The End of the Semiconductor Industry?

Bryzek says he has been influenced by the book Abundance authored by Peter Diamandis and Steven Kotler. An underlying theme of Abundance is that exponentially advancing technologies will make it possible to meet the basic needs of every human on the planet. Bryzek and Diamandis are convinced networked sensors, or sensor swarms, are part of that vision. Bryzek is also convinced that a related idea called Exponential Organizations could spell the end of several industries, including the semiconductor industry as we now know it. We recently chatted with Bryzek about these ideas.

MD: A trillion sensors will end many of humanity’s problems? Where did that idea come from?

Bryzek: Actually it is 45 trillion. The Abundance movement forecasts a need for this many sensors in about 20 years, many of which are not yet developed. The problem is that development cycles that take sensors from prototypes in academic labs to volume production typically span 30 years based on the historical cycles for the first MEMS devices that reached volume production. If we let sensor development happen according to historical cycles, slow commercialization would slow the Abundance. The Trillion Sensor (TSensor) movement aims to accelerate the development cycle of new sensors.

For my part, I have been involved in seven start-up companies that have brought sensor technologies to market. Based on my experience, I created an acceleration strategy.  It includes inventing new ultrahigh-volume sensor-based applications (TApps) likely to become commercial in the next decade; Selecting sensortechnology platforms that can best serve the largest number of TApps; And then accelerating development of these platforms using serial entrepreneurs with commercialization scars who know, for example, what not to do in start-up mode. Other experienced entrepreneurs and I can help accelerate the Abundance by preventing companies from making the mistakes we’ve all made in sensor start-ups.

MD: Is it just sensor technology that will make Abundance possible?

Bryzek: No, Abundance relies on what are called exponential technologies, and sensors are just one of those. Others include biotech and bioinformatics, medicine, nanomaterials, 3D printing and infinite computing, robotics, computational systems, and artificial intelligence. It also makes use of the DIY revolution as embodied in the achievements of innovators such as Burt Rattan and Elon Musk. All in all, Abundance becomes possible as a result of multiple emerging global economic tides such as the Internet of Things (IoT) and Digital Health.

MD: How will the TSensor movement advance the idea of Abundance?

Bryzek: I decided to invite visionaries to “invent” new sensor applications likely to drive ultrahigh demands for sensors. And by ultrahigh we mean more than a billion units annually because that’s what it takes to make an impact globally. The movement is really just getting off the ground. We held TSensor summits last year at Stanford University and UC Berkeley, and we just finished one in Tokyo. Later this year, we’ll convene in Munich and San Diego and again in Tokyo. There has been unbelievable excitement at these meetings. People are saying they are like TEDs talks in real time.

The TSensor summits were focused on discovery of new ultrahigh-volume sensor applications. As a follow-up, we started developing the TSensor Roadmap. The summit at Stanford resulted in 14 TApps.  Our working group has 120 volunteers developing recommendations for sensor technology platforms supporting these TApps. For example, one of the TApps is for minimally invasive medical care. Another is for personal imaging.

MD: How do exponential organizations fit into the picture?

Bryzek: Exponential organizations create new business models that disrupt the global economy, often taking advantage of crowdsourcing, crowdfunding, and incentive competitions. The Singularity University, cofounded by Peter Diamandis, maps out what’s called a 6D evolution process for exponential organizations. First, products or services get digitized, leading to a deceptive phase where the new technologies don’t seem to be good enough to upset the status quo. This leads to a disruptive phase where the technologies improve and do, in fact, disrupt existing players. Next comes dematerialization where physical products or services become digitized and are distributed as bits on widely available platforms, analogous to when downloads replaced DVDs and CDs. This leads to demonetization: The dematerialized products can be freely distributed globally, mandating new business models to reward players. Finally, you have democratization, which leads to growth everywhere. The classic example is that of mobile phones in Africa. Expectations are there will be over a billion of them there in two years.

MD: Why do you say the semiconductor industry is going to be a victim or beneficiary of the 6D evolution?

Bryzek: Consider that research labs are 3D printing sensors right now. For example, VTT, the technical research center in Finland, is doing chemical sensors. The University of Massachusetts at Amherst is developing environmental sensors. And two years ago, the IMEC nanotechnology center in Belgium 3D-printed a microprocessor with 4,000 transistors able to function about as well as Intel’s first 4004 microprocessor in 1971. Of course, the cycle time of the printed processor is lousy compared to the state of the art. Commercial products today include RFID tags, with 1,000 printed transistors in 5-µm process node. Multiple R&D centers are developing 130 and 70-nm transistors. There are a lot of developments under the radar. The performance of these devices is nowhere near that of standard electronics so most semiconductor companies are ignoring the implications.

MD: Are you saying companies will move to 3D printing to make electronics, or that 3D printing will eliminate electronic component makers?

Bryzek: 3D printing of electronics is in the deceptive phase where you don't see major threats to the biggest players. In a few years, the whole discipline will pop. Indications are that printed transistors will be 1,000 times cheaper than those fabbed conventionally in silicon. So suddenly, you don't have to strive for supersmall devices to get costs down. The point is that if you can print transistors with adequate performance on a 3D printer in anyone’s factory, there is less need for semiconductor companies.

MD: When could this happen?

Bryzek: Peter Diamandis says several industries could undergo demonetization in next 10 years. By then, he says, 40% of all Fortune 500 companies will disappear, with similar impacts on the rest of the world.

When you look at Imec’s 4004 microprocessor, its performance lag is 40 years.  I expect in 10 years the lag could drop to 20 years. So in 2021, printed electronics could achieve the performance of 2001-vintage ICs, but at 1/1,000^th the cost-per-area. If exponential growth continues, the gap could shrink another 2× in 10 years, meaning 3D-printed ICs in 2031 would have the performance of silicon ICs made in 2021.

MD: How do companies make any money when this happens?

Bryzek: Exponential Organizations will be deploying new business models. We already see many new business models as, for example, when Google provided free information paid for by advertisers.

Sensors are expected to generate Big Data. Through sensor-based services coming to market, for example, companies may get more revenue from selling subscriptions for sensor data postprocessing than for the product itself. One handheld spectrometer coming to market this year measures food quality and its sale price includes a one-year subscription for cloud-based processing. If you want to know what’s in your food, you pay for the data. That model can be applied to many other things. Take shoes, for example. You might someday get your shoes for free but pay for an analysis of your health based on sensors in your shoes.

MD: Do businesses in general understand how this drastic change might affect them?

Bryzek: A lot of big bosses in big corporations don't appreciate that this can be real. They will miss the boat, as they have in the past. An example may be photography; When Kodak went bankrupt, Google acquired Instagram for a lot of money. Some companies see the forthcoming changes and position themselves to ride the tide. Cisco, for example thinks the IoT will grow to $19 trillion by 2020, comprising over 20% of global GDP. Amazon, Cisco, GE, and IBM have been modifying their networks to support IoT. Many other companies, however, don’t know yet how to spell IoT.

And overall, these changes are a good thing, not a bad thing. IoT is expected to be the biggest business opportunity in the history of humans on earth. There are estimates that in the U.S., 50% of all employees (more than 70 million) will be eliminated in the next decade by robots. This is scary. But growth of IoT by $19 trillion would create about 170 million new jobs. The major task will be retraining; Can we create exponential education systems capable of supporting IoT growth? It is a major question.