsoft SEAL sustain a roughly $300M niche in banking, healthcare and government. But for AI inference, FHE remains 4–6 orders of magnitude slower than plaintext, and NVIDIA's confidential-computing TEEs on Hopper and Blackwell – AES-256-GCM with attestation at single-digit overhead – win nearly every practical workload. Hardware accelerators from Cornami, Optalysys and Intel HEXL are the swing factor through 2027–2028.
Duality + Zama + research grants sustain niche commercial through 2026
Duality (~$60M raised) stewards OpenFHE with DARPA + US government work; Inpher secure-MPC tech acquired. Zama (FHE unicorn) ships open-source TFHE-rs + Concrete. Roughly $300M commercial market in 2026, weighted to government, banking and blockchain. DARPA/IARPA/ARPA-H grants sustain the niche even without a performance breakthrough.
NIST PQC adjacency provides ecosystem tailwind
NIST PQC FIPS 203 ML-KEM Kyber uses lattice-based cryptography related to FHE. Combined PQC + FHE ecosystem could grow if hardware accelerators succeed. PQC migration 2024-2030 provides cryptographic ecosystem development tailwind for FHE.
Everything below is live inside CanaryIQ
The full analysis behind the verdict — the structure is real; the content unlocks when you log in.
Signal stack
Evidence stacked leading → lagging
Technology-native KPIs
Metrics that predict trajectory, tracked over time
Landscape map
Who builds what — and who depends on whom
Catalyst calendar
Dated events that will move the position
Technology roadmap
Milestones on the path to maturity
Watchlists
Companies, people and papers — each with a remove-by condition
Decision frameworks
The same call, framed for your desk
Thesis changelog
When our view changed, and why
Change our mind
5 disconfirming conditions
The rest is inside
You've read the verdict. The file is much deeper.
The full signal stack, technology-native KPIs tracked over time, the landscape of who depends on whom, the dated catalyst calendar, decision frameworks for every desk, live watchlists and the changelog of every time our call on Homomorphic encryption has changed — all live inside CanaryIQ.