Oct 12, 2018 · 1h 1m · y-combinator
A Conversation on Hard Tech with Eric Migicovsky · Y Combinator
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this Y Combinator discussion, partner and Pebble founder Eric Migicovsky shares essential strategies for hard tech startups, covering iterative hardware prototyping, early commercial validation, frugal capital management, and practical insights from Pebble's growth.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. How this is scored →
The partners as informed peer 3.0 Guest teaching 5.8 Guest disagreement 1.3 The partners pushing back 0.9
speaking balance: gold is the partners, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 6:34 Refuting the immediate post-YC success narrative
Eric directly pushes back on Adora's assumption that Pebble succeeded immediately after YC, clarifying that it took five painful years.
Hardest push from the partners ▶ 8:23 Questioning whether low-budget consumer hardware is viableAdora presses Eric on whether his improbable $40k consumer hardware path is genuinely replicable today or an artifact of the past.
Biggest teaching moment ▶ 19:47 Deconstructing the mandatory subscription service mythEric reframes Adora's question regarding subscription models into the nuanced concept of business-model company fit and hits-driven hardware.
The partners hold their own ▶ 10:37 Identifying how massive funding erodes startup scrappinessAdora astutely synthesizes the danger of overcapitalization, noting that founders lose creative scrappiness once they raise large funding rounds.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | The partners as informed peer | Guest teaching | Guest disagreement | The partners pushing back | Why |
|---|---|---|---|---|---|---|
| Pebble's Early Days and Iterative Manufacturing | 2 | 5 | 1 | 1 | Adora sets up conversational prompts about Pebble's beginnings and YC batch experience. Eric educates the audience and host on the benefits of garage-level one-by-one hand assembly to catch early hardware flaws. | |
| Overcoming Hardship, Timing Pitfalls, and Frugality | 4 | 6 | 2 | 1 | Eric gently corrects Adora's premise that Pebble took off right after YC, explaining it took five agonizing years. Adora adds a sharp observation about how raising too much capital makes founders lose scrappiness. | |
| Strategic Crowdfunding Frameworks for Hardware Startups | 3 | 6 | 1 | 1 | Adora inquires about the optimal timing and readiness needed for crowdfunding campaigns. Eric outlines two distinct strategic frameworks: early low-budget validation versus post-validation marketing scaling. | |
| Team Building and Avoiding Hardware Outsourcing | 2 | 6 | 2 | 1 | Adora asks about outsourcing specialized engineering roles. Eric strongly advises against contractors and agencies, arguing that generalist agility, trust, and alignment beat domain knowledge. | |
| Hardware Business Models and Business-Model Fit | 3 | 7 | 2 | 1 | Adora prompts Eric on the common belief that all hardware must have software subscriptions. Eric reframes this conventional dogma, explaining 'business-model company fit' and hits-driven market realities. | |
| Defining Hard Tech and Developing Minimum Viable Products | 3 | 6 | 1 | 1 | Adora asks for a concrete definition of hard tech and minimum viable products. Eric explains iteration cycle limitations and defines MVPs in heavily regulated or deep scientific domains. | |
| Balancing Technical Development with Early Sales Traction | 4 | 6 | 1 | 1 | Adora and Eric discuss Demo Day expectations for hard tech founders. Eric breaks down a case study of a university spinout shifting focus from laboratory tinkering to early customer sales. | |
| Commercializing University Research and Capital Efficiency | 3 | 6 | 1 | 1 | Adora asks about transitioning from PhD programs to commercial enterprise. Eric contrasts the capital-heavy Magic Leap approach against Oculus's lean, iterative crowdfunding trajectory. | |
| Case Study: Relativity Space and Technological Pivots | 4 | 5 | 1 | 1 | Eric shares the founding story of Relativity Space printing 3D rockets to eliminate integration testing. Adora adds context regarding the young founders' prior aerospace backgrounds. | |
| Audience Q&A: Customer Acquisition and Off-the-Shelf Hardware | 2 | 5 | 1 | 1 | An audience member asks how to find first customers. Eric advises software engineers to hack off-the-shelf hardware and shares grassroots marketing anecdotes from WhatsApp and Pebble. | |
| Audience Q&A: Location Strategy, IP Protection, and Niche Markets | 4 | 5 | 2 | 1 | Audience questions cover relocating to Shenzhen and patent moats. Adora and Eric actively collaborate to demystify IP, agreeing that product velocity and data network effects outweigh patents. | |
| Audience Q&A: Monetization Models, Hiring Tactics, and Custom Development | 4 | 6 | 1 | 1 | Eric answers questions on pricing enterprise hardware and scaling headcount rapidly through personal networks. Adora jumps in with practical advice on directly extending concrete job offers. | |
| Audience Q&A: VR Prototyping and Managing Hardware Cash Flow | 1 | 6 | 1 | 0 | Eric delivers an in-depth explanation of hardware cash conversion cycles, highlighting Xiaomi and OnePlus inventory models and the value of shortening lead times over optimizing unit margins. |
Statements from this episode (22)
Migicovsky: Pebble name was conceived two years before creating the smartwatch
“I actually thought of it while lying on a beach, and it was like two years before we actually made Pebble, and I just like wrote it down in Evernote on a list of Good names.”
Migicovsky: Seven of Pebble's first 10 watches broke apart during shipping
“We screwed up everything you could possibly screw up. In the first 10 watches we had a lot of problems, but we had a problem with RF. We couldn't get the signal out of the watch, because it's like a Faraday cage sitting on your wrist. And so we said, well, may…”
Migicovsky: Hand-assembling early hardware units limits catastrophic loss from manufacturing bugs
“It was a really good thing that we were making them one by one, or else we could have potentially made, like, hundreds with that same problem right at the beginning and had lost, you know, basically everything.”
Migicovsky: Pebble hand-manufactured 700 to 800 units before full automation
“We continued to manufacture hand by hand, one by one by hand, up until around 708 hundred units.”
Migicovsky: Pebble spent its entire post-YC $250k funding round on inventory
“We raised a bit of money about 250,000 dollars including this kind of large check that we got from, ah, one investor. And I took all that money and invested in inventory, because I thought at the time the biggest constraint on us growing was we didn't have a r…”
Migicovsky: Overhyped hardware crowdfunding campaigns usually fail to ship
“Most of those campaigns, even if they do succeed at raising money, usually fail to produce and actually ship their product.”
Migicovsky: Pebble sold around 1,500 watches before launching Kickstarter
“Most people don't know this, but when we launched our Kickstarter campaign, we had already made and sold around 1500 of our first watch.”
Migicovsky: Early hardware startups cannot pay contractors enough to care like founders
“Some of the most painful moments of my early career were, ah, not having full-time co-founders, and having to work with contractors or consultants who, honestly, like, you can't pay someone enough to care. I mean, firstly because you probably don't have that m…”
Migicovsky: Hire for trust, dependability, and general intelligence over deep domain expertise
“I actually opt for in the early days and pretty much all days trust and dependability and kind of, Smartness over very specific, deep, domain-specific knowledge in who I'm looking to hire.”
Migicovsky: Pebble might have succeeded by operating as a frugal, hits-driven business
“I don't think we ever Reached escape velocity to a hit with any of our later products, and I think if we had built a more frugal organization, if we were able to build products cheaper, and then launch them and test to see if they were successful, and if they …”
Migicovsky: A Medical Device MVP Requires Peer-Reviewed Science and Clearance Roadmap
“For medical devices, I think the MVP is successfully peer-reviewed science that supports what you're working on, that, that there's an, that there's some sort of opportunity here, and a kind of understanding of what your path to clearance is.”
Migicovsky: Oculus' iterative hardware approach beat Magic Leap's secretive, mega-funded R&D
“Magic Leap raised lots of money and spent just 11 years working on a product, versus Oculus, which raised two million dollars on Kickstarter, shipped a, you know, an early development kit that was not amazing, but they iterated and they got through, and now th…”
Migicovsky: Relativity Space's thesis was that 3D-printing eliminates integration testing costs
“Their thesis was it's not actually rocket assembly that costs a lot of money, it's the testing phase of integration, where if you have hundreds of different systems that have to work perfectly with each other, when you assemble and build the final rocket, you …”
Migicovsky: Pivot to building end products if customers won't buy your technology
“If you're in a situation where you're building something that's, like, mind-blowingly cool and revolutionary, but you're finding that customers don't want to buy it, pivot your company so you're actually doing what your customers would have done, and just use …”
Migicovsky: WhatsApp first launched on the FlyerTalk travel forums in 2009
“You know where the WhatsApp guy posted his app first in back in 2009?
[2266] Adora Cheung: No.
[2266] Eric Migicovsky: He posted it on FlyerTalk.
[2268] Eric Migicovsky: Forums, which is, like, a whole bunch of, like, geeky business people who fly a lot.
[22…”
Migicovsky: Early scooter startups began by gluing modems onto Alibaba Xiaomi scooters
“A good example of this would be kind of the scooter companies. They took Xiaomi scooters, you know, nine-bought scooters, and they just glued on cell modems. That's it. That, that's how scooter sharing started. It's like cell modems in GPS plus scooter. They d…”
Migicovsky: Hardware startups should hack existing products instead of hiring design firms
“Don't hire, like, An industrial design firm to build you some new design. Don't hire, like, an engineering firm that promises to do a custom design for you, because that'll cost hundreds of thousands of dollars, and it probably will be the wrong product, versu…”
Migicovsky: Startups rarely build successful moats using patents outside biotech
“So it's very rare that startups build a successful moat using patents. This happens much more in, like, biotech”
Migicovsky: Top investors prefer startups targeting small initial markets
“Honestly, the best investors will want to see a small market, because a small market at the beginning means you can identify who's in that market very cheaply, versus having to boil the ocean, saying, well, there's like, Two percent of all people in the world …”
Migicovsky: Pebble Recruited Seven College Friends in Days Post-Kickstarter
“Right after we kind of hit on Kickstarter, we were only two people at the company. Two full-time people and two interns, and so we had, like, nobody, and all of a sudden we needed to hire people, and we could have spent, you know, three to six months hiring a …”
Migicovsky: Xiaomi managed inventory by only selling phones on Tuesday mornings
“Xiaomi, up until two or three years ago, sold their phones on Tuesday mornings. Like, they had a, once the, you couldn't go on the website and buy Xiaomi phones. You could only buy them, like, Tuesday mornings, and so it was like a thing. Everyone would be at …”
Migicovsky: Hardware startups should trade unit margins for shorter component lead times
“Sometimes it's in your best interest to pay more for the product, pay more for the bomb, the bill of materials, pay more for the components, if The lead times are shorter because you may not actually have a margin problem. Maybe like you actually have quite a …”