An FPGA is often the right choice early in a development programme. It allows engineers to refine functionality and adapt designs as requirements evolve. The challenge comes when flexibility is no longer the primary requirement but still shapes the architecture. Here, Ross Turnbull, Director of Business Development at ASIC design and manufacturing specialist Swindon Silicon Systems, explains how to recognise when a design has outgrown an FPGA approach and when it is time to consider an ASIC.
As designs stabilise and production moves beyond initial deployment, the priorities begin to shift. Efficiency starts to carry more weight than flexibility. Cost becomes more visible and decisions at the device level begin to shape the wider system. These changes tend to appear through a few common signs.
1. Production volumes are increasing
The first signal is usually economic. FPGAs avoid the upfront non-recurring engineering investment associated with custom silicon, making them ideal for early development and lower-volume production.
The limitation is that the cost model does not always evolve favourably with volume. Each device retains the cost structure of a programmable platform, even after the architecture has stabilised. This tends to surface during cost-down exercises, when teams find there are few practical levers left within an FPGA-based design. At that point, the discussion shifts. ASICs introduce a different model, where upfront investment is higher, but per-unit cost reduces as production scales. For products entering sustained production, this trade-off can become increasingly attractive.
2. Efficiency constraints are affecting the design
Efficiency issues rarely appear as a single failing metric. They tend to emerge through tighter thermal margins, reduced performance targets, additional cooling requirements or increasing pressure to lower operating costs.
The reason for this comes back to how FPGAs are built. Their programmable architecture provides valuable flexibility, but it also introduces architectural overhead compared with hardware designed for a specific function. Although optimisation can improve performance and efficiency, it cannot remove the inherent characteristics of a programmable device. Over time, teams may find they are spending more effort managing these constraints than improving the wider system.
An ASIC gives designers greater scope to optimise efficiency at the system level. This changes what is achievable, particularly in applications where power consumption, thermal performance, form factor or operating costs have become important design priorities.
3. Integration pressure is increasing across the system
Power is not the only constraint that changes as products mature. System complexity tends to increase over time. Additional memory, interfaces and supporting components increase bill of materials costs, PCB complexity and validation requirements. Eventually, integration becomes less of an optimisation and more of a constraint.
An ASIC allows more functions to be integrated into a single device. This can reduce component count and remove some of the complexity that builds up as a design develops. The result can be a system that is simpler to test, manufacture and maintain.
4. Lifecycle requirements are outlasting component strategies
In many applications, the product is expected to last longer than the components it is built from. This is particularly true in industrial, medical and infrastructure systems, where deployments are measured over decades. FPGA portfolios continue to evolve. New devices replace older ones, and migration paths are typically available, but they are not without cost. Redesign, revalidation and, in some cases, recertification introduce risk that sits outside the core product architecture.
A full turnkey ASIC brings more of the lifecycle control in-house. It reduces reliance on external roadmaps and makes planning more predictable. For organisations working on long deployment cycles, this stability carries weight. This is particularly relevant where products require qualification, certification or controlled change processes, because even small hardware changes can have significant downstream impact.
5. The design is stable, but optimisation still matters
The final signal is less obvious, but often the most important. Early in development, flexibility is essential because engineers need to test and refine the system. Once the design is established, the focus often moves towards improving efficiency, reducing power consumption and getting more from the hardware already in place.
At this stage, an FPGA can start to become a limitation. An ASIC is designed around the specific requirements of the application, allowing the hardware to be optimised around the functions that matter most. It can improve efficiency, performance and integration, while also giving companies more control over how key functionality is implemented.
Moving from FPGA to ASIC is rarely about one issue alone. Usually, it is the combination of cost, power, integration and lifecycle pressures that leads teams to reconsider the architecture. FPGAs remain essential for prototyping, early production and applications requiring flexibility. The challenge is recognising when those priorities change. When this happens, the architecture that helped bring a product to market may no longer be the one best suited to take it forward.
Considering migration from FPGA to ASIC? Speak to Swindon Silicon’s ASIC specialists to explore whether custom silicon is the right approach for your application.
Read other recent instrumentation news: https://instrumentation.co.uk/category/news/
