Summary
Choosing between a chip-down design and a system-on-module (SOM) for an industrial product is really a lifecycle risk decision, not just a bill-of-materials comparison. Security obligations, regulatory exposure, supply chain longevity, post-launch support costs and time to market all matter over a 10-year or longer product life. For most moderate-volume industrial products, a SOM on a platform like Digi ConnectCore lowers that risk and gets the product to market sooner. Chip-down tends to make sense only at very high volumes, under extreme size or power constraints, or with a mature dedicated hardware and security team.
The build-vs-buy decision when designing and building embedded systems is no longer just about assembling embedded hardware and getting a prototype running. For industrial products, the chip-down design vs. system-on-module (SOM) decision determines your security posture, your regulatory exposure, your exposure to component shortages, and your support costs for the next 10 to 15 years. It’s also crucial for time to market. Choose the starting point that best manages those risks, not just the one with the lowest bill of materials.
Digi ConnectCore modules,
TrustFence security, Digi Embedded Yocto, and
ConnectCore Cloud and
Security Services are designed to take on much of that risk. The real payoff comes from getting to market sooner.
The Build vs. Buy Decision Has Changed
For years, the question of whether to build a custom PCB design around a processor (chip-down) or start with a SOM was framed as engineering effort versus unit cost. That framing no longer holds for industrial embedded products.
Industrial devices live in the field for a decade or more. They sit on networks that matter: factory floors, energy infrastructure, medical environments, transportation systems. They are subject to regulations that did not exist when many teams last chose a platform. And the last few years have shown how quickly a single unavailable component can stall a product line.
So the real question is: which starting point gives you the lowest total risk and total cost across the entire product lifecycle? That question concerns engineers, product managers, program leads and finance alike, and each group sees a different part of the answer.
It's a Spectrum, Not a Binary Choice
Most teams think of two options, but there are really three starting points:
| Starting point |
What you own |
Best fit |
| Chip-down |
Everything: schematics, layout, BSP, security, certification, sourcing |
Very high volumes, extreme cost or size constraints, deep in-house expertise |
| System-on-module (SOM) |
Carrier board, application software, product-level compliance |
Differentiated industrial products needing speed and long-term support |
| Single board computer (SBC) |
Enclosure, application, integration |
Rapid prototyping, lower-volume products, proof of concept |
Five Questions to Answer Before You Commit
1. Do you have the right engineering resources to build a chip-down product, and are they truly available?
With connected embedded product development, scope is almost always larger than anticipated during the ideation phase. Be honest about whether the team has built this class of product before, whether other projects can truly be set aside, and whether you have deep expertise in the less glamorous disciplines: high-speed layout, power management, RF design, secure boot, and long-term Linux maintenance. RF design complexity alone can be a showstopper if you don’t have the core RF engineering skills on your team. Chip-down means staffing all of these design, build and test competencies, not just the ones you enjoy.
How Digi helps: A ConnectCore SOM arrives with the processor, memory, power management and (on wireless variants) radios pre-certified, integrated and tested, and Digi Embedded Yocto provides a validated Linux board support package from day one. Your team can concentrate on the application and carrier board instead of the compute subsystem. Expert Technical Support and Particle Design Services are also available when you need deeper engineering help.
2. What does your timeline really look like?
Embedded system design projects without a defined timeline tend to be the most expensive, because teams work full time at less than peak efficiency for far longer than necessary. A pre-integrated, pre-tested module compresses the schedule in practical ways: software development can begin on day one on the target platform, application board design can follow a proven reference design, and hardware and software work can proceed in parallel instead of in sequence.
How Digi helps: Development kits and single board computers let software teams start on real target hardware while the custom carrier board is still being designed. Digi also provides schematics, Gerber files and bills of materials for its reference boards, so carrier board design starts from proven work rather than a blank page.
3. Have you modeled total cost of ownership, not just bill of materials?
A chip-down design may show a lower unit cost on paper. Pressure-test that assumption by including:
- Non-recurring engineering for hardware, software and test development
- Prototype spins and re-spins
- Certification and compliance costs such as radio, safety, and other industry-specific requirements
- Manufacturing test development and yield ramp
- Ongoing software maintenance and security patching over the product's life
- Obsolescence management and redesign risk
- The revenue you forgo for every month that your product is late to market (see below)
Where volumes are moderate and lifecycles long, the module's higher unit cost is frequently outweighed by the avoided engineering and lifecycle spend.
How Digi helps: Pre-certified Wi-Fi and Bluetooth on wireless SOMs remove much of the radio certification effort. ConnectCore Cloud Services are offered in Essential and Premium tiers, with subscription and pay-per-use options, so ongoing management costs are predictable and scale with the fleet.
4. Can you meet your security and regulatory obligations for the full product life?
This is where the table stakes have changed most. Security in connected industrial product development is no longer a feature to add before launch. It is a continuing obligation, starting with embedded design and prototyping and continuing through the product lifecycle in the field.
Regulations affecting connected products such as the EU Cyber Resilience Act, the EU Radio Equipment Directive cybersecurity requirements, and industry frameworks like IEC 62443 for industrial automation place responsibility on the manufacturer to build products securely, track and disclose vulnerabilities, and provide security updates throughout the support period.
For an IoT product development team, that translates into concrete work:
- Secure and authenticated boot, encrypted storage, and protected interfaces
- A trustworthy, signed over-the-air update mechanism
- A software bill of materials (SBOM) and a process for monitoring new vulnerabilities in every open-source component you ship
- A vulnerability handling and disclosure process
- Hardware-rooted device identity and key management
Building and sustaining all this in-house across a custom chip-down design is a significant, permanent commitment.
How Digi helps: Security is addressed at three levels: NXP and STMicroelectronics silicon with hardware security foundations, the Digi TrustFence framework (secure boot, encrypted file systems, authenticated access and secure firmware updates), and ConnectCore Security Services (SBOM, vulnerability scanning and CVE reporting) paired with Cloud Services for secure over-the-air patching. Digi positions this combination as supporting CRA Annex I requirements. Each layer is described below.
5. Can your supply chain support the product for 10 years or more?
The shortages of recent years left many teams with products they could not build. For industrial products, supply chain longevity needs to be a selection criterion, not an afterthought.
Questions to ask any platform vendor, and to ask yourself if you go with a chip-down vs SOM design:
- What is the committed availability window for the processor and the module?
- Does the processor come from a vendor with an industrial longevity program?
- How are end-of-life notices handled, and what is the last-time-buy process?
- Is there a clear migration path to a successor part with pin-compatible or software-compatible options?
- Who manages multi-sourcing and component-level risk: you, or your supplier?
With a SOM, much of the component sourcing, obsolescence tracking and lifecycle planning shifts to a supplier whose core business is exactly that. With chip-down, every capacitor, memory device and power IC on your board becomes your responsibility.
How Digi helps: Digi designs ConnectCore SOMs for industrial longevity, building on NXP and STMicroelectronics processor families that are positioned for long-term industrial supply, and manages module-level component sourcing, last-time-buy communication and successor planning so you do not have to. Modules are designed and verified for industrial conditions, including extended temperature ranges and reliability testing such as IEC 60068 and HALT on current designs.
The Costs that Arrive After Launch
Many teams budget carefully for development and then underestimate what follows. After launch you are still responsible for the following:
- Firmware updates
- Security patches
- Bug fixes and kernel updates
- Responding to new vulnerabilities in third-party components
- Managing fleet deployment
- Monitoring and remote updates
- Handling component changes without breaking certification
Over a 10-year lifecycle, these costs can rival or even exceed the original development investment. Product managers should ask for this lifecycle view up front, as it is the part of the business case most often missing.
This is also the area where a platform approach pays off most. Updates, vulnerability tracking and fleet management are not one-time engineering tasks; they are recurring services. Digi packages them as exactly that, as described below.
The Revenue Side: What a Faster Launch Is Worth
Most SOM vs chip-down design analyses are written as cost comparisons. That leaves out the largest number in the business case: the revenue and margin you earn, or forgo, depending on when your product reaches customers. Product managers and finance leaders should put this on the same page as the engineering estimate.
Every month of delay has a price. A simple way to frame it is: Cost of delay per month = expected monthly gross profit at steady state, adjusted for how quickly sales ramp.
The value of speed in rolling out your embedded solution comes from several sources:
- Earlier revenue. When the end of your market window is fixed (a competitor launch, a customer program date, a regulatory deadline), every month you start earlier is a month of sales you otherwise never get back.
- Competitive position. Industrial customers qualify suppliers and tend to stay with them. Arriving first can mean winning the design-in, and the installed base that follows.
- Pricing power. Early in a market, you are less exposed to price pressure from followers.
- Faster learning. Earlier field deployments produce real customer feedback that shapes the next release.
- Engineering capacity. Engineers who are not busy re-creating a compute platform are free to work on the features that differentiate your product, or on the next product.
- Market access. For products sold into regulated markets, missing a compliance date can block revenue entirely, not just delay it.
Get Our White Paper
Learn more about chip-down vs. SOM development
Download PDF
An Illustratrive Example of SOM vs Chip Down Design
| Assumption or result |
Hypothetical value |
| Annual volume |
20,000 units |
| Expected steady-state gross profit for the product |
$200,000 per month |
| Launch acceleration from module-based approach |
6 months earlier |
| Incremental gross profit from the earlier launch (market window fixed, ramp ignored) |
$1.2 million |
| Avoided NRE (compute subsystem hardware, BSP, security bring-up) |
~$1 million |
| Avoided certification and re-spin cost |
$150,000 |
| Avoided sustaining Linux and security maintenance (1 FTE at $200k/yr x 10 yrs |
~$2 million |
| Total avoided cost |
~$4.35 million |
| 10 Year Cost Analysis |
Hypothetical value |
| Module cost premium versus chip-down BOM |
$15 per unit, or $300,000 per year, $3 million over 10 years |
| Delta Cost (avoided cost – cost premium) |
~$1.35 million |
In this hypothetical case, the module cost premium totals about $3 million over ten years, but the earlier launch, the avoided engineering and certification work, and the avoided sustaining effort are worth about $4.35 million. The balance tips toward chip-down only above roughly 29,000 units per year. Your numbers will differ, but the method holds: compare the lifecycle premium with lifecycle savings plus the value of time to market, not BOM with BOM.
Teams should also model the chip down case honestly. At very high volumes with a long, stable market window, the per-unit premium can matter more.
When Does Chip-down Design Make Sense?
A modular approach is not always the answer. Chip-down product development can make sense when:
- Volumes are very high and per-unit cost dominates the business case
- Size, power or form factor constraints cannot be met by available modules
- You have a mature, dedicated hardware and security team with an established process
- You already maintain a long-lived software platform for the processor
Even then, many teams prototype with a system-on-module vs. a chip-down design first, and even launch on a module design. Doing so lets them validate the product and the market before committing to a custom design. Because Digi Embedded Yocto and the TrustFence framework run across the ConnectCore family and are built on standard NXP and STMicroelectronics silicon, software and security work done on a SOM carries forward if you later move to a custom board.