I’m taking an exhilarating class this semester – Tech Trek – one of the top undergraduate business classes in the country. It’s all about startups, technology, venture capital, and the like, and I was assigned a firm to research prior to our field study in two months. It was a venture capital firm known as Tallwood, which specializes in early-stage fabless semiconductor investments. Now that I’m nearly done my research, I’ve been enlightened on the semiconductor industry, and as with a lot of topics in the technology world, I find it rather fascinating. The fabless model is rather simple – a firm develops a new chip design, but then outsources its production. One of the first articles I read during my research outlined why this fabless ecosystem was weak, if not altogether beginning to collapse.
I always find it interesting when a firm outsources. My stubborn self would view that as a sign of weakness, a lack of ability to be resourceful and bootstrap. The reality, though, is that the fabless model simply works for semiconductor firms. Intel dominates the semiconductor market, and they’re more or less the only company that can afford to build their own chip “fabs” (factories where semiconductors are produced). Everybody else has to go fabless, for better or worse. The model creates (hopefully ethical) jobs out in Asia, and it reduces costs for firms trying to survive in a market where the risk and cost of failure are both absurdly high. Moore’s Law tells us that a firm is really only going to have one shot to make a good chip design every year or two – if they fail, their chip has already gone down in value significantly, and new technology is probably already out there, and that new technology is making millions for your rival.
But if the article I just referenced is true, and the fabless ecosystem may be on the brink of disaster, what does this mean for a market where nobody can afford to do anything but outsource? It could mean Intel will start taking advantage of being able to build their own fabs, and will hold a near monopoly in the future. However, I find that unlikely. ARM outsources, but I think if push comes to shove they’ll build their own fabs. The market will remain rather status-quo for the bigger players for now (ignoring the gradual demise of the PC and how that probably favors ARM, the company developing more chips for mobile devices). It’s the little guys that intrigue me. The pressure of any sort of tensions in the fabless ecosystem will lead to innovation. Why? Simply because it will be needed. The tech world will be in trouble if the semiconductor market is in trouble, and no matter how you feel about technology, it waits for nobody. The demand for semiconductors, both the core technology provided by Intel and ARM, and the unique, innovative chip technology that provides everything from high-speed wi-fi to noise cancellation will force the show to go on whether outsourcing is a viable option or not.
The dollars and cents will be tricky. I’m not naive enough to think that these smaller companies like Audience can just make the money to circumnavigate the fabless model appear out of nowhere. I simply think venture capital firms will see that chip innovation is what fuels any sort of technology innovation, and if new chip companies with great ideas need more money to combat whatever happens with the fabless model, venture capital firms will suck it up and throw whatever extra money is needed at these new firms. Intel and ARM are great, I’m sure, but I rarely trust larger firms to innovate. The little guys are the only ones that I really think can bring us the kinds of semiconductors that will change the world. And a collapse of the fabless model would mean that the little guys need more money. And to get more money from VCs, they need to pitch a better idea. It almost sounds like I’m hoping for the fabless model to disappear (trust me, I’m not, it’s a system that works), because I think it would cause a vicious cycle forcing semiconductor startups to really come up with game-changing ideas. In contrast to the article, I don’t see any major changes in the fabless model happening soon. But imagining life without the fabless model is a very interesting thought exercise. We’ll see innovation anyway (look to Tallwood VC’s investments to know where those innovators are – these guys know what they’re talking about), but it’s always fun to wonder whether outside pressures might cause greater innovation out of pure necessity.