July 15, 2004 – IBM’s director of physical sciences, Thomas Theis, discusses nano, disruptive technology and why physical research is a driving force at Big Blue.
He met with Small Times’ David Forman over lunch at the T.J. Watson Research Center in Yorktown Heights, N.Y., to discuss everything from next-generation magnetic memory (MRAM) to Millipede, IBM’s cutting-edge approach to data storage.
Despite the recently publicized foray into services, IBM is still doing what it does best: using technology to expand its existing markets and create new ones.
IBM recently explained that its new focus would be services. What does that mean for you?
What it means is our world is changing. We’re certainly an important part of what (physical research) goes on, but we are no longer necessarily seen as the driver for everything.
Twenty-five years ago, when I came here, hardware was completely dominant. That meant people in physical sciences. It meant people doing silicon technology and developing magnetic storage and all kinds of things.
I’ll just be a little bit provocative and say that the whole thrust of nanotechnology is to continue that trend by continuing to make ever more complex hardware systems, including IT systems, possible at ever-reducing costs.
Is nanotechnology therefore a big part of IBM’s strategy to stay away from what IBM Chief Executive Officer Sam Palmisano called commodity hell?
We want to be an innovator, yes. Maybe half of our total investment in what people would call nano is aimed at extending our established businesses, and that’s all about technology leadership. When Palmisano talks about that, he’s talking about services that aren’t just the same commodity services that everybody else is offering.
Are you saying the idea is to map IBM’s approach to technology onto services?
Yes, yes it is. That won’t change. Our business model won’t change just because we’re going into services more than we have in the past.
It’s sort of in our DNA, so to speak, that we would ask what are the parts of a business where we can provide a lot of value-add, something unique. And in fact, Sam Palmisano very much expects that to be based on the fact that IBM has this world-famous research division.
I view this move into services and software as potentially a huge opportunity for new markets in hardware. Look at any of the new IT systems that people are trying to implement. I’ll give you two examples.
One would be RFID tags. Another would be distributed sensors. We’re already talking to the companies. We’re already saying what does IBM have, how can we partner with you to implement that.
We can then go back to what we’re doing in nano, in the hardware arena, and say does any of what we have, any of the invention that we have, any of the intellectual property we’ve generated, tie into and help us with these new types of businesses that we’re trying to get in at the software and the services level.
So that becomes a services relationship because it’s consulting to begin with and then building a solution together?
Yes, but that’s not what dominates us. What dominates us is we continue to extend our established businesses in IT hardware and we continue to search for new methods or devices that will allow us to process, store or communicate information.
Millipede is an example. The way to look at Millipede is it’s an experimental systems-level engineering project. It’s neat to look at this nano part of the thing where we’re making little indentations in a polymer with what is the world’s most complex micromechanical system.
But none of that’s of much value unless you develop the multiplexed electronics that can address individually each of those tips, unless you have software people working on the coding of the data to make it error resistant. We’re working on all of those things.
Do these types of technologies — where you have speed of memory and the data retention of a hard drive — give you the ability to look at restructuring how we do computing?
MRAM is the one that has the possibility of giving you random access memory speed and nonvolatility (holding information without electricity). Millipede would still have some access time like a hard disk.
I think the way to think about Millipede is that it’s a much smaller form factor enabled by the fact that we’re storing information at an order of magnitude with more density than what we can get with a hard disk drive.
We’re working on more than one nonvolatile solid-state memory. These are all of course nano devices now. It’s exactly that idea. When you look at something that has the property that you can read it very fast and it can store the information in a nonvolatile way, it does allow you to re-architect.
It rewrites history a little bit in terms of what you can do.
Absolutely. And, in fact, people have a hard time thinking how it will work. The insights are sort of trickling in. There’s got to be some justification for the project, but these things tend to go toward things that have a chance to be revolutionary.
But you don’t understand all the advantages until people have had quite a bit of time to think about it.
Is that process infused in the culture or is it something you try to do in a more formal sense?
It’s both, but we do try to formalize it. We get the best technical experts we can find and ask a set of questions that people agree would elicit some insight: Where are we going? How’s the technology developing? If it develops this way, what would that enable?
And that comes back and informs the research?
Right. And we also ask what’s the business case? Suppose we had those system-level properties. How big is the market? How much do we want to invest?
It’s a circle. Technology drives thoughts about investment in the business and then that circles back.
At what stage is the Millipede project?
I think we’ll know enough about Millipede that we’ll be ready to make a business decision sometime this year.
That means do we invest in it as an IBM-funded project or do we partner with somebody? Or do we say we’re not interested in the markets we see because we’ve done these business cases and we’re going to license the technology?
What are the other nanoscale applications under research?
The things we’re looking at — mostly carbon nanotubes — are at a much more exploratory phase. Carbon nanotube transistors are very promising as devices but we still need to do further research just to confirm that it continues to be a potential competitor for the mainstream silicon device in terms of performance.
But for it to happen, a tremendous amount of invention has to take place that would be addressed to manufacturability. And we’re just starting to ramp up our research effort in that area.
Many of the early announcements clearly said the work was not done with any kind of scalable technique.
We’re just starting to work on that. Some of the research we’re doing with carbon nanotubes is meant to identify early niche applications even if they’re not applications IBM is likely to manufacture.
We would like to be a part of establishing some new niche market for that technology.
That’s the way many new things get started. In fact, most really new things have to find that niche where they don’t have to compete with the big established $100 billion-a-year technology to get started.
Patenting activity and strong intellectual property would seem to play a very pivotal role in that.
You know there’s a very interesting opinion piece that was just published by Nick Donofrio (IBM’s senior vice president of technology and manufacturing). I’m paraphrasing, but basically what he’s saying is that, yes, we get the most patents of any company, but that’s not what matters.
What matters is taking that stuff and innovating in the marketplace. Having an idea doesn’t make a new product that a customer actually will buy.
So not just a better mousetrap, but also a completely different type of trap?
Yes. Google’s an innovation that changed everybody’s life a little bit. But the hashing of data tables, which is what Google is based on, and all the other search services for the Internet, is an invention or a concept.
You have to do the basic research to have the base for that innovation.
So nanotech from this perspective is sort of incidental as a category? It happens to be that that’s where a lot of the enabling technologies for future innovations seem to be.
It’s the basis of the whole pyramid.
Because it’s at the cutting edge of these technologies that have disruptive potential?
Yes. Right. Because for a hundred years going back to mechanical tabulators, information technology has been about smaller, faster and cheaper.
All of the advances there at the device level, at the base level, are at the nanoscale. I would take an hour just to take you through what we do with the silicon transistor. That’s a nano device.
It’s an evolutionary path that’s fueled by invention. Evolutionary doesn’t mean easy. It means a lot of invention, a lot of failure. We see things like carbon nanotubes, Millipede, some of these non-volatile memory technologies, optical communications.
We’re going to publish stuff on optical communications this year that is going to make the world look.
That’s based on nano?
Very small optical communications devices with some nanoscale components, yes. Very small optical switches. People call it nanophotonics.
We’re already known in the scientific field and I think people can see that we’re doing very good work but we’re going to publish some real eye-opening stuff this year but I can’t talk to you about it. Nano is every advance, that’s what I’m saying.
How do you take groups working on competing technologies and work that competition to your benefit?
If it’s a small project, the competition is there on purpose because we don’t want to make theoretical guesses about what the advantages and disadvantages of various approaches are.
We want to have some laboratory data. So it’s the management’s responsibility to sort out which of them are so promising that they deserve long-term investment.
And then there’s that funnel of ever-increasing funding that you need to get something into the market. I won’t talk to you about individual companies but I would tell you this: All of the molecular memory companies have the problem that to get from what they have at a laboratory level to manufacturing a reliable product is going to cost them on the order of $50 to $60 million. That’s what it takes to develop a new memory.
There’s a lot of talk lately about how somebody can come along with a good enough solution, something that’s cheaper, kind of under the radar, and unseat a market leader. How does preventing that from happening factor into your strategy?
Well, it did happen to this company in the late 80s and early 90s (with the personal computer).
So there’s your case study?
We did not fully understand the implications, not really until (former CEO) Lou Gerstner came on board. That was when he got the company to say OK, let’s put those old models, the way we used to make a lot of money, let’s put that aside.
Let’s figure out how to make money in this new world that we’ve helped create but that we didn’t understand very well.
So yeah, it happened. I lived through that and hope it doesn’t happen again. We’re all conscious of it. We always are looking for things that might be disruptive to our business.
So innovate and let the bottom feeders bottom feed?
Yes. That’s basically what Palmisano was saying in that public statement. Don’t quote me as saying it that way (laughing). You said it.
What is IBM’s, and specifically, your role, bringing to the formulation of national strategy and legislation?
They look to us as a center for expertise. Actually it was Tom Kalil, who was in President Clinton’s science adviser’s office, who came to us. We worked with him to provide material, the literature that was generated to educate Congress.
Since then I’ve been involved in several NSF (National Science Foundation) events. They seem to like me to talk about societal implications of nano. So I’ve done a couple of those. I feel I must do those. I also do outreach stuff on that same subject.
I’m going to speak to what I hope is a fairly large group of high school teachers from this area and just try and help them to distinguish between the science fiction version of (nanotechnology visionary and author) Eric Drexler’s and what scientists are really doing in laboratories right now.
Is there anything you think was lacking from the legislation?
Not a lot. I am concerned about the fact that this is cross agency and we try to put a lot of our recommendations in as to how to strengthen that so that agencies don’t get into ignoring each other or competing with each other.
I’m waiting to see how that works out. I don’t have something that I would have put into the legislation now to avoid it or fix it.
We specifically recommended that some of this money go to fund research on societal consequences. I think the immediate problems are just that people are worried about these new materials.
People like me look at it and say there are new materials, tens of thousands of new chemicals synthesized every year. Some of them get into the manufacturing stream. We have OSHA (Occupational Safety and Health Administration) and EPA (Environmental Protection Agency).
They ought to be doing that, but I don’t personally see any nanomaterials that I’m aware of as fundamentally changing that picture.