A Conversation on Hard Tech

Migicovsky describes starting Pebble in 2008 and hand-building its first watches, including assembling components in a garage. Early batches exposed problems such as radio signals blocked by the metal case and electronics pushing through the back. Making watches one by one kept mistakes small while the team iterated.

Pebble took about five years to reach product-market fit, and a large manufacturing run went badly: production took six to eight months, costs rose, and customers moved on from BlackBerry. Migicovsky says crowdfunding can help either test early demand with a modest campaign or serve as a sales channel after a product has been tested with customers. He recommends completing an initial build-and-ship cycle before attempting a major campaign.

Migicovsky favors adaptable technical teams and hiring for trust and reliability over narrow expertise. For hard tech, he says the first milestone depends on the field: peer-reviewed science and a path to clearance for medical devices, or a subscale demonstration and customer interest for rockets and satellites. He urges founders to talk to customers early; one YC team won four or five million dollars in sales after shifting its focus toward selling.

Hard tech takes longer to iterate because of hardware development, fundamental research, and regulatory constraints. Migicovsky suggests starting with off-the-shelf products, such as a Raspberry Pi or a modified scooter, to test whether a solution works before commissioning custom hardware. He also advises scientists to leave the lab and speak with potential customers at industry conferences.

Migicovsky says product-market fit comes before choosing a business model, and warns that subscriptions should match what customers actually want. He describes consumer hardware as potentially hits-driven, where a company must stay frugal while testing new products. To reduce inventory risk, he recommends preorders or shorter production lead times, even if faster components cost more.