SoM vs SBC: How to Choose Between Core Boards and Single-Board Computers (2026)
Key takeaway: SoM and SBC are not better or worse — they are two different R&D investment models. An SBC buys "off-the-shelf speed"; an SoM buys "customization and lifecycle". Choosing wrong rarely hurts at the prototype stage — the cost surfaces at the enclosure and mass-production stages. Decide this at project kickoff.
1. Two concepts, clearly defined
- SBC (Single-Board Computer): CPU, memory, storage, power and I/O on one board — a complete "small computer" out of the box. Typical use: drop it into an enclosure, or develop on top of it.
- SoM (System on Module): the parts you cannot drop (CPU, memory, storage, PMU) on a small module, with I/O brought out through board-to-board connectors. Your team (or your vendor) designs a carrier board with exactly the interfaces, protection and mounting your project needs.
One line to remember: what changes goes on the carrier; what stays goes on the module. Once you map your project's "changing vs. fixed" parts, half of the selection is done.
2. Comparison across six dimensions
| Dimension | SBC | SoM + custom carrier |
|---|---|---|
| Time to start | Fastest — buy and develop | Carrier design and fab adds weeks |
| Upfront cost | Low (just buy boards) | Medium (carrier design + prototyping) |
| Customization | Fixed I/O; unused ports are dead weight, missing ports cannot be added | Tailored to the project: protection, power specs, mounting holes all customizable |
| Mechanical fit | Enclosure adapts to the board | Carrier adapts to the enclosure; better thermal and routing freedom |
| Production QC | Dependent on vendor batches | Module batches controlled + own carrier production line |
| Upgrade path | Swapping boards restarts the project | Keep the carrier, swap the module to upgrade the platform |
3. When an SBC is the right call
- Validation stage: software, algorithms and UX feasibility first; hardware not frozen yet;
- Very small batches: tens of units — carrier NRE cannot be amortized;
- Loose mechanical constraints: standard racks/cases, size-insensitive;
- Extremely tight budget and schedule: ship with unused ports rather than delay.
4. When SoM + carrier wins
- Hard mechanical constraints: wall-mount, embedded or shaped enclosures where port placement and holes must match;
- Interfaces need trimming and protection: industrial sites need isolation, surge protection and wide-range power that consumer SBCs cannot handle;
- Long product lifecycle: for 3+ year products, the SoM model lets you upgrade the platform without re-routing the carrier;
- Volume cost sensitivity: above several hundred units, removing unused components lowers unit cost;
- Multiple products on one platform: one SoM with different carriers becomes a product family.
5. The recommended hybrid path
In practice the two are not exclusive: validate software on the vendor's evaluation board while the carrier design proceeds in parallel. Software teams get a board in week 1–2 for driver and application work; hardware teams finish schematics, layout and fab simultaneously. You get the speed of SBC and the fit of SoM, and switch to the SoM solution seamlessly after the prototype stage.
Hongyin Technology supports exactly this path: validate on an evaluation board first, then we handle SoM selection and carrier design. You may also customize only the carrier and reuse our standard SoMs to further reduce investment.
6. Common pitfalls
- Myth 1: "SoM is always more expensive." Total cost includes carrier NRE and the hidden costs of unused SBC ports and later revisions. At volume, SoM solutions often win.
- Myth 2: "Start with an SBC, convert later." With hard mechanical constraints, SBC conclusions do not transfer — carrier work is unavoidable, so start it in parallel.
- Myth 3: "Any board-to-board connector will do." The connector defines signal integrity and mating life — it is the lifeline of an SoM solution and deserves careful evaluation.
7. FAQ
- Q1: Can carrier design be fully outsourced? Yes. Provide a requirement list (interfaces, power, mechanical constraints) and the vendor delivers schematics, layout, prototyping and bring-up — a carrier ready for assembly.
- Q2: What if the SoM is discontinued? Check the vendor's supply commitment and replacement parts. Another advantage of SoM: swap the module, keep the carrier — migration cost is far lower than with an SBC.
- Q3: Can we move from an SBC pilot to SoM later? Software (OS, drivers, apps) largely carries over within the same SoC family; mechanicals need re-evaluation.
- Q4: Should the carrier include redundancy? Reserve test points or pads for expandable buses (USB, UART, GPIO) — a revision costs far less than a new prototype run.
8. Related solutions & contact
Related solutions: Industrial Vision Inspection、Smart Retail Kiosk、Security NVR Recording、Camera Module Customization. Further reading: SoM Customization Cost & Timeline、MOQ & Low-Volume Production.
Shenzhen Hongyin Technology Co., Ltd. (HONGYIN TECH) provides embedded project customization on RK3588/RK3576/RK3568 platforms — free requirement review, engineer-direct communication and low-volume ordering. Send your requirement list and receive selection advice and quotes within 48 hours.
phone +86 13728726926 | email sales@hoyin-tech.com | website www.hoyin-tech.com | address: Room 301, Building A, Fanrong Science Park, Sanwei Community, Hangcheng Street, Bao'an District, Shenzhen, China.
Hongyin Tech provides full-stack SoM/SBC customization on Rockchip & HiSilicon platforms with Android/Linux. Engineers respond within 1 business day.
Get a Free Quote →