Optalysys raises £23M to accelerate photonic hardware for secure AI
UK deeptech startup Optalysys has secured a £23 million funding round to develop next-generation photonic hardware designed to run encrypted AI workloads at unprecedented speed and efficiency. The investment underlines growing investor conviction that optical computing will be critical to overcoming the performance, security and energy limits of today’s AI infrastructure.
While traditional processors rely on electrons moving through silicon, Optalysys is building systems that use light to perform complex mathematical operations central to machine learning and cryptography. By embedding computation directly into optical pathways, the company aims to deliver a platform where sensitive data can remain encrypted throughout processing, dramatically reducing the risk of exposure.
Bringing photonic computing to encrypted AI workloads
The core of Optalysys‘ value proposition lies in combining photonic processors with advanced privacy-preserving computation techniques. Many organisations in sectors such as defence, healthcare and finance are eager to adopt AI models but are constrained by strict data protection requirements. Conventional chips typically require data to be decrypted before processing, creating a vulnerable window in which information can be intercepted or misused.
Optalysys is engineering hardware that can support encrypted computation, including approaches aligned with homomorphic encryption and other advanced cryptographic schemes. In such architectures, data remains encrypted end-to-end, yet AI models can still be trained and executed on top of it. This capability is especially attractive for applications involving highly sensitive datasets, from medical records to classified intelligence.
Why photonics matters for AI and security
Photonics exploits the properties of light to perform operations that are either too slow or too power-hungry for conventional electronics. Optical systems can execute certain types of linear algebra—the backbone of modern deep learning—in a single pass, rather than through billions of discrete electronic operations.
For encrypted workloads, this is particularly significant. Homomorphic encryption and related techniques are notoriously computationally intensive, often rendering them impractical for real-time AI applications. By offloading key mathematical kernels to photonic accelerators, Optalysys aims to make strong cryptography and high-performance AI inference compatible at scale.
The company’s approach also targets the surging energy demands of AI data centres. Optical computation can, in principle, reduce power consumption per operation while increasing throughput. As regulators and enterprises scrutinise the environmental impact of large AI models, more efficient hardware platforms are becoming a strategic priority.
Strategic funding to move from lab to market
The £23 million injection will enable Optalysys to move its technology from advanced prototypes towards commercially deployable systems. The company plans to expand its engineering team, refine its photonic chip designs and deepen collaborations with early customers in high-security and high-performance sectors.
Part of the capital is expected to be channelled into building out a robust software stack that allows developers to integrate the hardware with existing AI frameworks and cloud infrastructure. For photonic computing to be adopted at scale, it must slot seamlessly into current workflows, from model training pipelines to inference services deployed at the edge and in the data centre.
Industry observers note that this round positions Optalysys among a small but influential group of companies racing to commercialise optical AI accelerators. While general-purpose GPU vendors have dominated the first wave of the AI hardware boom, demand is rising for specialised accelerators tailored to specific workloads, including secure computation and privacy-first AI.
Rising demand for secure, compliant AI
The timing of the raise aligns with intensifying regulatory and commercial pressure around data protection. In Europe, frameworks such as the GDPR and the forthcoming AI Act are forcing organisations to reassess how they collect, store and process personal data. Similar debates are unfolding in the US and across other major markets.
Organisations that handle classified or highly sensitive information—such as government agencies, critical infrastructure operators and major financial institutions—face an additional layer of scrutiny. For them, the ability to run advanced AI algorithms without ever exposing raw data could be transformative. Technologies that combine strong encryption with high throughput are increasingly viewed as essential enablers of compliant AI adoption rather than niche research projects.
Optalysys‘ photonic platform is being positioned squarely in this context. By enabling encrypted workloads to run at speeds closer to unencrypted processing, the company hopes to remove a key barrier that has historically limited the use of homomorphic encryption and related techniques in production environments.
Competition and the road ahead
The broader deeptech and quantum-safe computing landscape is becoming increasingly crowded, with startups and incumbents alike exploring photonics, quantum processors and novel semiconductor materials. However, Optalysys is differentiating itself by focusing explicitly on the intersection of AI acceleration and encrypted computation, rather than pursuing photonics as a general-purpose alternative to GPUs.
The newly raised funds will support pilot deployments with early adopters, where real-world performance, reliability and integration challenges can be addressed. Success in these pilots will be critical to convincing larger enterprises and governments to commit to the technology at scale.
If Optalysys delivers on its roadmap, its photonic hardware could become a foundational layer for the next generation of secure AI infrastructure, where sensitive data is analysed at high speed without being laid bare to the underlying systems. For industries seeking to harness the power of AI while maintaining strict confidentiality, that combination of performance and protection may prove decisive.
With £23 million now backing its vision, Optalysys is set to play a prominent role in the race to redefine how AI workloads are computed, secured and scaled in an era of mounting privacy, security and sustainability demands.