Optalysys

Accelerating the future: Photonic compute in transit

[Optalysys @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: https://t.co/kkYuvighzS

[Optalysys @Optalysys]

We're heading out to Colorado in a couple of weeks for the Optical Interconnects and Packaging (OIP) Conference 2026 ✈️

You can catch our Director of Photonics Imon Kundu on stage on the Wednesday (17th June) afternoon for his session 'Programmable Compute-in-Transit using Integrated Photonics', while CTO Robert Todd and CEO @NNOPT will be manning our exhibition stand.

If you'll be there and want to find out more about Compute-in-Transit, or just say hello to the team, drop us a message to schedule a meeting ✉️

Full event details & agenda ➡️ https://t.co/xEyH4ogfNn

[Optalysys @Optalysys]

How can the UK cement itself as a leader in the future of photonics? Our CEO @NNOPT explores for @UKTNofficial ⬇️

https://t.co/TxQySctCLp

[Optalysys @Optalysys]

We're thrilled to be named in @ProlificNorth's 2026 Tech Companies to Watch: Scale-up Edition list 🎉

The list "shines a light on ambitious companies that are not just building big ideas, but turning them into impact with real-world growth, funding success, and market traction whether that be through customers or subscriptions"

Thank you @ProlificNorth and congratulations to all other companies named!

[Optalysys @Optalysys]

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 → https://t.co/dhKYvmL9KV

[Optalysys @Optalysys]

@0xprimex0 Thank you @0xprimex0!

[Optalysys @Optalysys]

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 → https://t.co/pBNdF2oeoK

[Optalysys @Optalysys]

5/ From moving data to compute

to computing as data moves.

Read the paper → https://t.co/rf8curytZc

[Optalysys @Optalysys]

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.

[Optalysys @Optalysys]

4/ Photonics offers an ideal route to realising this model.

Optical systems can apply transformations directly to signals in transit, aligning computation with dataflow

[Optalysys @Optalysys]

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. https://t.co/FsQWqKW2rl

[Optalysys @Optalysys]

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 → https://t.co/YJpVybaa2b

[Optalysys @Optalysys]

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.

https://t.co/qqeCSLpnXK

[Optalysys @Optalysys]

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

[Optalysys @Optalysys]

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

[Optalysys @Optalysys]

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

[Optalysys @Optalysys]

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

[Optalysys @Optalysys]

𝗽𝗵𝗼𝘁𝗼𝗻𝗶𝗰𝘀 𝗶𝘀 𝗼𝗻𝗲 𝘄𝗮𝘆 𝗼𝘂𝘁: move data with photons, integrated on silicon for manufacturability: https://t.co/vgd2NGqCHt

[Optalysys @Optalysys]

https://t.co/hCR8sJygWC

[Optalysys @Optalysys]

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