YC for Hardware
blog.ycombinator.com
blog.ycombinator.com
Lowering the barrier to hardware startups is an awesome thing, of course. Hardware has a greater potential to directly impact lives at this point than software. There's tremendous software power locked up in the small, awkward-to-use computing devices we call phones, tablets, laptops, desktops, and servers.
Hardware and Software startups are sufficiently different in their needs and life cycles that it would be specious to use the success of a particular program for the one as indicative of its suitability for the other. Similarly, the sorts of b2c startups that YC typically funds are sufficiently different from b2g startups like Palantir that I don't think you could claim Palantir would have benefitted from YC based solely on the experiences of b2c startups.
That is true. A better way to think of it, might be as an experiment. Do you remember when YC was started funding startups? The tech group-thought at the time: a toy, a joke, nothing serious, never to gain traction.
>some
>arguably
>far enough along
>yet
I don't mean to insult you or criticize your post in any way, partly because I agree with you, but that is a lot of qualifiers for one short sentence. Diplomatic tones require them though, I suppose.
Even if they were building their own sheet metal lockers, I still don't think they were a "hardware" company. They had a great service and that's what they were selling.
Bufferbox offered a service and didn't sell you a product therefore I don't think they were "hardware" company. The hardware was means to an end.
I don't think it lessens what they did. It was a good service.
Walmart has a large IT infrastructure, with sys admins and in-house developers. They also manage a lot of buildings, and warehouses. I wouldn't consider them an IT company, a property management company, or a logistics company.
I agree with your points regarding BufferBox, but I wouldn't say they are a hardware company producing a hardware product that you can sell to a customer. They were in an entirely different space, with very different set of challenges.
The YC post concerned companies that sell hardware to end users. Further, the initial post in this chain was someone asking if YC has had other hardware companies beyond Pebble. I don't think BufferBox is in that category.
Your manufacturing pipeline, volumes, service issues are all very different. So as great as BufferBox was I don't think they were a hardware company and certainly not for the purposes of this thread. I am not discounting their abilities or what they made but based on my experience they are apples and oranges.
Most of the problems kickstarter companies run into are unique supply chain, and manufacturing issues that BufferBox wouldn't have had to deal with.
Additionally, a big part of what BufferBox offered would have been in their app, and website. A locker that you can remotely lock and unlock is not difficult from an electronics point of view and had been around before.
Didn't YC move?
sounds very scalable
So, let's say now you need 10,000 of them. For 1/4 the price. Delivered next month.
That's where Shenzhen starts looking really nice.
Especially if you've already got reliable contacts there, like Bolt does with Dragon Innovation.
That said, I was vastly more impressed with the infra in SZ than Bangalore. SZ is a modern metropolis where Bangalore, well... It has pockets of shiny corporate havens I the suburbs. (FWIW, I personally prefer the grit of downtown BGL to SZ anyway)
Your other comments are factually true, but i don't think they're totally fair. To play devil's advocate: China acknowledge they've got a pollution problem and introduced several methods to start addressing it [1][2][3]. USA also ignores human rights[4][5][6][7], supports dictators and totalitarian governments[8], fights proxy wars and arms terrorists[9][10]. "These are good reasons to not do business with" the USA.
[citations http://pastebin.com/d5UGRB1Z]
What about vias, or soldermasks?!
That machine is great for CNCing plastics and metals however.
You don't get soldermasks and vias are normally riveted (good enough). For most people, two layers is enough and if you can design the board well you shouldn't need much routing on the second layer.
Simple power/control boards (weird voltage, multiple outputs) that aren't available off the shelf from RS or Farnell. None of those boards need to be more than two layers. I've designed lens focus control boards that only required a single layer.
Yes of course if you're doing RF or fast serial you'd be daft to do it on a milled board, but they are convenient sometimes.
You can get two layer boards for $25-$40 depending on size in two three days, in US from many sources such as Pentalogix.
If anyone can recommend a good, cheap fab house in the UK I'd be interested to know about it!
Perfboard is fine up to a point, but I find it tends to get messy, even for simple boards, and it's not terribly optimal with regard to space. You're out of luck (without bodging) if any component is surface mount or isn't 2.54mm pitch.
Milling is a nice stopgap when you need a single copy of a board that's a bit too complicated for perfboard (i.e. different pitch/smt parts) but isn't worth spending 10x the cost for a professionally made board. With milled boards you can have custom shapes, cutouts and so on. If you're trying to design a board to fit inside some housing with a funny shape and standoff points, milling the board to fit is a nice and cheap.
Not even close.
I can't imagine mounting even a simple ATTiny, let alone a quad-flat anything or a bunch of 0603s without a solder mask.
I'm sure that some use cases can get by with all DIP, through-hole, and similar-sized parts, but it's hard to design a product that can later scale up cost-effectively without being able to use surface-mount parts. There's a reason the mass production all switched over to SMD->cost.
They have a similar focus on hardware, great numbers re the percent of each class that gets funded, and have had some big successes like the Skully AR-1 intelligent motorcycle helmet: http://www.skully.com/
WW recently toured the AQS facility in Fremont where a lot of well known products are being made like Makerbot and the Lift Labs tremor defeating spoon, and class members get frequent introductions to everyone from investors to Perkins offering billing for law services only on funding. It has been a pretty kickass program so far.
I still labour under the assumption that NREs make doing anything like that _extremely_ expensive.
The standard thing I have heard from VC's that don't know anything about semiconductors (Or have not been involved in it in this century): "It will cost $100 million and take 3 years to have your first silicon. You have a 50% chance of even making it that far, and even then you only have a 10% chance of success (e.g. making any sales)"
So basically, the VC assumption is that it is a minimum $100 million and 3 years to see ANYTHING come out of it... and even if you get to silicon, there is only a 10% chance you'll be able to sell it. This is ridiculously pessimistic if you actually have a real design innovation.
Some of the new developments: New tools: High Level Synthesis an "Hardware construction languages" (such as Chisel... http://chisel.eecs.berkeley.edu) radically bring RTL development time down. My team (consisting of just two of us) were able to complete the RTL of our new processor core in under 4 months (working on the weekends)... and that includes a lot of redesign and working on things other than the RTL itself. Chisel also brings the testing/verification time down significantly, as it can auto-generate Verilog optimized for either VLSI or FPGAs, and a high speed cycle accurate C++ simulator of the chip that is ~10x faster than the industry standard Verilog C Simulator. Even though Chisel really is this amazing, virtually no one outside of UC Berkeley knows about it.
2. FPGAs have gotten a lot better... I can say the synthesis tools still suck, but in terms of speed and testing ablility, you can test your chip on very cheap (Ours is $300) FPGAs these days.
3.Shuttle run/Multi project wafer services... a "block" (6 mm^2) on TSMC 28nm process is $90,000... and you get 100 chips back. A 16 core version of our chip is just two of these blocks.
4. Older process nodes are getting cheaper. Most fabs in the US are not going below 65nm or 45nm, and need to keep full utilization of their fabs. There are a lot of deals that can be had on these smaller fabs which are perfect for prototyping... A similar block size can be as cheap as $3 to 5k.
5. Moore's law dying is one of the best things that can happen to the industry. Instead of the ridiculous investment getting to the next process node, the fabs can invest in making the current process better and cheaper. This will bring down the cost of very good, proven processes (like 28nm, and potentially 14/16nm in ~3 to 5 years), making fabrication even cheaper and more widely available.
I'm happy to answer any questions!
EDIT/TLDR: Based on my own startup's current plan and estimates, it is possible to prototype (get 100 chips back) an entirely new processor architecture for under $2 million, and it is possible to go into mass production with that design for $10 to $15 million.
If you have a link to a source that explains how this works that would be awesome.
Edit: I misinterpreted the definition of fabless. This comment makes no sense.
I imagined something like 3D printing/laser cutting for silicon chips.
That's sort of what a fab already is =)
Wouldn't it be valuable to give interested folks who aren't yet in YC a place to do their prototyping, get discounts, and help, and then use it as lead gen for the next batch?
Think of all the companies that could exist but don't because they don't have access to these resources.
If you are hand assembling your prototypes then you only have PCB fabrication costs, which for a 2 to 4 layer board should be about $500 max for your protos (1 week turn). That should be doable for most start-ups, even if they need 3-4 spins.
I was thinking of more complex board, with non-standard finish and soldermask.
Even Pentalogix in Oregon offers 2-4 layer boards for about $25-$40 in two or three day turns.
Side note for companies like Indiegogo, success at attracting device makers to use their platform is key to their growth. At CES alone they have something like 14 booths.
(These were all on the Tilt/Open platform: https://open.tilt.com/)
As the first pre-order platform, Celery has powered successes like Pebble and Boosted Boards and helped them get to the next stage of their business.
The transformation over the past few years has been unreal and today is such an exciting time to be a HW startup. We personally can't wait to see what kinds of HW companies emerge next.
How does this make sense for YC to do?
Simply getting a functional prototype without easy access to a 3d printer, $$$ in compiler licenses for BT stacks, freakin custom batteries, etc etc.
Our thing: https://www.fitguard.me/
I'd recommend getting in touch with motorcycle helmet manufacturers (particularly MotoGP and WSBK sponsors) because concussions are a huge issue in motorcycle racing. As your site states, concussion severity is very abstractly diagnosed without data, and your product can make a big difference. The doctor has a tough time deciding whether the athlete is fit to race if he's had a crash and a concussion, and it ultimately comes down to a qualitative decision made by the rider and doctor. Data can definitely help save lives here.
If I have time for the track this season, I'll definitely sign up as a tester. I've never had a concussion, but I sure as hell would want to know how fast my head decelerates after a crash before getting back on the bike.
Good luck, and cheers on a great idea.
Data is my thing - there's a number of these types of products out there (Shockbox et al), the issue is that they all keep the algorithm quite close to their chest (fairly, I suppose!)
My role (in addition to helping out with the hardware) is the ecosystem around it. It's BLE capable, so while we can't make a diagnosis (all kinds of legal issues there), the hope is to aggregate impact data, improve the detection algorithm and be able to feed it back to the device as it improves - We're currently using a TI MCU that allows us to OTA update from the app.
Not sure on IG vs KS, not my department :D
I'm interested in the personal media/storage/app server space!
If you are hardware startup founder material and are interested, please get in touch.
*i think you meant staff
That would be really disruptive.