Back to blog
Company Manifesto Sri Lanka

Hello, world. Here's why we build hardware in Sri Lanka.

Our first post: what XD TRONICS is, and why we think Sri Lanka should design and build more of its own electronics.

· 10 min read · XD TRONICS
$ ./xdtronics --hello
studio    PCB · firmware · CAD
build     SMT to 0402 · FDM · resin
store     parts off our own bench
based     Sri Lanka
status    hello, world_

XD TRONICS

Every board we bring up starts the same way. Flash a few lines of firmware, open a serial monitor, and wait for hello, world. If it shows up, the board is alive and everything after that is detail.

This post is our hello, world. It's a long one, because before we tell you what XD TRONICS does, we want to tell you about a factory that almost happened.

The factory that almost happened

At the end of the 1970s, Sri Lanka opened up its economy and built its first free trade zone at Katunayake, next to the airport. American chipmakers were looking for places in Asia to assemble their parts. For a few years, Sri Lanka was one of those places.

In October 1980, Motorola registered a fully owned subsidiary in the zone to build an assembly plant for 2,624 workers. Harris Corporation, another American electronics company, registered its own subsidiary and started building a plant for 1,850 more. That's nearly 4,500 electronics jobs on the way.

Then came July 1983. Motorola pulled out that year, and Harris followed in 1984. Both took the work to Malaysia.

Penang had started a decade earlier, and small. Intel's first plant outside the United States opened there in 1972 as a five-acre assembly site. By 2013, Motorola's plant in Penang employed 6,500 people. Around it grew a cluster of local subcontractors, and some of them became independent companies with operations in other countries.

That last part is what Sri Lanka really lost. Not a factory. What a factory teaches: suppliers who learn to hold a tolerance, and engineers who learn to run a line and then leave to start something of their own.

Where that leaves us

Fast forward four decades. In 2024, Sri Lanka's electrical and electronics exports earned US$422 million, according to the Board of Investment. That's about 3.5% of national exports, made by roughly 37,000 people.

The same year, Sri Lankans working abroad sent home US$6.58 billion. More than 312,000 people left for jobs overseas in 2024, the highest number the Bureau of Foreign Employment has ever recorded.

Electronics exports vs. money sent home Sri Lanka, 2024, in US$ billions

Electrical & electronics exports

0.42

Remittances from Sri Lankans working abroad

6.58

Sources: Board of Investment of Sri Lanka (exports) · Central Bank of Sri Lanka (remittances)

For every dollar Sri Lanka earns exporting electronics, Sri Lankans abroad send home more than fifteen.

Engineers are part of that flow, and their degrees travel well. State universities teach them without tuition fees. Since 2014, when the Institution of Engineers Sri Lanka joined the Washington Accord, an accredited Sri Lankan engineering degree is recognised across every other signatory country, Australia and the UK included. By the Board of Investment's count, the country produces more than 2,000 engineering graduates a year. In 2015 alone, close to 2,500 engineers migrated.

We're not going to lecture anyone who left. People go for better pay and a future they can plan around, and after 2022 that's hard to argue with. The question we care about is a different one: what's here for the engineer who stays, or the one who wants to come back? If the only answer is support work on products designed somewhere else, the numbers above won't change.

2022 sent the bill

If you built hardware in Sri Lanka in 2022, you remember. The country defaulted on its foreign debt in April. In August, short of dollars, the government suspended imports of about 300 items, electrical and electronic goods among them. Inflation peaked at 69.8% in September, and over the year the rupee lost 44.8% of its value against the dollar.

A US$10 sensor module that cost about Rs. 2,000 in January cost about Rs. 3,600 by December. Nothing about the module had changed. Restrictions on more than 300 items, electronic equipment included, weren't lifted until June 2023.

Here's the lesson we took from it.

Import finished electronics and, when the supply stops, all you can do is wait. Design your own and you can redesign around whatever you can still buy.

That's not theory. During the global chip shortage of 2021 and 2022, STM32 microcontrollers became close to impossible to buy. Engineers who owned their schematics swapped in parts that were in stock (often an RP2040 or an ESP32), ported the firmware and kept shipping. Companies that only resold finished products had nothing to swap.

We'll be honest about the limits. The silicon is imported. Nobody in Sri Lanka fabs ESP32s, us included. But everything wrapped around the chip can happen here: the schematic, the layout, the firmware, the enclosure, the assembly, the testing, and the person who picks up the phone when it breaks. That's where most of the engineering hours go.

What XD TRONICS is

So here's what we built: three businesses that usually live apart, run as one team.

The studio

We design hardware. PCB layout in KiCad, Altium or EasyEDA, from a single-layer breakout to multilayer production boards. Firmware in C, C++, MicroPython and Arduino for ESP32, STM32, RP2040, ATmega and nRF52. Enclosures and mechanical parts in Fusion 360, SolidWorks and Onshape, with tolerances set for the machine that will actually make them.

Manufacturing

Then we build it, on our own bench and through partners we've vetted. SMT assembly down to 0402, BGA and QFN, on boards from 1 to 10 layers. FDM printing up to 300 × 300 × 400 mm, plus resin for fine detail, in PLA, PETG, ABS, ASA, TPU, nylon and carbon-fibre blends. Minimum order is one board. Every assembly gets a functional test before it leaves.

The store

And we stock the shelf: dev boards, sensors, robotics parts, RC and FPV gear, 3D printing supplies, tools, components and learning kits. One rule decides what goes on it. If we wouldn't use it on our own bench, we don't sell it.

Custom IoT builds

Put the three together and you get our favourite kind of job. Bring us a problem, leave with a finished product: board, firmware, enclosure and assembly, from one team.

Why we put it all under one roof

Three businesses are harder to run than one. We did it on purpose, and the reasons are economic before they're sentimental.

Every hand-off has a price

In 1937 the economist Ronald Coase asked an awkward question: if markets are so efficient, why do companies exist at all? His answer helped win him a Nobel Prize in 1991. Every time work crosses from one business to another, someone pays to find the supplier, brief them, check what comes back and argue about what went wrong. A company exists wherever doing the work inside costs less than all of that.

Hardware is mostly hand-offs, and here's a familiar one. A freelancer designs a board and emails the outline as a PDF. Someone else models an enclosure from that PDF. Two weeks later the boards arrive with the USB-C connector sitting 0.8 mm proud of the edge, because the library footprint didn't match the part that was actually bought. Now the lid won't close. The board designer blames the CAD, the CAD designer blames the board, and the client pays for both fixes.

When one team owns the schematic, the footprint, the STEP file and the printer, that argument never starts. There's nobody else to blame.

Margins stack up

Every independent business in a chain needs its own margin to survive. That's fair. But margins compound. Pass a part through four businesses that each add 20% to what they paid, and the final price isn't 80% over cost. It's 107% over, because every markup is charged on top of the last one.

Economists call the two-link version double marginalisation, and it's one of the oldest arguments for doing more under one roof. A single quote from us covers design, firmware, enclosure and assembly. One margin, not four.

Speed is how hardware learns

Hardware gets good one revision at a time. Each spin teaches you something the simulation didn't, so the real measure of a team isn't how clever version one is. It's how many versions it gets through before the money runs out.

Most of the waiting in hardware isn't machine time. It's the gaps between people: the file that sits in an inbox for two days, the format nobody can open, the courier, the customs hold. A team that owns the whole loop closes the gaps it controls. Small prints turn around the same day. PCB design prototypes start at five working days. More revisions a month means more learning per rupee.

Ask people who've built products in Shenzhen what makes it fast, and the answer usually isn't cheap labour. It's that everything is close: parts, boards, assembly and tooling, often within a short drive. Sri Lanka can't copy Shenzhen. It can copy the closeness.

Own both ends of the smile

In 1992, Acer founder Stan Shih drew what's now called the smile curve. Line up the stages of making an electronic product from left to right and plot how much value each one captures. You get a smile. The ends sit high, and assembly, in the middle, sits low.

The smile curve Value captured is high at research and design, lowest at assembly, and high again at sales and service. XD TRONICS covers the studio end, manufacturing in the middle, and the store end. Value captured Where most countries start R&D Design Assembly Sales Service Studio Manufacturing Store
The smile curve, after Stan Shih (Acer), 1992. Illustrative, not to scale. Brackets show where XD TRONICS works.

Most countries enter electronics at the bottom of that curve, assembling someone else's design in an export zone. It's honest work, and it's where Penang started. But the money and the know-how collect at the ends: the design files on the left, the customer on the right.

XD TRONICS is built to cover the whole curve. Our studio works the left end. Manufacturing sits in the middle. The store, and the support behind it, covers the right. With all three under one roof in Sri Lanka, the value from each stage stays here, and so do the people who know how to do the work.

The bench stocks the shelf

The parts in our store are the parts we design with. Buying for both lets us buy in bigger lots, keep things in stock and actually answer questions about what we sell. Kits start life as studio projects. And the student who buys a sensor from us for a final-year project might be the founder who needs a production run five years later. We'd like to be useful to both of them.

What it costs us

This isn't the easy route. Printers, assembly gear and stock all tie up cash. We have to be good at five crafts instead of one. And when the board doesn't fit the box, there's nobody else to blame.

That last part is the point.

What we believe

  1. 01

    Build it here.

    Design, firmware and finished hardware made in Sri Lanka, by people who live here. Silicon gets imported. The thinking doesn't have to be.

  2. 02

    One team owns the whole thing.

    Schematic to shipped box. If something doesn't fit, it's our problem, not a gap between vendors.

  3. 03

    Your files are yours.

    Schematics, Gerbers, firmware source, CAD. All of it is handed over at the end of a project. A startup that can't take its own design to another factory doesn't really own its product.

  4. 04

    The shelf is our bench.

    We only sell parts we'd use ourselves, and we can tell you why each one is there.

  5. 05

    Show the work.

    Open build logs, open-source designs where clients agree, and honest post-mortems when the bench wins. Nobody in a garage in Kurunegala should have to start from zero.

  6. 06

    Price for the person counting components.

    Pro-grade work without the enterprise quote. Fixed price where we can, transparent where we can't.

  7. 07

    Answer like a person.

    A real reply, usually within a day. No ticket numbers, no chatbot.

  8. 08

    Make staying worth it.

    Every serious hardware project done here is a reason for a good engineer not to leave. We want to be one of those reasons, and to help start a few more.

What's next

This blog is where we'll show the work. Expect build logs from the bench, PCB and firmware deep-dives, print settings you can copy, and the occasional post-mortem. Designs we can open up will go on GitHub.

If you're a student with a final-year project, a startup with a prototype that needs to become a product, a teacher setting up a robotics club, or an engineer who came home and wants to build something real, we'd like to hear from you.

In 1980, two American companies picked Sri Lanka to build electronics. Within four years both were gone, and the work went with them. This time we're not waiting for someone to pick us. We're building here, one board at a time.

The XD TRONICS team


Building something? Tell us about it, or browse the shelf. A real person reads every email.

Sources

Company Manifesto Sri Lanka
More posts

Got a project in mind?

We design, build, and ship hardware in Sri Lanka. One team, zero hand-offs.

Start a conversation