As I understood it, from having worked at a (successful) startup that sold hardware, the big problem with hardware isn't that it's hard to make. Rather, the two big problems are margin and, worse for a startup, inventory costs. Airbnb can add 1,000 new customers with no infrastructure changes, but for a hardware startup to take on 1,000 new customers, someone will have to finance the inventory, and someone will need to predict the amount of widgets to stock in that inventory, and that gets very expensive quickly.
I don't know how much the inventory issue is mitigated by the fact that YC companies with working offerings seem to be immediately able to conjure up 500k-1MM in funding.
Also, YC's major successes haven't been hardware companies, have they? The last essay I read before this one suggested --- in agreement with the conventional wisdom of VC's --- that a company needs to be Dropbox-successful to move the needle for YC. Not that YC isn't, I'm sure, thrilled to have hardware product companies with traction in their portfolio.
Let me just add a banal point: YC's business strategy is, obviously, "throw everything we can at the wall and see what sticks". If you're considering your first company, that's probably not your best strategy. Even putting aside the big-ticket problems like inventory and margin, there are a lot of other things that suck about hardware: lead times, managing supply chain, QA and managing defects, field recalls, shipping. These problems are so big that major hardware companies have people who don't just have one of those tasks as full-time jobs, but are also famous for being able to deal with them.
I'm talking about a new trend. The hardware companies we've funded haven't had time to grow into Dropboxes. But as I wrote, hardware companies are overrepresented among the most promising startups from the summer batch, as far as we can tell this early.
True, hardware involves lots of schleps. But that means good ideas of this type are often lying around in plain sight, while all the other would-be founders are fixated on making the latest mobile social commerce app. (http://paulgraham.com/schlep.html)
Maybe it's particularly good to be a hardware company when the solution to the problem you're tackling requires new hardware. For instance, if 50 years from now we're all going to be getting around on motorized skateboards. But if you watch Kickstarter, there are hardware proposals that are "hardware for its own sake"; those offerings might be structurally disadvantaged compared to iOS or Android software, and you might not want to start that kind of company as your first startup.
Schlep was originally a Yiddish word but has passed
into general use in the US. It means a tedious,
unpleasant task.
Interesting linguistic tangent. I'm curious where you encountered schlep as a noun. My experience and various dictionaries define it as a verb ("to schlep"). E.g. "Oy vey. Do we really have to schlep over at this hour?"
You clearly haven't hung out with my grandmother.
Schlep can definitely be used as a noun, and commonly is by native Yiddish speakers. (As in "Oy! What a shlep!" or "Picking you up during rush hour would be a shlep, how about you just take the bus?")
There's a natural tendency in English to use as nouns words that were originally verbs. Examples: I'm going for a walk. I'm having a talk with someone.
Dictionary definition or not, a google search for the phrase "a schlep" returns >60,000 results.
If the (recently added) part-of-speech tagging on Google Books is accurate, the chronology is the other way around: it used to be equally used as a noun and verb, but sometime in the 1990s the verb form became dominant: http://books.google.com/ngrams/graph?content=schlep_VERB%2Cs...
As applied to English, it refers to the use of a verb (in its -ing form) as a noun (for example, the verb "learning" in the sentence "Learning is an easy process for some").
I've sensed a bit of a trend (e.g., with YC startups like BoostedBoards and Double, and popular KickStarter projects like Ouya), but I'm having a hard time wrapping my head around the capital requirements and ultimate profitability of these startups as they attempt to scale up into larger businesses. Maybe I just have to shake off naive preconceptions about hardware businesses being hard and difficult-to-pivot.
pg: if possible, I'd love to read your thoughts on how the economics of these businesses might evolve as they attempt to grow, especially as compared to software-only businesses.
The path-dependency (or difficulty of pivot) is an interesting angle. Certainly worth considering. Most product needs iteration if not outright pivoting. Perhaps the mitigating factor is that if the product is tangible, the need is also. SW is so powerful, as a concept, that i wonder if SW some startups may suffer from "too many degrees of freedom". Thus the ~inevitvble pivot -- combined with the ability to re-purpose a core of valuable skill-- is almost a more inherent characteristic of SW startup vs HW one. When considered in context of problem solving and product/market fit, &tc.
The issue, as I understand it, is not so much path-dependency per se, but the long lead times involved in the manufacture of new hardware at scale. Software has path-dependency too, but changing source code to modify a new software product is faster than changing physical machinery, processes, materials, suppliers, and logistics to modify a new hardware product.
This is a good comment, and sheds more light on the subject. The long lead time is a question of agility. This is subtly but importantly different that path-dependency (considered more of a lock-in to an existing technology). The link is that the ways of making the company more agile (investing in specialized manufacturing/prototyping and/or inventory) tend to create a more path-dependent trajectory (or opportunity set). So, perhaps this is better said that HW has 'interial' dependency on initial trajectory if not true path dependency per-se. Even provided that the startup is low-capital/outsourcing model. I hope this does some justice to incorporating your point.
Nice comment. I think there is indeed an inherent "inertial" in hardware business.
A software/website service can have different usage. Users may prefer to use a smaller part of the service, or using the main service for purposes not originally conceived by the creators. In this way there is a chance of pivoting.
However a piece of hardware has a very focused usage. All parts are integrated to provide a single function for the user.
Also, unlike software, there is no straightforward way to collect real-time data on how a user uses their hardware.
On some level, Dropbox is a hardware company [1]. I think that once you have installed software at a clients location, you are in a different realm than simply being a cloud/software/service provider. The customer has full ownership of the device. I think that the move to hardware is tri-fold (a) hardware is easier now than ever (b) hardware is a good strategy and a natural barrier to entry for a startup (in general, less people grok hardware and systems software than they do web programming[2]) (c) a small/mini subconscious push-back against the Cloud[3] where more and more customers are being turned off by not owning something at the end of the day[4]
[1] The hardware is the customers PC.
[2] iOS device programming, w/ human interaction (gyroscope) has also been a natural training ground.
[3] Of course, many Hardware companies these days Square, Dropbox, etc, will be paired with a cloud service.
[4] Acqui-hires..."We are shutting Down. Please Download your stuff by next week. Thanks for being a User and getting me interest from a big fish, see you on my next startup"
8% of such a tiny sample isn't anywhere near close to being able to extract trends. In such a volatile arena as investor funding even if the number of hardware companies doubled from 4% last year to 8% this, it's still noise.
We (as a tech community) need to careful thinking so small. YC is tiny. If the article was the result of talking to other investors and a similar trend was identified across the board it might be newsworthy, but to anyone with any statistical background this isn't a trend, it's noise. That doesn't mean there isn't a trend, it just means this isn't in and of itself, evidence of one.
Perhaps hardware companies don't have the "dropbox" effect of such quick growth. Do you think though that hardware companies may be a less risky proposition in some ways?
For example they are less likely to have the sort of network effects problem that some social media based application would.
Of all the companies they are funding, I could totally see Lockitron going viral. Once you have the app on your phone, you have the first key in your keychain. Once the price drops to like $50 a unit, you are in the realm of being super competitive with your standard physical lock, enough so that people would start looking at each key on their keychain as one that can be replaced.
The company I founded has custom hardware and a custom OS (not, e.g. Linux or Windows -- closer to an embedded RT OS), and our "for sale" product is SaaS.
We actually have four different hardware systems, ranging in size from a 1U box that goes on site, to a full rack that's basically a turn-key cluster with no external configuration (just plug in ethernet and power, and assign static IP addresses).
All of our hardware is based on Intel motherboards and chips, especially the E3/E5 series. This eliminates the long prototype cycles, since we use off-the-shelf hardware. The only real prototype cycle for us is choosing components and the case (!), which we did with 3D CAD software and a machine-shop-as-a-service we found on the Internet.
We went this route for reliability and performance (especially, latency), and then getting those two at a reasonable cost. By only paying for the hardware we need, and configuring it to exactly what our product/service needs, we're able to meet our product and business needs while still maintaining healthy margins at a cost our customers can afford. I wouldn't change a thing.
We also face all of the problems you mentioned (how to deal with inventory -- our hardware has a comprehensive 3 week burn in time), how to deal with repairs/replacements in the field, etc, and how to get the capital needed to sustain startup-scale growth. They're real problems.
Like most hardware-based companies, we're able to do this only with outside funding, but at least in our case (the enterprise space), we don't need any up front funding to stay cash-flow positive with slow growth, and we developed a rather innovative way to get the funding that we do need to fund fast frowth (hint: not through VCs). This took a lot of effort at the business-design level, something I suspect most tech startups don't take the time to do. It's pretty much mandatory if you're going the custom hardware route and want to grow fast.
One major positive of going the custom hardware route is the elimination of competitors. No YC-backed company going the standard AWS route can possibly compete with us -- not on cost, not on performance, not on latency, and not on reliability -- not even close. Controlling the whole stack, from hardware to OS to software is why we can do what we do at the price we do it, and grow at the rate we expect to grow.
I'm not sure we could make our approach work in the consumer hardware space today, but in the enterprise market, it's ideal for us.
I do remember the lament of Fohr on investment (re: social web startups vs hardware / (non vaporware)), and though I can't find the relevant thread, I am glad Fohr is still kicking around in the last two years since I last read about your troubles :)
Well, the particular system I described is for an operations management system we built for our film studio (Fohr).
We're in the process of packaging that up as a standalone product/service and building a business around it. As one of the commenters noted, we failed to get funding for the studio itself last year, and this is Plan B.
We've got some crazy technology in it, and I'm hoping to get some papers written up describing it all. AFAIK, it's the only Google Spanner-like design outside of Google right now.
Inventory costs are a problem, but a bigger problem is production costs.
To create a PCB you need artwork. This includes the electronic files; the drill plots, the photoplots, the etch resist masks and the silkscreens. You then need the production tools for the PCB. You'll need a stencil for the screen-printing of the solder; a pick and place machine program (and you hope your PCB software can provide something that the P&P machine can read); if there's any through hole component you might want a cropping plate made.
A team goes through the prototyping process, using hand built small batch stuff. But then they need to convert this into production ready stuff. So the above costs happen when you hope -but are not sure- that you're ready to build a full scale batch. You get ten PCBs made for test, and discover a missing trace. That means you need either another set of stuff or to modify every board as it's produced.
Without very careful internal quality procedures and excellent engineering standards it'd be very easy for a company to burn through $10k before producing any saleable product.
It is difficult to follow lean startup principles with hardware products. The feedback loop to design, build and test hardware is long and expensive. Then once you've got a workable design you'll need a lot of capital to do a production run.
Before we can have a true renaissance of hardware startups there needs to a be a platform for cheap rapid prototyping with push button production scaling.
In effect, contract manufacturers can enable "push button production scaling." They'll source parts, do engineering, pack, and ship--basically whatever you pay them to do. A high tech copy machine. The interface could certainly use some work though...
There were plenty of web app startups long before Heroku and EC2 were available, these just lowered the bar even further.
Things like arduino can lower the friction between hardware and software design quite substantially for hackers.
I wouldn't necessarily expect startups to be rushing out to compete with the dishwasher any time soon but the point is that it now becomes much simpler to think about "hooks" in your software into the physical world.
The way you explain it means two things. First the investment needed to built the necessary production capacity, but that is not production cost (production cost are a result of this via hourly rates and costs per piece). The ability to cale from a prototype stage to serial production is crucial here, and yes damn expensive and risky. One wrong step and the company doesn't even have to care about inventories.
Inventory kicks in once the company has scaled successfuly. And then iventories are more curial to success than production costs in terms of company survival. Market entry is won with product costs low enough to make a profit. Once in the market you live and die by cash expanses, inventories are huge part of this. Espacially since hardware becomes outdated pretty fast, high levels of exess stock can basically mean death for a young company.
But all that is one of the schleps coming with hardware, and bulding such an operation from scratch is a really interesting thing to do.
I suspect this would be less of a worry if you focussed on low volume high value products. If you are selling solutions that are $10,000+ then you can afford a less efficient manufacturing process. I know a few people in my local area who are doing just that in some interesting niches.
To make a PCB you design artwork, usually from the circuit diagram. This artwork is transferred onto clear sheets. You overlay the clear sheet onto a board - a sheet of fibreglass, a sheet of copper, and a layer of UV sensitive etch-resist. You expose it to UV, you fix it, you wash off the etch-resist that isn't fixed (thus exposing copper) and then etch the exposed copper off, leaving the traces.
This process involves a couple of one-off costs (design of the PCB; transfer of that design to photo artwork). It also involves some setup costs that you get every time you need more product - setting jigs to the right size to hold the stuff. The artwork costs a couple hundred dollars, and that does not include any of the design and engineering costs, that's just the price if you hand a bunch of electronic data to a PCB house.
A stencil for screen-printing solder? That's steel, laser cut, can only be used for your PCB, and if you change the design of the PCB you might need a new stencil. Each stencil costs a couple hundred dollars.
A cropping plate is steel, or sometimes fibre glass, plate that has holes drilled where you have through-hole components fitted. An assembly-operative would have the cropping plate held in a jig, with PCBs on it, and they'd stuff the PCB. When they've fitted all the through hole components the plate is put on a machine which slides a blade under the plate and crops the leads. The operative then takes the PCBs off the plate, and sends them off for (probably) wave soldering. Each cropping plate is individual, and cannot be used for any other product. You'll need as many cropping plates as you want people to build PCBs. Each plate is going to be a couple of hundred dollars (steel) maybe a hundred dollar (fibreglass).
So far you haven't got any product that you can sell to the end customer! You've just got setup costs that you hopefully amortise over the number of product you produce.
You're right that an electronic assembly house would tend to outsource all this stuff to other people rather than doing everything in house.
Tooling waived when re-ordered emphasises what I've said - this is a one off up front cost that you pay before you get any product that can be sold to customers.
Note that I haven't said that PCB manf is hard. I've even said that companies don't do this stuff in house, but that they outsource it.
I'm pretty sure Sunstone doesn't charge for tooling if you use their proprietary software. You're locked in, but the costs aren't too bad for development. I'm sure getting the actual gerbers is a pain. ExpressPCB has a similar deal where they will make three index card size boards for $51 total. Not too bad, but they charge ~$60 for the gerbers.
If you're using "normal" software that can generate gerbers, there are other options. If you don't mind long turnaround and overseas fab, Seeed Studio will make boards for ~$1 each, for 5cmx5cm [1]. Slightly more for larger boards.
For hobby projects and prototypes, OSH Park can't be beat. Started as a board pooling coop for open hardware, they now send out batch jobs a few times a week to Amitron in IL and produce high quality boards (ENIG, with silk screen and soldermask) for $5/in^2 (for three copies, so a 1.5in^2 board would be $3.75 each)[2]. Their ordering process couldn't be better--upload gerbers, approve an on-screen rendering, specify the number of boards, and enter your billing info. No confusing options to go through. Seriously, try it. It's awesome.
So board fab is not that expensive anymore, and perhaps cheaper than when you last looked. The turnaround times aren't great for the inexpensive options, but with luck or good EDA it's possible to get by without too many iterations.
Board fab is over an order of magnitude cheaper than that these days. Yes, I have old circuit boards that I designed 15+ years ago that cost me $400 for two bare PC board prototypes, but today I can get the same two for about $50 these days, shipped express from somewhere in China.
To go even more extreme, I just ordered a handful of an existing design from OSH park just to see what their quality is. Three tiny boards cost me $5. Five bucks!!!
Regarding your comment about inventory costs, it seems that if this trend for creating hardware items takes off, there might be an opportunity for a company that does volume manufacturing on-demand. Recently, most volume manufacturing of physical goods has happened in China because the labor costs there are much lower. But I'm thinking that a company which does mass production on demand, cheaply and _well_ would be very successful. It would be like an infrastructure company for physical goods.
No reason why this company couldn't be located in Asia as well, but having very close and accessible communication with the people who use its services might also turn out to be an idea that was stupid until it wasn't.
Recession hits these companies hard, because of the costs of stock; labour; etc, and because volatility of customers makes long term planning tricky.
And, really, the components that are common over many projects are cheap. It's the stuff that's specific to a build that is expensive. The PCB, the case and tooling for the case, the connector (sometimes), these are the big costs.
Indeed, looking at capital costs for our product the injection mold tooling is one of the most expensive aspects of production. Rapid prototyping (CNC milling/laser cutting/3D printing/etc.) reduces the cost of iterative development, but there's no way around what it costs to ramp up for production.
Margin is an issue when you are making just another router.
Cost of inventory is an issue, but there are several ways around it, including dropshipping and createing on demand. You can hire an MBA to optimize in this area (this is the area where they are supposed to do well).
Basically if you could honestly imagine your product doing well on kickstarter, you don't need to worry about margins.
If you can imagine your customers being willing to wait 3 weeks on the product you don't have to worry too much about inventory.
I very much doubt that YC will back a new factory to produce hardware for others.
I very much believe that YC will back a company that produces an alarm clock that makes you want to start the day, or a device to insert into your toilet that will analyze your urin and tell you if the measurements are out of whack.
I have a passing familiarity with some well-regarded MBA curricula and disagree that these problems are things that b-school grads leave school equipped to handle.
Here's a Slate story about a Kickstarter hardware project that failed spectacularly. The income Kickstarter generated was counterfeit, because the guy running the business was swamped by expenses he hadn't predicted out the outset and would only learn when he needed to ship in volume. Spoiler: all he was trying to sell was lockpicks, little shims of metal.
This is where good business people shine - whether you call them MBA's or not. Often they're not.
Good business people will anticipate the expenses needed and make sure there's a way to pay them (or delay payments and make sure everything is in a company with limited liability), borrow money in the bank with collateral in your inventory, get good deals with suppliers (If you do it right you'll get a 90 days credit line thus nullifying much of your cashflow problem), etc. etc.
That article is a blurb, when compared to the length and detail I've come to expect from Slate articles. Really, it starts by telling how a bunch of lockpicks broke at a lockpick convention, then jumps into the excitement of Kickstarter.
The article discusses how this guy abandoned his friends, launched their project without them, then went money-crazy when the project was funded. This guy is spending on conferences, unlimited air passes, and $24,000 worth of locks and Open Locksport swag, but doesn't even have a source for the lock picks he's "sold" to people.
Maybe the best part is where he has to spend $2000 to hire a go-between to talk to lockpick suppliers, as he was too insecure to do so himself. Maybe one of his friends could have done that, if he hadn't abandoned them. No...the best part is where all the conferences and flying and vacations has caused him to become depressed, so he checks himself into a hospital. sigh
Oh, and I think the friends he abandoned came back to sell his locks and sell this story. I'm not completely sure if they're his friends, because the Slate article is very unclear about this.
Those are from the top of my head. Anyone from software who is jumping into hardware needs to stop and think: it's different, and that needs to be reckoned with. It's not impossible, of course. :-)
Do you have experience with the FCC certification process? We're working on an electronic instrument, and it would be cool to hear where the pitfalls lie on the route to getting certified. Wouldn't want to spend too many trips visiting an anechoic RF chamber...
It really depends on how complex your product is, and how good your PCB designer is. We are getting into some decently high-speed designs (LVDS/USB) and it is definitely a learning process with every board...
As far as pitfalls, they range from minor product alterations like requiring the addition of ferrite beads/elements on power lines/other wires inside the enclosure or altering component values on the board, to major alterations that require another board spin (adding extra passives like ferrites, decoupling caps, or just better PCB trace layouts), to very bad problems that require major redesign and rearchitecture of parts of the product.
If you are still in the board design/prototype stage, try to add spots on data lines and high speed traces for: extra capacitors in parallel for decoupling and to target power supply ripples and harmonics, and 0 ohm resistors in series to later replace with ferrites to absorb EMI or low-value resistors to correct for trace impedance mismatches. You can always remove these when the design gets closer to production.
Budget ~$1-2k extra as well if you want the testing house engineer to try various ferrites and shielding to get a design a few dB out of whack to pass while it is in their test environment. Of course, some of their recommendations may be too expensive/unworkable and some redesign is in order.
Unfortunately, all of this costs time, money, and enormous amounts of energy cursing the gods of physics and RF when things go wrong... but that's part of the fun of designing new things, right? :)
And if you're doing anything in the critical infrastructure space, certifications are ABSOLUTELY KEY. Things like ISO9001 are no joke in the hardware space. And those can take ages to get and sometimes get very political.
Barriers to entry are a good thing for incumbents and for the market leader. They work against all new entrants to a market, including the entrants YC picks.
> Barriers to entry are a good thing for incumbents and for the market leader.
There was a time long long ago where it was also considered a good thing for startups, so that young funded companies would not be competing with bootstrapped/small-angel deals.
There is a kind of startup that comes pre-packaged with a solution for some perceived barrier to entry. For instance, your team could include a world expert on some heretofore intractable problem in your startup's problem space. Or, you could be the first mover in a market with significant network effects, and you could be raising when you already have significant traction. In these cases, barriers to entry are a good thing.
In other words: the barriers to entry behind you are defenses for your success. The barriers to entry ahead of you are obstacles to your success.
> The barriers to entry ahead of you are obstacles to your success.
If you choose it right, they are obstacles to everyone's success, which is kind of the point. A well-funded company that can make a biz work with lower-margins and inventory costs will likely not be competing with bootstrapped (for example) companies.
Edit: pg is more pithy below -
Barriers to entry are good for those who can overcome them.
Why on earth would barriers to entry be a good thing? As far as I can tell they are just a proxy method of limiting the true competition capitalism is supposed to offer.
Different kinds of "good". If you don't limit the true competition that capitalism is supposed to offer, then you can't make economic profit in the long term. If your goal, in starting a company, is to continue to make out-sized profits, then there being barriers to entry that you can overcome better than others is a good thing. One way, but not the only way, of overcoming obstacles better is simply overcoming them first: if the cost of overcoming the obstacles is more than the upside in a competitive environment, then you can rake in monopoly profits without it being worth it for others to join you.
Most of this is not necessarily "good" for the consumer or the economy as a whole - although if the barriers to entry are natural and it means that problems are being addressed that otherwise wouldn't be addressed at all, then it might be.
It's also noteworthy that startups who can overcome these barriers have not only the advantage over other startups, but also the incumbents who are clunkier and less agile. sweetspot
15 years ago I did a hardware startup, development was extremely hard and I woved never to do a hardware startup again.
Now I'm not so sure.
Our problem was twofold: First hardware development was not generally something you just did in your sparetime, it was for the big boys. This meant a lot of convoluted processes for dealing with suppliers, expensive and unreliable dev kits and tools, long lead times and all sorts of other hassles.
Second the turnaround time for a prototype was at least two weeks. If you made a small mistake you'd find out two weeks later when your prototype arrived. This adds up quickly and slows you down tremendously. Not because we made a lot of mistakes, but we had to be absolutely sure that something worked before sending it off to be prototyped. No testing a new idea in an afternoon or two.
These things have totally changed with the commodisation of hardware and the looming 3D revolution.
With a makerbot, a raspberry Pi, an arduino and a shelf full of components you're prety much ready to go and can hammer together a working prototype in no time. If you feel cheap you can buy a nice box for your arduino and call it a finished product.
> With a makerbot, a raspberry Pi, an arduino and a shelf full of components
From recent experience working on a hardware side-project: these are extremely cool and useful development/indie tools. But when you get to the stage of large scale production we are still lacking easy entry.
I've been working on an XBEE home automation project intended to be run from a Rasp Pi "server" (but designed to run from anything that can boot linux and has GPIO). It's looks great, has a nice enclosure and is something any hobbiest could get working with the right plug and play bits.
But.
Packaging this for mainstream, where a user can simply buy "black boxes" and connect them up is tough. Circuit design and fabrication on any scale requires both skill (i.e. in designing the layout - I am an Electronic Engineer but of the solder-stuff-together ilk rather than CAD) Admittedly places such as Batchpcb etc. make prototyping a lot easier, but the hurdles you mention are still there in many forms.
The maker/hacker collective is very well serviced nowadays. Next we need someone to disrupt the small-scale development arena: off hand, a community of PCB designers/hackers available to make bespoke PCB layout which links into a batch production system would rock! I've hacked together my prototypes, but have minimal confidence they will work first time :)
Plus... a lot of these companies are side-stepping CE / FCC certifications by clinging to the "we only sell to 'developers'." When you want to sell to non-DIY consumers, I'm not so sure this is an option anymore -- especially for projects like home automation where there are serious safety concerns (fire and electrocution!).
I've never gone through that process. Anyone know ballpark costs for certifications?
The way we got a CE certification for our product was by hacking the system. It turns out that the way it works is that some legal entity of your own choice puts a CE-approved sticker on the product and can then be held responsible if the product doesn't meet the regulations laid forth in the directive that deals with whatever your product was (In our case the machine directive).
Of course the way it's meant to work is that your company hires an approval company that knows a lot about certifications, tests the product and then approves it knowing that they're on the hook if it doesn't meet requirements and regulations.
Nowhere does it say that the approving company can't be the same company as the one making the product, so basically we bought a few CE stickers put them on the product and it was approved. By us.
* The product may be placed on the market only if it complies with the provisions of all applicable directives and if the conformity assessment procedure has been carried out accordingly.
* If stipulated in the directive(s), an authorized third party (Notified Body) must be involved in the conformity assessment procedure.
I'm not sure if the machine directive stipulates it, but "authorized third party" seems to imply that the manufacturer can't perform the certification themselves.
Even if you can count yourself as the approving company, did you make sure that your product complies with all the provisions? Or were you taking advantage of some loophole somewhere to skip the entire process?
If stipulated in the directive(s), an authorized third party (Notified Body) must be involved in the conformity assessment procedure.
The "If stipulated" part seems to be the key. Below your quotes in the wikipedia entry there is a section that covers self-certification. While it doesn't list the directives comprehensively, the steps there do seem to indicate that self-certification is not unanticipated.
It's not really "Hacking the system" - it is more than acceptable for a company to self certify so long as you are happy with the documentation you collect from the component manufactures - it's just that you assume all the risk if you are fund to be in breach of the unifying directives (the legals that sit behind the legislation).
One of the reasons the test houses exist is that not everyone wants to assume the responsibility of self certification and so they outsource it to experts in the test industry.
The most important thing to make sure, regardless of the route you choose, is to always have your certifications and technical documents to hand should one of the monitoring bodies ask for it
Our recent projects through the FCC were about $5-6k for a single product, and around 1-2 week turnaround time.
Granted, we have intentional radiators (like wifi radios, cellular modems) in our products so it may be more expensive. YMMV of course.
Edit: with the comment by thomaslangston, I should clarify that the included radios already had their own FCC certifications; this was just Part 15 for the whole device.
Well. In terms of technical difficulty CEcertification is not ridiculous. However it does require not-insignificant amounts of time (writing the documentation) and money (certification).
I have read somewhere, that using pre-certified parts (wifi, boards, etc.) can dramatically streamline that process. But I have not done this personally so treat this advice as urban myth until you find further verification.
I give you, the Open Pandora project - a bunch of open-source hackers who are spearheading the hardware rennaissance and showing everyone, along the way, how to do it - and not do it (also: volcano):
One day, probably soon, there will be books written about this project. It is an awesome thing - the device itself, the people involved (EvilDragon and his merry band VERSUS Craigix' Reality Distortion Field [really, he has one!]), and the whole community. The Open Pandora Community is a very, very, very fertile y-combinator with proven track record of delivering, within its own realm of course, much value.
So, check it out, yo. Open Pandora .. 2.0 .. could be a YC project? Or .. the inevitable spinoffs? (The Open Pandora is old enough now that there will soon be splinter factions establishing new product, btw..)
> Second the turnaround time for a prototype was at least two weeks. If you made a small mistake you'd find out two weeks later when your prototype arrived.
[Cue Four Yorkshiremen]
Try an fabless IC company: 8-10 week turn times on fabrication, each run was 6 digits $.
(this is from a throwaway to avoid naming the company I worked for)
I worked at a startup with a hardware product in 2011 (nb: I didn't work on the hardware personally, but was aware of the difficulties we faced), and the issues you describe are all very familiar.
It may be the case that the technologies you describe allow you to push the problems to later in the product development cycle, but from what I observed, they all still happen once you try to transform that product into a consumer product that can be manufactured and sold at a viable price-point.
It's true for example, that the hardware engineer founder built a prototype relatively quickly using off the shelf components, but the resulting product was so large, the housing so ugly, and those components so expensive that it would never have been economically viable.
Once they were past the prototype stage, they faced all the problems you describe: parts manufacturers and suppliers only willing to deal in huge quantities and make sales well in advance; finding manufacturers who can make plastics that are of high quality and meet safety standards is surprisingly difficult and slow [1]; much of the tooling for small, cheap, low-power-consumption microcontrollers seems to be from the dark ages [2].
What made it even worse was that many VCs we spoke to seemed to be unfamiliar with or unwilling to take the risks associated with hardware startups: much higher up-front costs to scale the business, and much longer product development times. It's possible the company I worked for didn't take advantage of all the modern developments available to hardware startups, or that they had the wrong strategy (trying to scale too fast or seeking the wrong customers, for example), but I certainly have enormous respect for the challenges hardware startups must overcome, even in these times.
1: Some anecdotes I remember from the plastics saga: the first overseas manufacturers they tried in an early attempt to make the housings cheaper absolutely would not, and maybe could not produce a housing in the correct shape. We would send them models in a form designed to be structurally sound and aesthetically appealing, and they'd come back as basically ugly square boxes. We ended up using a much more expensive US-based manufacturer for a long time because of this. (IIRC, they were more expensive mostly because their facility was really meant for rapid development and not mass production.)
I also remember a significant delay at a crucial time because we couldn't get the fireproof coloring agent we needed in the right color.
2: One of the hardware engineers spent a grueling couple of weeks hunched over an oscilloscope looking for a particular pulse that would indicate that the microcontroller had found and booted our company's software on the chip.
It's quite curious how hardware design lost its cachet.
When I was in Computer Engineering at Waterloo (class of '98), many of my classmates were vying for coop jobs in FPGA, ASIC or board-level design. I happened to be one of the lucky ones and it set me on a path to a career in ASIC verification.
Today when I look around at my industry, it's downright shocking how little young blood there is around. The youngest ASIC guy I've met in the last 6 or 7 years had a Master's and three years experience - and we all regarded him as the newbie.
Only in the last few months have we seen the occasional new-grad's resume cross our threshold. I'd not yet call it a trend, but I hope it becomes one. We have horrible languages, horrible libraries, horrible tools - a huge, shaky mess of infrastructure built on technologies stretched far beyond what they were originally built for - and it's so deeply entrenched that few of my colleagues seem to recognize it. We really need a new generation of fresh perspectives to shake us all up.
Let's hope it is indeed a new renaissance that Mr. Graham heralds.
I have a friend who has been trying to convince me that it is feasible to create a language that can let a general purpose programmer write useful software and "compile" it down to a FPGA to run really, really fast. (He has a lot of specific thoughts that I won't share about what such a language should look like.)
In principle he thinks that it should be possible to compile such a language down to an ASIC representation, but that would only make sense for very high volume runs. (He's specifically interested in software to run financial models, so those would not be high volume runs.)
He makes sense to me, but I have no idea what practical difficulties there would be. But does something like this sound doable in principle from your point of view?
There are a number of solutions that do that kind of thing to a degree - and they are improving. For certain types of designs, you can actually compile fairly unrestricted C and C++ into gates that are faster and lower gate-count than hand-coded Verilog designs. Your friend's domain might well be one of them.
I think you need some specialized knowledge to use this existing generation of tools, but having never used such tools, I couldn't quantify that.
I'm also not sure how well these tools perform on arbitrary hardware designs.
But no - I certainly don't think he's out to lunch.
The problem is that C and C++ do not implicitly parallelize well. To get maximum bang for the bunk you want to extract lots of parallelism. Which requires a different kind of language.
The interesting flip side is that if you design for easy parallelism, you are better able to take advantage of multiple cores in standard CPUs. Also you can then turn it around and also distribute the same program across a cluster of commodity hardware on the cloud.
Bluespec is trying to do something similar to this, but they focus on synthesizing hardware, software is a secondary thing. However, it's designed to run well in software too.
Originally implemented by one of the major Haskell gurus, Augustusson.
You can already do this with Labview and National Instruments hardware. It is quite easy to program and Labview is inherently a parallel language so you get that for free, but the price is way too high.
One FPGA module costs around 3000 $ and a Labview licence costs around 10000 $, at least in Denmark. It is probably cheaper in the US.
Commercial EDA vendors have solutions starting from C, Matlab and LISA for synthesis of HW logic into FPGA or ASICs. One such list of products is here:
"We have horrible languages, horrible libraries, horrible tools - a huge, shaky mess of infrastructure built on technologies stretched far beyond what they were originally built for - and it's so deeply entrenched that few of my colleagues seem to recognize it. We really need a new generation of fresh perspectives to shake us all up."
I have somewhat recently transitioned from embedded software to verification, and you have hit the nail right on the head. We should talk.
What, you don't think systemverilog is the cat's meow? </sarcasm>
I've been in the industry for about 15 years. I agree the languages are horrible. The tools can be buggy, but I don't feel they are all horrible.
There are open source verilog and vhdl simulators out there, although I haven't used them yet [1]. They could be a good starting point for tool improvement. It mentions synthesis as well, but anything practical for hacking would have to target FPGAs and that means vendor tools have to be involved.
While the hardware implementation languages could stand improvement, verification is the largest consumer of resources right now. Being able to work in something like python or clojure for testbenches would be really interesting.
A matter of degrees, I guess. They're certainly improving, but we run into bugs and missing core language features in all three major vendors. I work for a consulting company, and we switch simulators and coding styles from one gig to another, so that may amplify my perception of bugginess. But just as examples, in the last couple of years, my coworkers and I have discovered:
- a case where one simulator gets the xor of a two-state and a four-state logic vector wrong.
- a case where another simulator converts from 4-state to 2-state value with the wrong result.
- one of the simulators doesn't implement a single-precision floating point type
Those are pretty fundamental bugs - not the class of bug I expect to find in a tool that has been in the wild for 5 or 6 years and costs more than my car for a single one-year license. And we've found many, many more less-blatant ones.
I couldn't get pyhvl to work, reported a bug, and got no response.
For verification there should have been effort expended on creating libraries for existing high-level software languages and on interop between those languages and simulators, not on creating our own iffy language (SystemVerilog), and then a huge library to work around all the deficiencies and impedance mismatches between the language and the problems we are trying to solve (UVM).
I'm in Comp Eng. at the University of Ottawa right now, and I'm having a hell of a time getting any hands on experience with FPGAs or board-level design. I worked for one company on Co-op which was good enough to let me try to layout a simple board, but I'd simply never seen the best practices at work, and nobody really had the time to teach. I found VHDL really interesting, but I've only had about two classes worth, and I haven't really has a hobby project. I've though about picking up an Altera dev board and emulating an old computer architecture; is that the kind of practice you think would be worthwhile? Do you have any tips about learning board design and what kinds of projects/hardware are good to get started on?
I'd heartily recommend to make the most of co-op opportunities if you can. I learned way more in co-op than I ever did in school. I suppose it's harder these days, as resources are stretched further so it's difficult for staff to allocate spare cycles for mentoring. It's unfortunately myopic, as we really need to invest more in guys like you for our future.
Real board-level skills are difficult to acquire on your own. I've only done a little - and that was last millenium - but a lot of it is a black art: rules of thumb here and there about what works, what doesn't - and they're best learned as an apprentice to someone who's been doing it a while.
VHDL is probably a cleaner language than Verilog/SystemVerilog, but it's used less in industry - at least in ASIC designs. There are free Verilog compilers/simulators (Icarus Verilog being a common one); there are probably no free SV implementations. Not sure what the status of free VHDL is. There are probably decent cheap options for development boards from Altera, Xilinx and co.
Feel free to drop me a line if you want to discuss further - email in profile.
Having looked at this site, these boards seem very affordable and extensible. The Spartan-3, in particular, seems to offer everything I could want, and their development environment looks to support VHDL and Verilog. My only reservation is that the free tools they offer from their website are probably not very much like what's really used in industry, but it's a very cool start. Thanks!
For simple board design projects, I recommend starting with a microcontroller, such as AVR (as used by the Arduino). The chip manufacturer usually provides reference boards as a starting point / cheat-sheet for your design.
If you'd like to go deep into embedded system design, I've also accumulated all of notes at: http://bringuplabs.com/guide/smartphone/
I'm actually looking for students to give feedback on this curriculum format, so feel free to message me if you have any questions!
This is addressing a different level of hardware. I don't think the primary drivers of a hardware revolution (new tools, cheap and rapid prototyping, etc) extend to ASIC design. As a recently graduated Masters EE student, none of my fellow students considered VLSI and chip design to be a new or hot area or seriously considered it as a career path. I don't doubt that there's plenty of innovation there, but it's rightly or wrongly perceived as being less sexy than other subfields of EE.
While I admire your moxy, you need to understand that a full suite of tools we use to get ASICs designed, verified, laid out and ready for fab cost 5 or 6 figures per license, depending on what you're doing. The very use of the tools themselves, is subject to non-disclosure agreements. So are the details of their pricing.
I do think the status quo will be radically reshaped by an external force one day, but thinking one is going to mount an assault without knowing who the three main EDA vendors are - let alone without real-world ASIC experience - is quixotic.
It is also not obvious to me how the economics of a viable competitor would actually work.
[EDIT: I hope this didn't come across as excessively abrasive. You are quite right - it's an industry that needs shaking up, but its not obvious how to do it. Fresh faces and new ideas would help us at least realize that there is a problem]
My biggest gripe would be with the abomination of a language we have for functional design and verification, SystemVerilog.
- It has been developed without public discussion, now that the IEEE has shut down their mailing lists to non-paying entities and persons
- The steering committee are bright, well-meaning individuals... who don't appear to have any language design experience between them, and whose knowledge of programming languages appear to be limited to Verilog, Java and a daring dash of Tcl. They have never seen a feature they didn't think was worth adding as new syntax.
- It retains backward compatibility with Verilog, which was a cool hack in 1983.
- It basically looks like Java with a whole bunch of additional (inconsistent) syntax for stuff that should probably be in a standard library, but...
- It doesn't have a standard library
- Oh, and by Java, I mean Java as it was in 1995. Or at least, it will have sort of the same feature set, once they add interface inheritance - later this year. Yup, 2012.
- Its metaprogramming facilities are weaker than ANSI C, which is why most large ASIC shops end up brewing their own metalanguage syntax, usually by having embedded snippets of Perl or Tcl.
- And this inexpressive, crippled language somehow has a spawling mess of a spec that runs a whopping 1200+ pages.
I used Verilog a little bit in college, while studying for my EE degree. It didn't seem like a bad language. It was a weird blend of register transfer-level stuff and C-like pseudocode stuff.
It sounds like Verilog fell prey to the C++ philosophy of language design-- just keep adding anything that sounds remotely like a good idea to the core language. If it interacts poorly with what's already there, that's just too bad.
Unfortunately, new programming languages are NOT a good startup idea, in any field. Also, if people are paying 5 and 6 figures for tools in SystemVerilog, change is going to be slow.
Some of the tools we used in my computer engineering curriculum were HDL Designer from Mentor Graphics and some tools from Xylinx that were similar. If I remember correctly the reps from Caterpillar said they invested in HDL Designer pretty heavily as it was used widely there. I mention Caterpillar specifically because their world headquarters was about an hour and a half down the road and they recruited pretty heavily from UIUC.
HDL Designer was cool but it was pretty outdated looking even 5 years ago when I was at school. It was all motif running on Sparc Solaris boxes.
"And in particular, don't be deterred from applying to Y
Combinator with a hardware idea, because we're especially
interested in hardware startups."
How are you planning on funding hardware startups? From your site:
"Usually $11,000 + $3000 per founder. So $17,000 for two
founders, $20,000 for three or more. Occasionally we
invest more. The goal is usually to give you enough
money to build an impressive prototype or version 1,
which you can then use to get further funding."
A real hardware startup would require at least an order of magnitude more money than this. Unless they walk in with all the required tools, you could burn $20K just in software licenses (Solidworks, Altium Designer, Xilinx Foundation, Keil, etc.) and not have much left for other stuff. Heck, my DSO alone cost me about $20K. The computer I am typing this on probably has $50K in hardware and licenses on it.
I could personally consider the idea of presenting a hardware startup to YC, but I would need to know that this is not about (with the utmost respect) finding a few starving 20-year-olds that will kill themselves for a $20K investment. From my vantage point, if you are not throwing $250K+ into a hardware startup it just isn't going to happen. Of course there are exceptions to every rule. Then again, this ain't my first rodeo.
Now, if the idea is to throw some money at a project to cobble-together a smoke-and-mirrors prototype and then go raise a few million, well, for the right project this could work.
Again, I say the above with respect for what you do. I have done a lot of hardware/software/multidisciplinary development. It's very different from pure software web/mobile startups. Very different.
This is an important point and needs to be addressed. $20k doesn't go very far at all in the hardware startup world.
That said, it is possible to be frugal--to a point. My team and I try to use inexpensive solutions where we can. We're using Rhino instead of SolidWorks (OK until we need the hardcore simulation functionality), KiCAD instead of Altium (same, until we need the simulation), MikroElektronika's C compiler for ARM rather than Keil, MeshCAM rather than Mastercam, a cobbled together kit CNC mill rather than "real" mill, a Chinese-made laser cutter rather than an Epilog, and a wide array of machine shop equipment sourced from Craigslist.
Our product also touches biotech, so we have outrageous reagent costs too. We still have shell out for antibodies, fluorescent probes, proper biosafety lab space, etc.
I'd love to see YCombinator comment on accepting biotech/healthcare companies.
As you said, to a point. In my experience some of the approaches to frugality can quickly become counterproductive and eat away critically needed development clock cycles.
Here's an concrete example: I made due for years with cobbled-together CNC equipment of various kinds. I bought an old Bridgeport CNC mill really cheap. Within a month of use the spindle transmission had to be rebuilt. It did not have an automatic tool changer or tool probing. After that I took a Bridgeport knee mill and converted it to CNC. No tool changer or tool probe. It did OK. Both of these represented HUGE time drains and detracted from the main mission in major ways.
Later on I decided to do it right and leased a Haas VF6-SS. If I remember correctly the lease ran about $2,000 per month. Best decision I ever made. All of a sudden CNC was not a research project. The thing just worked and it could work 24/7 if required. It had an automatic tool changer and automatic tool probing. When properly married to Solidworks and a CAM program the setup was smooth as silk. This improvement in productivity was well worth the $2,000 per month and other costs (tooling, setup, training) in that it quickly returned everyone to being focused on what we were actually trying to design rather. It also opened-up design possibilities that did not exist before this machine arrived due to what was lacking on the cobbled-up machines (fourth axis being one example).
That's not to say that I am not for being frugal and creative whenever possible. For example, we used the CNC machine to manufacture custom assembly tooling that would have cost tens of thousands of dollars to purchase. We also made our own sub-table for the machine and saved tons of money.
You'll have to raise more money, of course. Currently, all YC companies are offered $150k automatically and companies that look promising to investors are able to raise a lot more. I believe it's possible to get most good ideas to the stage of "promising" on $170k.
As another poster said, there are starting to be alternatives to these expensive and hard to learn software tools for the early prototyping work. I'm the founder of Tinkercad and we offer a web based solid CAD tool that is significantly quicker to get started with than Solidworks. There are similar tools emerging for circuit design.
A lot of our users are working on hardware startups. Once they converge on a working prototype many move forward to more traditional packages to optimize the product for larger production runs. The cloud based offerings are also developing very fast, simulation is a good example of a problem that benefits greatly from being run in a large cluster.
I applaud you for creating such tools. They obviously have a place where they fit very well.
As for what you term "hard to learn" tools, well, I'll have to disagree with you on that one. At a professional level tools such as SolidWorks are an absolute necessity, not a luxury. The list of needs is endless, from parametric modeling to assembly management, mechanical, thermal and fluid simulation, interference evaluation, draft analysis, configurations, documentation, sheet metal tools, molding tools, CAM integration, DFM integration, programmability, etc.
Parametric design alone (and table-driven parametrics) is a huge advantage in these tools.
And, as you get into complex assemblies of assemblies with hundreds or thousands of parts and thousands of fasteners things get interesting too.
The same is true of circuit design tools. As an example, once you get into serious high-speed digital design you need tools that can facilitate the design, simulation and layout of complex multi-layer controlled impedance data paths as well as power distribution systems. At the higher end you have to get into field solvers and various SI (signal integrity) type tools. With FPGA's you need tools that, at the very least, make the task of working with devices that have thousands of pins as easy as possible. Bank swapping, for example, is a nightmare without the right tools. Heck, dealing with any IC's with more than, oh, 100 pins per package on a less-than-capable EDA tool is a recipe for disaster. I did a PCB with an FPGA that had over 1100 pins. The design started on a lesser EDA tool because, well, long story, but the client didn't know any better and didn't want to spend the money for a better tool (I use Altium Designer, but there are others). After 16 hours spent defining the pinout and PCB footprint for the FPGA (with verification, etc.) he got the message. The same job on Designer took about 30 minutes tops.
As for simulation run on clusters. Yes, of course. There are services that will take SolidWorks fluid simulations and spread them across multiple nodes for speed. I've fluid/heat-flow simulations that have taken upwards of 18 hours on 64 nodes. Fun stuff.
Like I said above, there's a place for what I will term "entry level tools". This is not intended as a derogatory term at all. That's just what they are in the context of what "pro-level" (another grab for a term) tools can do.
I will definitely keep Tinkercad in mind to recommend to friends with projects that might fit what the sofware can do. I wish you huge success. Get it into schools. Kids need to learn to create this way.
I'm not sure if hardware hackers want to be Steve Jobs... they want to be Woz!
"Woz soon followed with the machine that made the company, the Apple II. He single-handedly designed all its hardware and software—an extraordinary feat even for the time. And what's more, he did it all while working at his day job at Hewlett-Packard"
If pg is correct and hardware startups are making a renaissance then we can expect to see startups soon that sell shovels for hardware startups. For web startups we see shovel-style products that help with customer support, A/B testing, virtualized servers and so on.
In the same way we can expect startups to pop up that make life easier for hardware startups. For instance startups that make prototyping hardware easier. Or that simplify shipping goods all across the globe. Or startups that make it easier to find the right suppliers and get good deals with them.
Don't know why hardware is suddenly rising like a phoenix? One word: Apple.
Apple used to talk about ease of use and "computer for the rest of us", and they still throw some of that into every presentation, but it's no longer the lead. After Steve Jobs rejoined Apple, it was all about the look and feel of the hardware, not ease of use of the software. "Look, candy colors!" "Look how small! Feel how light! Look at those curves! Oooh, you can never be too thin. Did I mention thin? I meant thinnest!"
Then the small, thin, candy-colored devices emerged, and they made Apple the most valuable company in the world, with the media hanging on its every move. And what are they talking about, the software? No, silly, a Windows machine can do anything a Mac can do. Apple wouldn't be so successful without the full package of hardware, software, and services, but it's the HARDWARE more than anything else that everyone talks about.
This has given Google, Microsoft, Samsung, and other big companies serious Apple envy, which shows up in their "strategies". How can it not affect small startups as well? It's the environment these little companies are born into.
And in particular, don't be deterred from applying to Y Combinator with a hardware idea, because we're especially interested in hardware startups.
This is great news. I felt a bit old and stupid at Startup school, because I want to make hardware (albeit for a niche rather than a mass market), but anyone I talked to seemed to find the idea weird and most of the concepts from the stage were along the lines that software - particularly internet software - was, is, and ever shall be the sole basis of a successful startup. It was a relief at the end of the day to hear Joel Spolsky discuss the viability and possible desirability of building a nice little $10m company in just as much detail as building a $1b one.
I'm still struggling to understand the logistics of starting a hardware company.
Let's say I have an idea for a portable mini-fridge. What kind of engineers do I need to contract to design it? Where do I find them? Do I need approval, testing? What kind of regulations do I need to follow? And finally how do I get it manufactured? Do I actually need to go to China and meet with factories?
I wish someone would put together a step by step guide for these kinds of questions.
Regulations differ across different industries and most have an industry association that you can become a part of to do research. Manufacturing is very old-hat in North America.
Truly rapid prototyping is bringing costs down, like a great dev framework cuts dev time down.
There are many possible paths to succeeding in manufacturing, one size does not fit all. It largely depends on A) what you're trying to build; and B) who you're going to sell it to.
For example, if you're going to sell to consumers you're most likely going to have to come up with capital to cover your manufacturing costs before you have a single unit sold.
If you're selling to businesses you might be able to get a Letter of Intent or a Letter of Credit and then get a bridge loan to cover the costs. If I were building a manufacturing company, this is probably the route I would take.
Like software though, you need the right people. Good mechanical engineers are as rare as good software engineers and the programs they use in their craft are far more expensive (SolidWorks, etc.). You'll also probably need an Electrical Engineer. It's kind of like the difference between front-end/back-end devs.
The simple answer is that it's no different than a software-only startup: spend 10 years learning how to build [software/hardware], then the logistics are straightforward.
The trend towards more hardware-product startups in the sf bay area is real and there is more evidence than just the increase and quality of yc hardware companies that pg wrote about. There are new hardware focused incubators (like Lemnos Labs), more hardware hackathons, and there's been growth in the sensor and hardware meetups for the past couple years.
However, it's possible that innovation in hardware technologies is out of reach of startups. I'm differentiating here between hardware-tech and hardware-product startups. Of course there are hardware startups that are attempting to bring new technology to market (go Integrated Plasmonics, and 3Scan!). Interestingly, a good indicator of whether a startup is developing an innovative technology is that Peter Thiel is invested in it--many of his fund-ees are slaving away in labs scattered around SF. The failure modes of these hardware-tech companies will be more interesting than the hardware-product startups. These companies may take a much longer time to develop tech, then the product using that tech, and then fail, because the market they try to disrupt with their technology may be disrupted by other technologies with better economics--energy tech is full of these sorts of baby elephant skeletons. Still, successful investment in real hardware innovation that wins leads to companies like the next GE, Intel, or AT&T so it makes sense for investors with deep enough pockets to aim for these.
Hardware products that simply integrate existing commodity components (like Blossom Coffee) have a pretty well understood binary risk profile, they either they hit or they don't. The new hardware renaissance PG is observing is based on the reduced cost of production, and new ways to crowd fund these sorts of companies: things you might prototype in TechShop (shared tools space) and sell via Kickstarter.
I'd suggest that there is a huge opening for a startup that does manufacturing process. This company could earn money by selling its process to others, and dogfooding its own product in a few lucrative niches.
For example, the company might start by pulling all the open source EDA, CAD, CNC, and similar, software into a cohesive whole. It would then bring in one of the open source ERP systems and integrate it with the engineering toolchain. Keeping on doing this for every part of the process: ordering, inventory, manufacturing/robotics, testing, sales, distribution, support, financials. The idea is to (as nearly as possible) completely automate the process of scaling hardware based business from a prototype to a product. Achieve this goal and hardware becomes as easy to scale as software.
I've been building hardware for the last 30 years, and invariably most of the work goes into making the process run smoothly (ie. designing the process) rather than designing the product.
I think it's fascinating how much money people are willing to pay for hardware compared to software. It's hard to convince people to pay anything for software, but these kickstarter projects are getting lots of people to pay hundreds of dollars for hardware.
it's fascinating how much money people are willing to pay for hardware
-- Sometimes there are social resons for this.
Tangibility. Evident complexity. Somthing to show off/talk about. You can sell art-work to hang on the wall for a much higher price than a digital file only, etc. (Software's intangibility and ease of replication in these cases are not per-se advantages.)
> It's hard to convince people to pay anything for software,
If that software is for "entertainment" (e.g. games, social networks, cool email clients), yes. But most companies and professionals will gladly pay for software that makes them money or, equivalently, saves them time.
Patents. That is what is stopping hardware. Go ahead and try to have something listed in Best Buy without having an army of lawyers breathing down your neck.
Every. Single. Hardware. Startup. I know, has either gone bankrupt due to lawsuits, or hasn't ever reached product market fit.
I'm a software guy at a hardware company so I know a little about this. From my perspective I think there's room for a startup that acts as a kind of project manager for the actual "making the physical stuff" part of a hardware company. A group of people that knows the companies who do make circuit boards, design the enclosures, produce the plastics, get FCC/UL/etc. signoff if needed, all that kind of thing. It's all well and good to have a great idea for some piece of hardware but you're going to get screwed on price and timelines if you don't know the right companies to make the stuff for you and it's a pretty complicated process.
* the ability to construct circuits as part of 3D printing
* a much wider range of materials available.
Essentially, you'd need to be able to ship hardware as a digital good. Which raises the huge question of DRM. The recent years of software startups side-stepped that by going the SaaS approach. (You'll notice desktop software is pretty much dead. There are no major new projects in that area since probably ~2008 or so.)
HaaS is going to be much more difficult. (Not necessarily undoable. Minds on too much coffee have suggested going a biological route, where devices just decompose after a certain time)
I don't get the fetish about 3D printing when CNC milling machine are cheaper and better in most aspects. Between a Makerbot and a Roland iModela I know what I want, and it's not the Makerbot...
Software can be remade by reloading the page on your dev server. Hardware has to be rebuilt. You'd need to provide a tighter iteration loop on designs.
Tighter iteration loops does seem like an area where startups could help. Maybe not just with 3D printing, but in other areas of coordinating the workflow as well.
I wonder too if there's some streamlining to be done in the supply chain. Look at how Amazon streamlined retail to a point that they serve not only as a catalog/payment service, but also a fulfillment house. Is there something like that to be done with supply chain? Manufacturing?
I know there's companies out there in those areas, but that doesn't necessarily mean there's not still big gains to be made somehow.
I suppose a startup could make it easier for other startups to access contracted design and manufacturing, but I think there's a lot more regulatory barriers too.
As soon as you're selling hardware, you're dealing with customs, FCC (or equivalent) regulations, a lot more safety regulations, etc.
Actually, I did bootstrap a hardware product: www.growguard.net. Agree, it's not the best looking product, but engineering, designing and puting it in manufacturing cost less than 5k!
A couple of years ago, I was the CTO and cofounder of another hw startup (FTTH space), we did get this thing off teh ground for less than 300k and it is now doing very well and one of the top-3 in it's niche
It looks really cool, and I'm extremely impressed that you did it for less than 5k.
I think I found an interesting problem to solve, and while I'm looking at software-only solutions, I believe that it would be better solved by (simple) hardware.
Your post gives me some encouragement to try it out :) (not right now, but I think I'll be in position to do it next year).
Edit: I see you had both a "Show HN" and a blog post, it's a shame I didn't see them earlier and they didn't get traction here.
I think Shenzhen is more the hardware SV than the bay area. Or more "Designed in California, Made in China". Look at this article by bunnie huang illustrating this fact:
I find these essays really refreshing not only because of their content, but because they're evidence of someone searching for points of uncertainty and investigating them. PG seems to be always looking for what does not make sense and trying to figure it out.
This is a good read with some great points... more and more people are coming into the fold and realizing what you can do with platforms like the Arduino in the space of physical/embedded computing. This is where real innovation is happening. Coming from a software dev background, I picked up an Arduino about a year and a half ago - and it is awesome...it's just so much fun...it's like the feeling you get when you make a great web app but times 100 because it's a real physical thing that moves, senses the world around it and lights up (or whatever). Even if you don't believe it's the "next big thing", if you are a sofware dev, you should get into it because it's so much fun and it makes you better developer when you learn about things about digital logic like shift registers.
<editorial_hyperbole>
I really think the culmination of all this will be what I see as the next logical step after "mobility" in computing (i.e. the idea that you "take your computer everywhere with you"): "ubiquity" - or, in other words, the idea that everything around you - the table you are sitting at in your restaurant, the walls in your office, things that farm your food and control your air conditioner, are computers..and perhaps the most exceiting thing is that this next revolution will (hopefully) be ushered in and controlled by startups. Exciting times.
</editorial_hyperbole>
It seems like the unification of hardware and software (as Apple has done) is where a lot of real value lies. I'd love to see more startups selling innovative hardware/software products.
I agree. If you're not scared of designing hardware and you have been exposed to software, it puts you in a position where you can build a physical object and also program its behavior. You can begin product level design just as a big player like Apple does with its large dedicated teams. With the advent of SOCs, its possible to do most of the hardware side of the design at the chip level. So if you can combine SOC design with high level programming (if needed) you have everything you need to holistically design a gadget. Can one person do this? Yes, I think so. As designers, with the kind of support we are now getting from SOC design tools and software development tools, the only thing missing is your product level idea and your desire to execute it. If you do this a few times, it almost becomes as simple as writing software. This is why I agree why with Paul, we truly are at the threshold of a renaissance in hardware product design.
If you're doing something approaching consumer hardware, no, one person can't do this (until someone does, of course). Being good at the quadrant of hard/soft and design/engineering takes a da Vinci, and he didn't mass produce any of it. Apple still had to hire an industrial designer to put a case on Woz's work. But two people, yes. :)
If you start a startup based on a hardware gadget idea, you can do the initial design of both the hardware and the software and then as you grow you can farm out parts of the process. I am not suggesting you would build a whole gadget all by yourself, but rather that you can conceptualize and drill down into the initial stages of the design enough that you can get your company started with a hardware idea.
Of course we are assuming you are not scared oh hardware(chip) design and you have a background in software. So it would require someone comfortable with both software and hardware. A lot of the students studying Electrical Engineering take up jobs as software engineers. These kinds of people would fit the bill.
This is great that PG is talking Hardware Startups - I seem to remember him being less than enthusiastic before, but either way I'm glad that he's seeing a lot of the same things we on the ground are seeing.
A few months back I announced on HN that I started a http://www.reddit.com/r/hwstartups, so I wanted to mention that we're growing a pretty good community over there as well.
I am surprised there was no mention of the Rapsberry Pi. Open sourcing hardware and allowing one to build their own "widget" is a game changer. You don't have to know about board layouts anymore, you now have a product/package ready for production that is $25/$35 that has HDMI, 2 usb (easy wifi enabled), ethernet, and audio. Just build software and put on SD card.
There are a TON of software products you can build off of this and sell.
One of these hardware startups would grow into an YC of its own. If a hardware incubator can solve the problems of inventory, distribution and shipping, it would pump out products at a much faster rate, with a higher chance of success. You don't doubt Apple's newest product for a reason.
Did anyone else find the bottom line about the (250 word) piece being reviewed by 7+ folks a bit excessive? Honest question. Just feels like a very heavy burden to meet before getting public word out on an idea.
Physical things are great. They just haven't been as great a way to start a rapidly growing business as software.
The same issue applied to software not all that long ago. As with software, the wide availability of consumer-level, cheap or even free tools to make stuff is driving a renaissance for hardware.
And with so much hardware hacking then, as with software, some will see explosive growth.
It's good to see in my mind because hardware can be infinitely more important/useful than software.
I am just glad YC is realising the potential of HW companies. Being a Hardware Engineer, I have been feeling out of place in Silicon Valley Startup Scenes.
People have mentioned lead times in this thread. Just so I can calibrate my expectations - what do people reading HN consider to be a "long lead time" for hardware?
Difficulties of a hardware startup, compared to a software one, is mainly of storage, selling channels and after-sales services.
The inability to iterate quickly is also one of its problem.
Unless someone come up with a solution for these problems, I don't see a renaissance of hardware in the form of many startups (but can be in other forms, since now Win 8 is encouraging hardware creativity)
Personally I think the biofeedback thing sounds very cool. I ordered a biofeedback device from some outfit a few years ago and never came close to getting the thing to work.
> And one of most conspicuous trends in the last batch was the large number of hardware startups.
OT, but please be kind to a learner of the english language: should that sentence be "one of _the_ most conspicuous" or there is some rule by which the article can be used or not in this form?
I'm not sure 8% is really indicative of any trend, also I think the sample size is way too small to draw any kind of conclusions. And no examples of what this 8% are up to, at least in broad terms?
If you look at the bigger picture, rather than the little world of YC you'll see lots of hw in medical, aviation, military, communications etc etc.
Saying that the number of companies that would be attracted to YC for money has risen from X to 8% doesn't really speak of any trend.
There are some examples from kickstarter that spring to mind, like for example, Brydge (bluetooth keyboard/speakers for iPad) and the game controller world has its fair share of hw innovation (like Connect for example) but real hw innovation is out of reach of startups, you need chip fab plants and cross disciplinary skills that are expensive and
locked up in the big company r&d departments. The real companies doing hw are not going to YC for money, they go to their boss.
It's impossible to bootstrap a hardware company. You need capital, otherwise you'll never get out of the tar pit that is small-scale production. So hardware companies will always be much harder than software companies, if only for this reason.
Good observation. On top of that, it may be nice to find a solution where you can easily monitor data flowing out of electronic devices in real time. Certainly a lot of software products you can build on top of the coming hardware era.
pg, do you have a quantified list of hardware startup business models that you can share? For example, as a startup, what are the realistic sales & distribution channels that we can model, financially?
What we're worried about is designing the wrong product at the wrong price for what seems like a limited set of distribution channels with some rather onerous fixed costs and low margins -- certainly compared to software/SaaS sales. How do you wade through that swamp?
What would be the advantages of applying to YC rather than Lemonos or a hardware specific incubator? I would think initial pool of capital is the biggest hindrance.
Well, a train is the most likely place to learn about the combination of electrical and mechanical engineering that is found on rails. The Arduino isn't a bad place to expose new people to electronics though.
Spot on PG. I think we are qualified to add value to this discussion here. We have bootstrapped our plug device (TonidoPlug) couple of years back. Lot of people ridiculed and cautioned us when we have started. Now it is one of the top rated NAS product in amazon without any recognizable brand name. We have built the entire software stack around it. Now, We OEM our software stack to other Consumer Electronic companies. We have done all without any external investments. Also there are fundamental changes happening in the industry that will aid provide impetus to the growth story. 41% of CE sales are done via online now. So we don't need to pay hefty margins required by Best Buy OR other brick and mortar companies. We can sell directly to end customers and pay the amazon sales tax :)
Also many CE companies (Exception of Apple) stopped innovating or don't have access to the exceptional software skills. Typically they buy units from Taiwan ODMs, buy software stack from companies like us, put their brand and sell to their existing channels(Staple, Best Buy).
There is no value addition provided by these companies in the entire supply chain except their distribution. Now even the distribution is coming under attack because of the direct sales model. For instance we ship units to more than 75 countries.This situation is so true when you take a look at home networking products (Consumer Routers, NAS, DVRs). You can do a heck a lot of innovation there and people use it daily.
Now we are sitting with exceptional know-how, ODM relationships, a software stack that can run on all the embedded device (Routers, DVRs to NAS products) and mobile apps on all the popular mobile OS'es. In couple of years we can challenge the incumbents and build a company that can generate 50-100 million in sales and challenge the incumbents. Even if the margins are tight, we can still make at least 20% net margin. We are ready to accept investment from YC if PG is interested.
I don't know how much the inventory issue is mitigated by the fact that YC companies with working offerings seem to be immediately able to conjure up 500k-1MM in funding.
Also, YC's major successes haven't been hardware companies, have they? The last essay I read before this one suggested --- in agreement with the conventional wisdom of VC's --- that a company needs to be Dropbox-successful to move the needle for YC. Not that YC isn't, I'm sure, thrilled to have hardware product companies with traction in their portfolio.
Let me just add a banal point: YC's business strategy is, obviously, "throw everything we can at the wall and see what sticks". If you're considering your first company, that's probably not your best strategy. Even putting aside the big-ticket problems like inventory and margin, there are a lot of other things that suck about hardware: lead times, managing supply chain, QA and managing defects, field recalls, shipping. These problems are so big that major hardware companies have people who don't just have one of those tasks as full-time jobs, but are also famous for being able to deal with them.