As you have probably seen, software startup luminary and Y Combinator co-founder Paul Graham posted an interesting article on his highly respected blog yesterday: YC is now encouraging hardware companies to apply to their 90-day program. As much as I respect Paul, be very careful of attending YC with a hardware product. Here’s why:

Not Enough Time/Cash/Staff…

Building a hardware product takes a significant amount of additional resources (time, capital, person-hours, expertise, etc) than a typical software or web service product. Y Combinator and pretty much every other seed acceleration program are built around the constraints of the sw/ws development cycle. They provide two primary pieces to satisfy this part of the life-cycle: mentorship and investor relations. These resources can be applied to companies extremely quickly and across their entire portfolio in sufficient quantities (although some disagree with the latter). When it comes to hardware, it’s simply not enough.

Limited Expertise

There are a few mentors in the YC network that know about hardware, but it’s a tiny (really tiny) minority of their community. The amount of exposure a portfolio company gets to this kind of expertise is extremely minimal, which is exactly the opposite of what hardware startups need (more than sw/ws). Asking one of the YC mentors about what the draft angle on your injection mold tool may not yield a prompt (or correct) response.

No Space

Founders at YC usually camp out in coffee shops and rented apartments around the Bay Area. Completely sufficient if you spend all day in front of TextMate and the command line;  not nearly enough when you’re building prototypes of physical goods. Hardware startups need equipment and space, even if they’re sending prototype parts to be made elsewhere. TechShop and others can help satisfy this need, but their equipment isn’t suited or within tolerance for most later-stage hardware prototypes.

Investor Expectations

Demo Day for YC is an adrenaline-fueled day with rapid-fire pitches to the who’s-who of the investment community. When investors come to the event, they expect widespread traction, minimal burn, continuous builds, lots of metrics, and other web startup buzz-words. Sorry folks, hardware has a few differences. Mileage may vary.

Community Focus

As much as I would love YC and other programs to build communities around hardware folks, it’s not going to happen anytime soon. Seed accelerators will remain software and web-service centric. This means hardware companies take a backseat and programs/events/meetings/expertise and pretty much everything else is catered around software.

So if you’re a hardware startup looking for some early stage help, I strongly suggest avoiding Y Combinator and the other seed accelerators and going with like-minded experts. In case you don’t them: Lemnos Labs in the Bay Area, HAXLR8R in China and of course us at Bolt in Boston.