Defining innovation
Wow. Judging from the response (in the blog comments, emails and phone calls I got) to the earlier post I think I hit a nerve. Nobody in semiconductors or EDA likes to be told they are not innovating. But the VCs that responded, as well as some people in EDA, where generally in agreement with me. I think it may come down to the definition of the term.
I did some looking and found a blog that actually puts it in a way that meets my definition. It's called Broken Bulbs. Simply put, it is the "the profitable implementation of ideas."
That's where I think the semi industry is missing it, because they have not been profitable in a long time. The reason it isn't profitable, from the top end custom chip all the way down to the EDA tool, is because they are not concentrating on innovation. They are patching holes and it may be time just to throw the cloth away.
But let's look at the profitability issue. Let's look at the big daddy, Intel. Since 2000, Intel's stock is down 50 points and there hasn't been a stock split in over a decade. If you want to make a growth investment, Intel is not the place you want to do it. Over the same time frame, Apple's stock has split twice and is not multiplied in value 6 times over 2000 prices. Google hasn't done as well as Intel over the past decade. It's value has dropped rather precipitously and not its valued only 300x that of Intel. So strictly from perception, Intel is not seen as a very innovative company -- except by people in the semiconductor industry.
Semiconductor profitability as an industry is starting to rise again, with the economy. I'm sure EDA will also see a rise of revenue. But there are many companies that have not just survived, but thrived during the economic downturn. They did it through innovation. That's a catch phrase even in the semi industry. good companies innovate their way out of bad markets. The semi industry did not. They just held the line.
Here's another way of looking at it: My partner in Austin, Joe Basques, said the other day. If you don't know what the problem is, you can't innovate. The semi industry is dealing with lots of problems, but the haven't figured out THE problem. They are a sales driven industry -- the entire industry-- not market driven. So every time a customer bitches about a glitch, they throw their 30 percent R&D budget at the glitch, hoping to keep the customer happy... until the next glitch hits.
In the meantime, when they get approached with a revolutionary way of doing something--in other words, something that solves THE problem -- They kick it to the curb because it might kill their core product line.
Intel use to be good at solving THE problem. They would just, as a matter of culture, obsolete the previous product. They have been trying to do this for a while, but they haven't figured out what THE problem is to make that possible. ARM, on the other hand, has been doing pretty good by having systems designers (their customers) figure out how to innovate with their platform. But ARM isn't the innovator. their customers are.
There is a large semi firm (I mean like one of the top 5) that has been talking with an innovative company about what could be defined as THE problem in the semi company's world. The little company has the answer. All the engineers say the innovator has the answer. The bean counters say that's the answer. But for two years, they bean counters and engineers have been hunkered down trying to reverse engineer the answer so they don't have to pay for it. They estimate the 2-year wait has cost them $1B in sales. But still they haggle over the price of the answer, that would be 0.0001 percent of that loss.
There is a large semi-equipment company that developed a new technology for testing semiconductors. It would reduce the cost of test by a factor of 10. And when I say cost, I mean reduction of manpower, slashing the cost of equipment and making the entire process more efficient. There were customers who had purchased the first generation of the product and are still using it. But the company realized that the new tech would kill their primary product line that was, essentially, 20 years old. So they killed the project, destroyed the remaining stock and have locked the patents away.
I have a bunch more examples that essentially would make the semi industry profitable, and I mean like 50 percent increases in revenue, in two years. But they won't do it. They are too comfortable with the way things are.
That's not a culture of innovation. That's survival mode in my definition.
Lou,
ReplyDeleteI see good points from you and your readers. May I offer my own.
First, I get it that you are referring to everyone connected with the design, manufacture and test of semiconductor devices.
Second, within this superset, I do see innovation, however, it's focused on the single most important goal of the semi industry - efficiency.
The mantra of Moore's Law calls for a 2x efficiency improvement with every technology generation, and that includes design tools and fabs as well as density and performance.
Incredibly, the industry has managed to innovate to this mantra with great success for 50 years. But it's an inwardly focused kind of innovation, similar to what you see in an auto assembly plant.
Clearly, it's a rare (but fascinating) article or video that covers auto manufacturing efficiency, and that holds for semi as well.
I don't want to pull your discussion in a different direction, but I would observe that this manufacturing-focused efficiency is starting to crap out.
In fact I'm willing to say that whatever technology development gets us to 16nm and below with reasonable yields, will likely be the most significant innovation in 20 years. Will it be new? Maybe not.
It may have been around for a while, but the industry just didn't need it yet. That's typical of the semi market. There's a lot of great ideas just waiting to be used.
Can these ideas be used to innovate the semi industry out of its self-imposed box? Perhaps, but only if they are disguised as improvements in efficiency. That's how focused this industry is.
It takes a Cool Hand Luke type of commitment to escape this market. It's been done, but only by a precious few, and most don't even know they are out yet.
Actually, Kevin, now we are on the same page... I think. Everything you described in your eighth graf is what I am talking about. The innovation comes from the designer of the system, not the designer of the chip. The chip is just a commodity in the process.
ReplyDeleteI agree that the invention of the semiconductor was a great innovation. The application of the semiconductor into an SoC was an innovation. The reprogrammability of the FPGA was an innovation. But there hasn't been an innovative step forward in semiconductors or EDA in 10 years. The entire industry is lagging behind the rest of all their markets, when 20 years ago, it was driving them.
And let's not focus only on the VCs. Wall Street is not that thrilled with semi companies. The government isn't much interested in keeping them in the country. College students prefer careers in software design than electronic.
Maybe the problem is all perceptual. I've been courted out of the semi industry by nanotech, green, social media, medical tech, non-profits and transportation sectors. The semi world has been, for more than a decade, satisfied with the internal knowledge that they are the be all and end all of the world and they don't care if people pay attention to them... until it comes time for an exit strategy or a stock price. Then they complain that they get no respect.
I spent a lot of time and effort trying to help them with that problem, but they don't value that effort and think it should be donated to them. You and I both know that to be true.
Deep down, I really want to see the industry go back to being in a leadership position and creating technology worthy of real attention. I just don't see it happening in my lifetime.
Also, I don't think you can point to off shoring and downsizing as profitable exercises, because through all of that, no company has become any more profitable then they were before. In fact, many companies are returning operations to their home countries because offshoring has proven to be more costly, especially in the area of IP theft. The reason companies went offshore in the first place was to stop the bleeding... which I blame on ineffective efforts to innovate.
ReplyDeleteWe seem to think that Moore's Law is a force of nature and must be adhered to, when it was only supposed to be a guideline. By forcing adherence we've blown through the problems of power consumption and loss, cost of production and basic laws of physics that should be dealt with before moving to a smaller node. We've buried small, innovative companies in the name of maintaining market share and the status quo.
I take a look at a big company, like SAP, that is investing in new companies and spinning off divisions to keep them from being burdened by corporate ennui and as a result, I see them growing, evolving and innovating. But I look at semi and EDA companies sitting on truckloads of cash, and denying engineers the opportunity to solve some of the basic problems and, yes, innovate.