Summary
- Opdo has raised $2.3 million in pre-seed funding led by Angular Ventures, with participation from Inflection and Tiny.vc.
- Its AI agents operate within established optical engineering applications, including Zemax and CODE V, while engineers retain control of design decisions.
- Manufacturing measurements can inform subsequent simulations, although the company's claims of substantially shorter development cycles have not been independently benchmarked.
Optical engineering startup Opdo has raised $2.3 million to develop artificial intelligence tools that automate design and simulation work within established engineering software, while connecting the results to physical manufacturing and component testing.
Operating from Berlin and Los Angeles, the company is developing AI agents capable of working with optical design applications including Zemax and CODE V. Rather than replacing specialist simulation tools, its software performs calculations and design tasks within them, allowing engineers to review the resulting models and retain responsibility for technical decisions.
Angular Ventures led the pre-seed financing, with Inflection and Tiny.vc participating. The investment will support additional engineering capacity, development of the company’s AI tools, and expansion of customer projects involving optical systems.
Opdo has spent more than a year working with customers on applications including lasers, imaging equipment, micro-optics, and nanostructured surfaces. The company says its approach has reduced some development and manufacturing cycles from months to weeks, although it has not published independently verified comparative results across those projects.
AI operates within specialist engineering software
Optical engineering involves designing components that control how light travels through a system. Applications range from lenses and laser equipment to industrial sensors, aerospace technology, and systems used in quantum computing.
Developing these components requires engineers to evaluate variables including resolution, focal length, aberrations, thermal behaviour, and manufacturing tolerances. A design that satisfies one technical requirement may perform less effectively against another, requiring repeated adjustments and simulation.
Established optical software provides the numerical tools needed to assess these relationships. Engineers can model how light propagates through a proposed arrangement, examine its performance under different conditions, and determine whether the design is likely to meet the intended specification.
Opdo’s agents work within those applications, helping prepare simulation runs, explore alternative configurations, and check results. Its software can identify missing requirements, establish proposed design steps, and explain calculations so engineers can inspect and challenge the output.
The company says its approach includes independent checks and records of the decisions made during the design process. These are important safeguards because a plausible recommendation generated by an AI model is not equivalent to a result established through a specialist simulation.
Engineers remain responsible for reviewing the design and determining whether it is suitable for the intended application. Opdo also provides access to its own optical engineers, allowing customers to combine automated calculations with additional technical expertise.
The arrangement is particularly relevant to organisations whose optical design requirements are substantial but do not justify maintaining large permanent specialist teams. External engineering capacity can help them progress a project without transferring the entire development process into unfamiliar software.
Manufacturing constraints shape the design
A significant part of Opdo’s proposition involves connecting digital design with physical production. Optical components can perform well within a mathematical model while proving difficult or expensive to manufacture consistently.
Fabrication techniques impose limits on achievable shapes, materials, surface finishes, and tolerances. Temperature changes, assembly conditions, and mechanical stresses can also alter how a completed component behaves.
Opdo incorporates manufacturing considerations into its design process and offers customers access to prototype and component sourcing. The company says finished parts are measured against the relevant design specification before delivery.
Those measurements can then be used to assess the relationship between predicted and actual performance, providing information that may improve subsequent design decisions.
Its longer-term approach involves feeding manufacturing and testing results back into the AI systems used for simulation. This could help identify recurring differences between models and physical components, although the company has not demonstrated that the process produces consistent improvements across every class of optical system.
The distinction between simulation accuracy and manufacturing reliability remains important. Software may accelerate the selection of a promising design, but it cannot eliminate material limitations, production defects, or the need to validate components under their intended operating conditions.
Opdo’s commercial model combines software automation with engineering support and physical component delivery. Consequently, reported reductions in development time may reflect several changes, including faster calculations, additional specialist capacity, and closer coordination with manufacturers.
Without a detailed breakdown of individual projects, those contributions cannot be separated or used to establish a general productivity benchmark.
Specialist AI faces different implementation requirements
The company was founded by Stephan Kuehr, its chief executive, and Yousef Arzhangnia, its chief technology officer. Their experience spans manufacturing software, automation, artificial intelligence, and optical engineering.
Kuehr previously co-founded 3YOURMIND, which develops software for distributed and additive manufacturing. That background connects Opdo’s approach to an established engineering problem: design information must remain useful when a component moves from digital development into physical production.
Unlike general-purpose workplace assistants, specialised engineering agents must interact with applications whose outputs depend on mathematical models and precise technical assumptions. Integration therefore involves more than supplying prompts or retrieving documents.
Opdo says its software can work with additional applications including FRED, Lumerical, and COMSOL. Its technical offering covers simulation orchestration, tolerance analysis, manufacturing checks, and the development of models adapted to individual customers.
Where companies possess commercially sensitive designs, the location and use of engineering information become important procurement considerations. Opdo offers deployment arrangements that can include customer-controlled environments and locally hosted models, alongside commitments concerning the use of customer data for model training.
The suitability of those arrangements will depend on the security requirements of individual projects and the extent to which third-party applications remain involved in simulation and development.
Customers also need to determine how AI-generated decisions will be recorded and reviewed. An incorrect design assumption can create costs further along the manufacturing process, particularly where tooling or specialist materials must be ordered before physical testing can begin.
Opdo has not disclosed the number of paying customers, the value of its commercial contracts, or independently validated performance results. The new financing will support development of its engineering tools and additional customer work, with the company’s commercial progress dependent on demonstrating reliable components and repeatable improvements in the development process.












