The hardware equivalent of an Internet deal

I had a great lunch recently with a successful young Internet entrepreneur, and he posed the strong belief that universities should stop wasting time teaching hardware and focus only on software. Those of us who helped build the infrastructure that enabled the Internet will disagree with that position, but there is certainly a sentiment that the business model has fundamentally changed because you can build an Internet business on very little cash, and this deserves a new VC model.

Rather than debate that one, I would rather show how old-school venture does hardware in a capital efficient way, rather like Internet deals ... and preferably without credit cards ;-)

I know two entrepreneurs who are in their 60s who recently gave a group of 20-something year old Internet entrepreneurs a run for their money in terms of drive, energy, and entrepreneurship. Now these guys do materials--in the laser industry, anytime someone mentioned a project that was good to go except for a slight materials problem, we figured it was 10-20 years from a product, so this is about as far an extreme as I can think of in hardware from the Internet. If you can remember when people first started extolling the virtues of vanadate (Nd:YVO4) as a laser crystal, heralded as the replacement for Nd:YAG, it took pretty close to 20 years to actually make a dent in the market.

Just to make matters worse, let's also recognize that these guys are attacking something slower to respond even than the telecom market, by trying to get the semiconductor market to make a change to their Fab process. I certainly can’t think of anything further from the Internet, with worse customers and more capital intensive, than the combination of semi Fabs and materials :-)

So how do they do it? Well, the theory of deep tech is pretty simple, you have something unique and so valuable to your potential customer, that they are willing to invest in it to get the unfair advantage it offers. They can't get it from anyone else, and while the pain of change will be costly and time consuming for them, the pain of not adapting could well be fatal.

So they set up a simple lab, and beg, borrow, or otherwise get whatever equipment they can, everyone works for equity, and they develop their material solution. Pretty much identical to an Internet deal, except fewer PCs and programmers, more white coats and lab gear. But sometimes late at night, they multiplay with their web peers on Warcraft.

Unlike the web deal, what they create here cannot be copied without running afoul of the patents, or deep knowledge of both the materials technology and the real customer needs. The equivalent burn of this team is about $50k/month, if they were paying salaries, which they are not, at least to start. When they do, it's still about $50k/month because their customers are supporting most of their growth in resources and facilities. If the value proposition is compelling enough, even those stodgy Fab customers will get enthused and throw internal resources at evaluating the material. They run wafers, put them through a raft of tests that would cost literally millions of dollars to do if you were contracting the work.

It's this customer validation, in the absence of revenue, which is a really long time off, that is the equivalent of users or subscribers or eyeballs on a web deal. Now to be sure, if the dedicated user count gets into the millions, then there is money to be made--and quickly--in the web deal, but the dark side of that is the lack of stickiness of those customers, who are easily lured away by the next shiny object. The Fab customers can be lured away, but the more time and money they spend on verifying the material, the harder it is to walk away from, and the more sticky the traction.

Where the web deal is really compelling of course, is in its ability to deliver meteoric rises in users, and possibly even revenue ;-) If it goes viral, web based products and services can rapidly rack up $100M in revenue. Often this is actually someone else's revenue and the web business is taking a 5% clip of it, but sometimes it's all theirs and the company is wildly successful.

In the materials business example, this is going to take a long time--could be as long as seven years to get through the entire Fab cycle, and certainly more than three years to wait for the adoption of the next node (design change). Now they do also have to share some of that revenue with distribution channels, and there is come COGS that limit gross margin to say 80% ... Of course if the material is accepted, then its about $100M per line, on the order of $500M per Fab, and depending on how many Fabs adopt it, it can quickly become serious money.

You won’t see it on Facebook, but it will be in the mobile device you are using to look at Facebook, in the wireless and optical network elements that are bringing you the data, and in the servers that host the web app--so I think they are pretty intimately connected.

Isn’t it wonderful to see the confused look on a 22 year old web entrepreneur’s face when he learns that 2 guys in their 60s are on their 10th successful startup ;-)