In the second in our series of investigations into the lesser known R&D activities of beastly old Intel, we meet the man behind a revolutionary new bioelectric diagnostic chip.
Medical diagnostics is a pretty time consuming and expensive endeavour. Testing bodily fluids for various markers of ailments and disease, spooling up those high tech scanners... it all adds to the cost of health care. It's a cost that is becoming increasingly crippling in developed nations.
But what if there was a technology that was not only massively cheaper, but also much faster, as well as more sensitive and more capable? Well, there is.
At least that's what Ilan Levy, one of Intel's big brains at its research centre in Israel (yup, that'll be the same Israeli outfit that saved Intel's bacon with first, the Pentium M, and then, Core 2 CPU architectures).
The basic idea is simple enough to grasp. To use Intel's peerless silicon chip expertise to mass produce a computer chip festooned with diagnostic sensors. "We have developed a single-die chip with 148 different sensors capable of multiple levels of analysis," Levy explains.
A chip thats cheap
Thanks to the use of cutting-edge silicon technology, the final production chip is likely to be very small, and hence extremely cheap. That in turn should allow it to be integrated into a low cost, disposable, single-use cartridge that plugs into a larger reusable device.
For this usage model, bodily fluids are passed over the chip and the resulting signal or data is wirelessly sent to a control system. Simply replace the cartridge and repeat for each test subject.
In theory, the single chip Intel has developed is capable of performing any bodily-fluid based diagnostic currently available. That includes tests that require cultures to be grown.
Levy explains: "The standard method for detecting some infectious agents is to grow cultures. Take hepatitis. For that, it's important to know how many viral particles are present in one cc of blood."
Currently, the medical industry uses a molecular detection method, which is based on the DNA composition of the virus. The idea is to amplify the amount of genetic material to detect it. "But with our chip, using an electrical detection method, it is possible to detect a single virus."
Deluge of data
What's more, apart from potentially replacing a long list of conventional medical tests with a cheaper and more sensitive alternative, the chip spews out a mountain of additional data that will only add to diagnostic accuracy and insight.
"The immediate interaction with the chip produces one-dimensional data, a peak or spike in the signal. But there are further dynamics in the fall off following that initial signal.
A huge amount of data can be captured. Quite frankly, nobody has a clue what to do with it right now. But eventually, it will be enormously valuable."
What's more, the bioelectric chip's capabilities extend beyond medical uses. "The food industry is a good example. Let's say you want to look for pesticide contamination in dairy food. You can do it with this chip. Or water and other types of environmental pollution. You can do it with this chip," Levy says.
Needless to say, the possibilities are incredibly exciting. There is, however, a catch. Bringing the chip to market will be a monumental task. Although Levy says the chip itself is essentially finished and ready for action, there's much work to be done on the supporting systems and software.
As Levy says, Intel is not an established player in this market and would need to either court or acquire a production partner outright to develop the accompanying cartridge, device and supporting software.
Blood, sweat and tears
Thanks to reliability concerns, that would be no mean feat. "In a way, it's a crazy idea, throwing electronics into an environment of bodily fluids - tears, sweat, blood, urine and semen," he says.
Then there are the regulatory barriers. The US is chip's likely biggest market. To bring it to market there requires FDA certification. That could be a serious headache. "Let's say we go to FDA tomorrow. It would take between 3-6 six years to get approval," Levy explains.
However, because of the way certification is regulated, the permissions would only apply to the complete system the chip was housed in, not the chip itself. Any change in the supporting device would require new certification.
It gets worse. "The big catch is that we are developing a platform that can simultaneously detect many biomarkers. But this system would need to be certified for each different diagnostic application."
Bung in years of field trials and accompanying research and you have a long, costly and risky process. Indeed, Levy is not convinced it's one that Intel has the stomach for.
"If you show the business groups the numbers," he says, "the market is just huge in terms of units. But what about the margins? How much silicon are you going to sell and at what margins?"
The disposable cartridge implementation, in other words, pretty much demands low pricing and tight margins on the chip itself. And if it is only the chip Intel is supplying, the business model suddenly looks a little less spectacular.
That said, it might just be a project like this that Intel needs to pursue. Certainly Intel's stock price has been largely unmoved by the roaring success of its core PC processor business in recent years.
If the world is tiring of the quest for the perfect PC processor, maybe a more diversified approach is what it will take to get investors back on Intel's side.