How to launch a hardware startupLet me cover a few points as the Founder of a hardware start-up and as a NPI professional that has brought over 30+ hardware products to consumer release for major companies.
1) Costs: Hardware startups are only slightly more expensive then software start-ups. My consumer electronics startup is on Version 10 Rev 3 for PCBs and I am on Version D of my plastics. Including components and plastics I am out of pocket for about $2K for hardware related costs. I have also worked on high-speed telecommunications (Fiber to the Business) and a cellphone startup. In both cases the hardware required to get funded was less than $5K. Yes the costs go up when you get ready to go to production (probably another $10K-50K for tooling and approbations), but that is not like the millions of dollars that hardware startups used to cost.
2) Lean, lean, lean...: Let me say that again Lean. Lean startup practices are a MUST for hardware startups. The leading cause of hardware startups failing is not getting market/product fit prior to paying for tooling. Your design is probably going to change so don't pay for tooling and such until you are sure you have product/market fit. Build a few, get customer feedback and iterate. The margins are probably going to suck at first, but you can absorb the costs into what (should) be the distributor and/or wholesale markups.
3) Avoid China (at first): People are going to try to pressure you into producing in volume and going to the Far East to get the best price. This can be THE KISS OF DEATH! It is stupid and introduces so much more risk to your company. Going to China (or Thailand or Vietnam or Malaysia) is for high volume products that have a well-defined manufacturing and testing process. It is a cost reduction effort, not an NPI route. Never have your product built for the first time ever at a factory that you can't drive to in less than a day if there is an issue (and there will be issues).
4) Never buy more than you know you can sell in a reasonable timeframe: A lot of start-up costs are in inventory. Most people buy more than they have orders for (or at least honest forecasts) due to high MOQs (see 3), driving the price down to profitable levels, long lead times (see 3 again) and/or being overly optimistic. Inventory ties up capital that could be better used elsewhere and also reduces agility as it is very costly to scrap or rework product that has already been produced. It also costs money to ship and store and can cause accounting nightmares. You can also order low volume components through places like Mouser and Digikey. Just watch your lead times on some items as there are components that have 16 week lead times.
5) DIY: Learn to make do with what you have for prototypes and low volume production. There are plenty of low volume/cost batch PCB processing services ( I personally use OSH Park ~ Welcome), frameless cut metal stencils and a rewired toaster oven (Page on instructables.com). Yes it is a pain placing parts by hand. I can place down to 0402 and QFP by hand without to many problems. BGA are hit and miss and you will probably need a low volume assembly house, but those are easy to find.
One final note: I am about to launch my first Kickstarter next months and it represents a large capital investment. Both the photography/video and advertising budget are pretty substantial ($2K total and I might be increasing my ad budget). Don't think of it as some free magic bullet that will take care of your funding problems. Make sure you have enough cash on hand to have pretty high production values as well as a decent marketing plan that you can execute.