OpenLight raised $34 million in Series A funding in August 2025, with the round co-led by Xora Innovation and Capricorn Investment Group. The funding round assigned the company a post-money valuation of approximately $49 million.

Additional participants in the Series A included Mayfield, HPE's Juniper Networks, Lam Capital (the corporate venture arm of Lam Research Corporation), New Legacy Ventures, and K2 Access. This marks OpenLight's first institutional funding round since becoming an independent company.

The company's technology platform originated at Aurrion, a University of California Santa Barbara spin-out acquired by Juniper Networks in 2016. OpenLight was re-established as an independent entity in 2022 with initial funding from Juniper and Synopsys prior to the Series A round.

OpenLight develops photonic application-specific integrated circuits (PASICs) that transmit data using light rather than electrons, achieving higher speeds with reduced heat generation. The company offers three primary solutions to translate this technology into manufacturable hardware.

The first is an open process design kit, which operates as a drag-and-drop parts library within industry-standard design tools. Engineers can access pre-qualified components such as indium-phosphide lasers, electro-absorption modulators, optical amplifiers, photodiodes, waveguides, and couplers. These components can be schematically connected, simulated, and used to auto-generate the mask set for manufacturing.

The second offering includes reference photonic integrated circuit (PIC) designs, such as 800G DR8 transmitters and 1.6T two-fiber platforms. These designs provide validated templates, enabling module manufacturers to proceed directly to transceiver layouts without needing to design each component individually.

The third solution ensures manufacturing through partnerships with semiconductor foundries. OpenLight integrates thin films of indium phosphide onto Tower Semiconductor's silicon-photonics wafer process, delivering warranted first-article performance and facilitating volume production runs without requiring customers to establish direct fab relationships.

The design workflow consists of eight steps, from specification to tape-out. OpenLight delivers fully tested PICs within 8-12 weeks and offers engineering support to address any performance deviations from specifications.

OpenLight operates a platform-as-a-service model in the photonic integrated circuit space, generating revenue through three primary mechanisms: licensing its Process Design Kit (PDK) to customers for one-time fees, providing design and engineering services, and collecting production royalties when customer designs enter volume manufacturing.

The business employs a B2B go-to-market strategy, targeting AI datacenter companies, optical module manufacturers, and applications in LiDAR and quantum computing. OpenLight's open foundry model distinguishes it from vertically integrated competitors by licensing its component library and technology to semiconductor foundries rather than maintaining ownership of manufacturing assets.

The company's cost structure leverages an asset-light approach, outsourcing manufacturing to foundry partners such as Tower Semiconductor. This strategy enables OpenLight to allocate resources toward R&D and customer support while avoiding the capital expenditures associated with owning fabrication facilities.

Revenue growth is driven by both customer acquisition and expansion within existing accounts. As customers progress from the design phase to production, OpenLight generates recurring royalty revenue that scales with manufacturing volumes. The company also increases its addressable market by adding new components to its PDK library and developing reference designs for higher-speed applications.

Intel has shipped over 8 million PICs containing 32 million on-chip lasers and recently introduced fully integrated Optical I/O chiplets co-packaged with Intel CPUs for 800G to 1.6T links. Its competitive advantage stems from captive CPU attach rates and extensive reliability data, though its closed platform restricts third-party access.

Marvell offers 1.6T Light Engines that integrate linear drivers, TIAs, and silicon photonics dies into single packages for its switch and NIC ecosystems. This strategy exerts downward pressure on merchant optics pricing by consolidating the optical chain within Marvell's ecosystem.

Cisco, through its Acacia division, supplies pluggable DSP-plus-silicon photonics transceivers up to 1.6T and collaborates with module manufacturers like Accelink on OSFP-XD DR8 products. This indicates a willingness to license silicon photonics IP beyond internal applications.

GlobalFoundries Fotonix directly competes with OpenLight's foundry model by providing silicon photonics manufacturing services to third-party chip designers. Its established semiconductor manufacturing scale offers cost advantages, though it lacks OpenLight's heterogeneous indium phosphide integration capabilities.

Ayar Labs targets chiplet-scale optical I/O for CPU and GPU co-packaging applications. Its approach prioritizes integration with processor architectures, competing for the same AI datacenter applications as OpenLight.

Tower Semiconductor, while serving as OpenLight's manufacturing partner, also provides competing photonics foundry services to other platform companies, creating potential conflicts as the market expands.

Celestial AI and Lightmatter represent emerging competitors that could bypass discrete optical I/O chips entirely. These companies develop custom PICs to offload memory or matrix math operations, potentially reducing the need for standalone optical interconnect solutions.

Lightmatter's photonic fabric integrates compute and communication functions, while Celestial AI focuses on optical memory interfaces. Both approaches could diminish demand for OpenLight's datacenter interconnect solutions if their integrated technologies prove more effective.

OpenLight is expanding its portfolio from 100G and 200G devices to 400G-per-lane products, enabling 3.2T links for co-packaged optics in AI clusters. This aligns with the AI-cluster optics segment, projected to reach $10 billion by 2026, with linear-drive and co-packaged optics expected to dominate after 2027.

The Series A funding will support the expansion of the PDK to include standard 1.6T and 3.2T reference designs. This allows module manufacturers to avoid custom NRE costs and shorten time-to-market. The expanded component catalog will feature additional laser types, semiconductor optical amplifiers, and monitoring devices.

OpenLight is also developing embedded modulator and laser IP for direct integration onto switch or GPU substrates. This positions the company to capture value in the co-packaged optics market, which represents a new addressable segment exceeding $10 billion.

Hyperscale AI datacenter operators, including GPU manufacturers and cloud providers, represent a key growth opportunity due to their need for terabit optical I/O to connect GPU racks. OpenLight's guaranteed production service model mitigates risk for these customers compared to developing in-house capabilities.

Partnerships with Jabil and TFC establish silicon-to-solution pathways that appeal to automotive and industrial LiDAR OEMs lacking photonics manufacturing expertise. This expands OpenLight's applications beyond datacenters into autonomous vehicle and sensing markets.

The company is also pursuing opportunities in quantum computing and healthcare sensing through partnerships with design houses such as Epiphany. These collaborations enable OpenLight to reach new PIC users without requiring a direct sales force presence in these specialized verticals.

The TFC partnership provides back-end processing capabilities in China, reducing logistics costs for customers serving Asian markets. This geographic expansion addresses the growing demand for optical components in the Asia-Pacific region while adhering to export regulations.

European datacenter growth and regional photonics initiatives present additional opportunities, particularly as customers seek alternatives to fully integrated solutions from US-based competitors.

OpenLight's open foundry model supports geographic scaling through foundry partnerships rather than direct facility investments, enabling rapid entry into new regions as demand arises.

Manufacturing dependence: OpenLight depends on Tower Semiconductor for its heterogeneous indium phosphide integration process, introducing a single-point-of-failure risk. Disruptions to Tower's PH18DA line or shifts in their foundry priorities could materially affect OpenLight's ability to meet customer commitments and maintain differentiation from competitors relying on standard silicon photonics processes.

Vertical integration: Companies such as Intel, Marvell, and Cisco are increasingly incorporating photonics capabilities internally instead of sourcing from independent platforms. This shift toward vertical integration may reduce OpenLight's addressable market for foundry services as large customers develop competing in-house solutions.

Technology obsolescence: Photonic compute companies like Celestial AI and Lightmatter are advancing integrated solutions that combine processing and optical communication functions, potentially reducing the need for discrete optical interconnect chips. If these integrated approaches achieve superior performance, demand for OpenLight's datacenter interconnect solutions could decline significantly.