Hardware and hard tech startups: lessons from Pebble's founder

Executive overview

Most hardware founders over-invest in product before proving demand, then mis-read their business model once they find it. Eric Migicovsky built Pebble through hand-assembled units, a Kickstarter that shipped because of 1,500 prior sales, and a painful lesson about frugality after fundraising.

Hard tech compounds the problem: regulation, slow iteration, and large capital requirements all mask whether customers actually want what you're building.

Start selling before the product is ready — customers will build it with you.

The Pebble origin story

• Applied to YC in 2011 with ~300 hand-assembled watches; watch worked on Blackberry
• Assembled units one-by-one until ~700–800 units, catching defects early (exploding backs, RF shielding failures)
• First big mistake: raised $250k after YC, spent it all on a 3,000-unit contract manufacturing run
• That run took 6–8 months, cost more than planned, and landed when customers had switched from Blackberry to iPhone
• Second big mistake: raised a larger round, hired rapidly, lost frugality — happened again three years later
• Pebble's Kickstarter launched after 1,500 units were already sold; it was marketing, not discovery

Two distinct crowdfunding moments

• Early stage: raise $25k–$50k from friends and niche communities to fund a first prototype
• Do not run a polished marketing campaign at this stage — most big early campaigns fail to ship
• Use Kickstarter, Indiegogo, Crowd Supply, Tindy, Discord, Reddit, or niche forums
• Later stage: use crowdfunding as a sales and marketing channel once you have a proven product
• Mixing up these two modes is a common failure pattern

Frugality as strategy

• Creative constraint forces better decisions; money removes the pressure to figure things out
• Hardware as a hits-driven business: launch cheap, test if it's a hit, scale only if it is
• Pebble never identified its second hit because the org grew too expensive to run lean experiments
• "Business model company fit" matters as much as product-market fit — the wrong model can kill a good product

What hard tech MVPs look like

• Hard tech = slow iteration cycles due to hardware, fundamental technology development, or regulation
• Medical devices: peer-reviewed science supporting the opportunity + a clear path to regulatory clearance
• Rockets / propulsion: subscale working prototype + early customer contracts or LOIs
• In all cases: customers paying or committed money is a stronger signal than technical progress alone
• Do not delay sales until the product is "ready" — engineers treat sales as a muscle they haven't trained

Selling before building

• Companies that spend the first two months on prototypes and the last month on sales consistently underperform
• Flip the order: talk to customers first, build for what they confirm
• A YC company that had been iterating for two years went from zero sales to $4–5M in contracts within three months of being forced to sell
• If a customer isn't excited when you explain their problem and your solution, they probably don't have the problem

Advice for scientists and engineers leaving academia

• University ecosystems default to grants and long-horizon research; commercial urgency is not in the culture
• Start small: go to industry (not academic) conferences, talk to potential customers
• If customers respond with "where has this been all my life," that pull is the momentum worth following
• Most hard tech companies do not need a binary $100M round to start — find the smallest fundable milestone first

Hardware for software engineers

• Buy an off-the-shelf product that is 50–75% of what you need; hack it rather than designing from scratch
• Scooter companies example: Xiaomi scooters + cellular modems + GPS — no custom hardware
• Raspberry Pi + Pi camera + 3D-printed case can be a viable commercial product
• Custom industrial design before product-market fit is a trap: it costs hundreds of thousands and produces the wrong product
• Move closer to customers before product-market fit; only move near factories if iteration speed requires it

Hiring in hardware startups

• Hire for trust and reliability over domain-specific expertise
• After Kickstarter blew up, Pebble hired seven people in 3–4 days by calling university friends — faster and more cohesive than a formal hiring process
• Contractors and consultants cannot be paid enough to care the same way a founder-level team member does
• Slowly scan your LinkedIn and Facebook connections; overlooked people from school or past jobs are often the right first hires

Moats and IP

• Network effects and data lock-in are the two strongest moats for hardware companies
• Patents are expensive to obtain and even more expensive to enforce — rarely the right tool for a startup
• Building a great product is the primary moat for 95% of companies
• Small initial markets are a feature: easier to identify customers, faster feedback on whether the product works

Inventory and financing

• Inventory is the core cash drain for hardware companies: cash goes in, revenue comes out months later
• Preorders break the cycle — use customer money to fund production rather than investor or bank money
• Reduce lead times even at the cost of higher unit cost: less capital tied up in stock, lower risk per batch
• Banks will not fund R&D; VCs and grants do
• Xiaomi and OnePlus used weekly "drops" to match production exactly to demand — a structural inventory solution