I Sold 2,500 MIDI Recorders Alone. Hardware Wasn't the Hard Part.

Ask a software developer why they’ve never built a physical product and you’ll get the same answer every time: hardware is hard. Inventory rots. Tooling costs are brutal. Returns pile up. You end up with a garage full of boxes and a drained bank account. But over the past few years, a steady stream of indie builders has been quietly demolishing that story — solo operators shipping thousands of units of deeply niche products like standalone MIDI recorders. Their retrospectives converge on something uncomfortable: the part everyone fears isn’t the part that kills you.

What “Hardware Is Not So Hard” Actually Means

The phrase has become something of a meme in indie hardware circles, and it’s routinely misread. It does not mean hardware is easy. It means the specific thing you’re afraid of isn’t the thing that’s difficult.

Run down the typical developer’s list of hardware anxieties. I can’t design a circuit. I don’t know how to negotiate with a factory in Shenzhen. Certification will eat me alive. Now read the postmortems from people who actually shipped. Almost every one of those items lands in the “learned it faster than expected” column.

The reason is structural. The barrier to entry collapsed over the last decade. KiCad made professional PCB design free. JLCPCB and PCBWay will take a five-board order without blinking. Distributors like Digi-Key and Mouser sell single components to individuals. Work that once required a 10,000-unit minimum order quantity now starts with a $15 prototype run. The tooling caught up. The fear didn’t.

The Bottleneck Is Everything Around Manufacturing

Founders who’ve crossed the thousand-unit mark point at the same choke points, and none of them are on the factory floor.

Fulfillment comes first. Building 100 units is a solved problem. Shipping 100 units to 30 countries is not. Every destination has its own customs regime. Packages get crushed. Customers mistype their addresses. This isn’t engineering — it’s operations, and for a solo founder operations scale multiplicatively, not linearly. Every unit sold is a potential support ticket for the next five years.

Support is the second wall. In software, you fix the bug and push. In hardware, the broken thing is sitting on someone’s desk in Melbourne. A firmware update is the lucky case. A bad capacitor means you’re shipping a replacement out of your own pocket or issuing a refund — either way, that sale is now negative revenue.

The third one is the killer.

Inventory Is a Cash Flow Black Hole

This is where software people get blindsided. Software has near-zero marginal cost. One user or ten thousand, the difference is a server bill.

Hardware inverts that completely. You spend the money before you make any. You buy components, pay for assembly, warehouse the boxes — and only then do you start selling. If a chip shortage stretches lead times to 40 weeks mid-run, your plan is already dead.

Do the arithmetic. A product with $50 in bill-of-materials cost, ordered at 500 units, means $25,000 leaves your account up front. If it takes six months to sell through, that’s six months of your life financed without that capital. This is the entire reason Kickstarter and preorders became load-bearing infrastructure for indie hardware. They don’t just validate demand — they reverse the order of the cash flow, and that reversal is worth more than the marketing.

Niche Isn’t a Limitation. It’s the Moat.

Here’s why a product like a standalone MIDI recorder works for one person: the market is small.

That sounds like a weakness. It’s the whole strategy. A market that moves a few thousand units a year is invisible to any company with a headcount — the salaries alone wouldn’t clear. Nobody at a major audio brand is pitching a product line that tops out at 3,000 units. But for a solo founder, 2,500 units at $100 margin is $250,000. That’s not a side project. That’s a living.

Niche customers also come pre-qualified. Musicians can articulate exactly what they want — capture MIDI performances without hauling a laptop to every session — because they’ve been annoyed by its absence for years. You reach them through forums, YouTube reviewers, and Reddit threads, not paid acquisition. Try that with a mass-market gadget and customer acquisition cost devours your margin before the first unit ships.

So Should You Do It?

The message for developers sitting on a hardware idea is unambiguous. The technical skills are learnable. What you actually need is capital and patience.

If you’ve been deferring on the grounds that you can’t design a circuit board, that excuse expired sometime around 2018. The real questions are harsher: do you have enough runway to survive until unit 500 sells, and are you genuinely willing to answer support email every single day for years?

Above all, have you found something a small group of people wants badly? That’s the only part nobody can teach you. So — in whatever field you work in, is there a thing you keep wishing existed? That itch is probably the whole business.