Optalysys

❯❯❯ 𝘁𝗵𝗲 𝗾𝘂𝗮𝗻𝘁𝘂𝗺 𝗲𝗿𝗮 𝗶𝘀 𝗮𝗰𝗰𝗲𝗹𝗲𝗿𝗮𝘁𝗶𝗻𝗴 — 𝗮𝗻𝗱 𝗶𝘁’𝘀 𝗮𝗯𝗼𝘂𝘁 𝘁𝗼 𝗿𝗲𝘄𝗿𝗶𝘁𝗲 𝘁𝗵𝗲 𝗿𝘂𝗹𝗲𝘀 𝗼𝗳 𝘀𝗲𝗰𝘂𝗿𝗶𝘁𝘆 today’s post breaks down how post‑quantum cryptography ⚛️ + photonic compute 💡 unlock a defence model built for a world where classical encryption won’t survive quantum‑vulnerable chains, HNDL attacks, silicon photonics… this is what orgs need to understand now to stay ahead of Q‑Day full deep dive: eu1.hubs.ly/H0tm_vB0

If FHE is to work at scale, the question won’t be: “How fast can we compute?” It'll be: “How efficiently can encrypted data move?” Find out why photonic compute-in-transit is the critical enabler → eu1.hubs.ly/H0vxPXW0

@0xprimex0 Thank you @0xprimex0!

To anyone landing on this tweet: this paper is an incredibly quick read, only 7 pages total (including references). If you’re working on fully homomorphic encryption (FHE) or quantum-resistant cryptography, take the few minutes to read it. You’ll immediately see why “compute-in-transit” via photonics is such a powerful architectural shift for overcoming the data-movement bottlenecks that currently limit both fields. Worth every second. cc: @veorq @QSECDEF @octra @h33aiofficial @DualityTech @FabricCrypto @enveil_inc @fhenix @zama (and other FHE leaders) * Not a paid promotion, just love the company and work they do.

PQC is a rapidly rising priority in cybersecurity strategies, but it can only be effective against quantum threats if it’s scalable and usable at a global scale. That is where silicon photonics comes in. Read our latest piece for @ITSecurityWire → eu1.hubs.ly/H0vxJbk0

5/ From moving data to compute to computing as data moves. Read the paper → eprint.iacr.org/2026/807

4/ Photonics offers an ideal route to realising this model. Optical systems can apply transformations directly to signals in transit, aligning computation with dataflow

1/ Modern workloads are facing a new compute challenge. Across AI, FHE and post-quantum cryptography, performance is increasingly constrained by data movement, not just arithmetic. This is known as the data movement bottleneck.

FHE is famously compute-intensive, but it’s also incredibly data-movement intensive. As encrypted workloads scale, moving large ciphertexts through digital systems becomes a major bottleneck. Photonics offers a new path: compute while data moves. eu1.hubs.ly/H0v8BTg0

Q-Day isn’t just a cryptography problem... It’s an infrastructure problem. PQC must be secure, scalable and efficient enough for real-world deployment. That is where silicon photonics comes in. Read our latest piece for @ITSecurityWire → eu1.hubs.ly/H0v8qzx0

FHE, PQC and AI may look like different workloads, but they share a common constraint: Large intermediate representations are repeatedly transformed and moved. This results in the data movement bottleneck – where systems are limited by the cost of moving data, rather than computing on it. Our latest paper explores compute-in-transit as a revolutionary architectural solution enabled by photonics. eu1.hubs.ly/H0v8lHB0

that’s why we’re building silicon photonics for sustainable + secure compute and why we’re looking for partners to test/validate in real workloads (ai, privacy-preserving compute, scientific modeling) let's talk... get in touch to find out more

we’ve optimised “compute with electrons” for ~60 years. the next step-change likely comes from changing the carrier... light gives you a different set of trade-offs: bandwidth/parallelism with minimal heat, and a path to higher performance-per-watt at system scale

for context on why chip-level performance-per-watt matters (and why photonics is interesting), check out our last article (link in profile)

data centres already consume a meaningful share of global electricity. and AI demand is pushing the curve up fast 𝗾𝘂𝗲𝘀𝘁𝗶𝗼𝗻: in your org, what’s the real bottleneck to sustainable scaling right now? power availability cooling/thermals networking bandwidth hardware capex something else

𝗽𝗵𝗼𝘁𝗼𝗻𝗶𝗰𝘀 𝗶𝘀 𝗼𝗻𝗲 𝘄𝗮𝘆 𝗼𝘂𝘁: move data with photons, integrated on silicon for manufacturability: x.com/Optalysys/stat…

compute scaling isn’t “ending” so much as changing constraint: power + thermals are now the ceiling moore’s law slowing + dennard scaling broken ⇒ “dark silicon”: you can pack in transistors but you can’t power them all concurrently without blowing the thermal budget

Q Day is a date we don’t control. what we can control is how quickly we move from awareness → execution... hot take: the organisations that win will treat pqc as both a security upgrade and a compute architecture problem (power, latency, scale)

the post-quantum era shouldn’t force a trade-off between security and performance! if pqc-fhe adoption makes systems slower and more expensive, it won’t reach the scale the world needs. that’s the motivation for pursuing photonic acceleration: 𝘮𝘢𝘬𝘦 𝘵𝘩𝘦 𝘴𝘵𝘳𝘰𝘯𝘨𝘦𝘴𝘵 𝘤𝘳𝘺𝘱𝘵𝘰𝘨𝘳𝘢𝘱𝘩𝘺 𝘱𝘳𝘢𝘤𝘵𝘪𝘤𝘢𝘭 𝘶𝘯𝘥𝘦𝘳 𝘳𝘦𝘢𝘭 𝘱𝘰𝘸𝘦𝘳 + 𝘭𝘢𝘵𝘦𝘯𝘤𝘺 𝘤𝘰𝘯𝘴𝘵𝘳𝘢𝘪𝘯𝘵𝘴. eu1.hubs.ly/H0tQlwC0

𝗾𝘂𝗲𝘀𝘁𝗶𝗼𝗻: when you prioritise PQC migration, do you start with (a) long-lived sensitive data, (b) internet-facing protocols, or (c) internal identity/signing infrastructure?

“𝘩𝘢𝘳𝘷𝘦𝘴𝘵 𝘯𝘰𝘸, 𝘥𝘦𝘤𝘳𝘺𝘱𝘵 𝘭𝘢𝘵𝘦𝘳” flips the ROI maths on quantum readiness: the data you encrypt today can become the breach headline years from now.