There is always room for a product start up. Force10, Fire-eye, etc., prove that. And products are monumentally cool. My brother-in-law is the product start

There is always room for a product start up. Force10, Fire-eye, etc., prove that. And products are monumentally cool. My brother-in-law is the product start up guy, and we always marvel at the low barrier to entry I have (as a services startup guy). The issue is that a hardware start up requires an order of magnitude or more more resources to get off the ground than a software startup. Consider the process for developing the product. For software, that's a development issue. For hardware, it's (usually) also a development issue, but it's also a logistics issue. Where do you get the components? Who is going to assemble it? How are you going to make sure they don't suck? How are you going to get over the regulatory compliance issues in all the markets you want to sell to (like FCC certs, ROHS in the EU)? How are you going to package and ship. Do you use a channel model to support sales/deployment/support or do you build that yourself? Consider the support issues. For software, it's a test/patch/deploy issue. For hardware, it's probably that, plus how do you stock spares? How do you deploy a tech to service a customer? How do you handle a bad production run? Where is the feedback loop between reliability in the field vs the test environment? If you use one, how do you support the channel? Product just has more of everything.