First off, let me say at the outset of this column that I’m a fan of intelligent globalization. My understanding of what turned the 1929 market crash into a depression was protectionism under Smoot-Hawley. That point has been argued back and forth repeatedly, and I weigh in on the side that it was bad.

Now that I’ve made clear my vantage point, I’m hearing a lot of complaints from engineers these days that too many jobs are disappearing to Asia and that companies are only using H1-B visas to cut more expensive jobs and replace them with lower-paying jobs.

Let’s take a look at several scenarios—and I welcome your comments because I don’t pretend to have all the answers—and examine some consequences.

·      The U.S. government does nothing. The problem is that total inaction on the part of the federal government has been shown to be disastrous. In 1837, and again in 1839, president Martin Van Buren sided with the camp that no action by the federal government to control the successive panics was the right approach. Van Buren became known as Martin Van Ruin and the downturn lasted until 1845, which was the longest downturn in the history of the United States. What eventually brought the U.S. out of the Great Depression in 1929—a misnomer in comparison to the Panic of 1839—was federal government involvement and debt financing. So the government has to do something—but what?

·      The government can create protectionist barriers. That reduces overall sales while increasing inflation because the price of manufacturing domestically is higher than internationally. Offering subsidies and rebates is the same thing in reverse, which virtually every major economy does for industries it believes it needs to be strong. China did that with its foundries and design houses; the United States does that for solar energy companies and everything Green. Canada does it for farmers. Subsidies are much harder to measure than tariffs, however, and often fall under the radar of market watchers. But in real terms they’re basically the same thing. The problem for electronics companies in the United States is they don’t get subsidies, while in other countries they frequently do—sometimes in the way of cheap land or utilities or low taxes for 50 years. In a global economy, the field needs to be level. Otherwise, it needs to be explained loudly and clearly that you can’t do business with a country until they clean up their act.

·      H1-B visas are a blessing and a curse, and the blame for the latter can be split among lots of different groups. Companies are complaining they’re not getting enough expertise in the United States, so they have to either hire more foreign students or go overseas. There is some truth to that. Part of it is also disingenuous. If engineering and science were billed as the only way to solve global warming and save the planet—and they are, short of everyone giving up their cars, televisions, cell phones, etc.—then we’d have no shortage of students pounding on the doors of universities. This is a major shortcoming of both universities and teachers in k-12, and it’s something that needs to be driven by electronics companies and politicians. At the same time, if we extended more visas for foreign students, we’d have plenty of them sticking around and adding something to our melting pot of ideas—and enough to tide us over until the numbers rise in the schools.

Each of these problems is more complicated than anyone can comprehend, and there are many sides to each of them. But tough times demand well-understood solutions, and knee-jerk reactions are bad for business in the short term and the long term.

What do you think?

–Ed Sperling

The wall is in sight.

Moore’s Law has propelled the semiconductor industry at an amazing velocity since it was first introduced in 1965, and despite some minor changes from 18 months to two years, we have pretty much stayed on course. In the past, most people thought we would hit the wall at 1 micron, and they thought it would happen again at 32nm. The road map appears pretty solid down well beyond that.

But at 22nm—or the half node of 20nm—things seem to get a bit fuzzier than in the past. You can’t just fix one problem anymore. You have to fix a bunch of them simultaneously. And those problems become thornier as you progress down to 10nm. Even the fastest ASICs and processors will begin to look more like systems designs than chips, and designs will expand well beyond the physical limits of the silicon to include software, other components on a board and the manufacturing processes to create them.

This used to be so far in the future that no one really gave it much thought. It was something you talked about over a beer. But with companies now working on 32 nanometer IP blocks and manufacturing processes, it won’t be that long before we start seeing the wall. There will be ways around the wall, of course, but the path will hardly be a straight line.

At the very basic level, lithography technology will have to change. The move to extreme ultraviolet lithography has been talked about for years. TSMC already has committed publicly to immersion lithography. But which way the industry ultimately heads is the subject of lots of research at the moment. So far, there are no clear answers. Both technologies are subject to defects, and while those defects may not be significant at 180nm, they will ruin a chip at 20nm.

On top of that, new transistor designs will be needed. Chip designs already have started going vertical. Memory makers are using stacked die to create their chips. But we’re also starting to see the need for new transistor designs such as FinFETS, which have 3D fins resembling a 1958 Cadillac.

Add to that such technologies as air gap, new materials and substrates such as silicon-on-insulator, and suddenly the wall begins to take shape. From a distance it looks opaque, but up close it’s porous. The only problem is that getting through it requires a huge investment in new technology.

Moore’s Law originally was created as an economic statement of manufacturing economies of scale. The further down the road map, the more the economies of scale begin rolling out in reverse. Many companies have been wondering where the tipping point will be for Moore’s Law, and it varies by company. But the end of the road may be when the last company no longer gets a benefit from putting more transistors on a piece of silicon—no matter what shape they take or how exotic the substrate.

–Ed Sperling

Intel’s decision to invest $7 billion in 32nm fabs is possibly the best example yet that in spite of a drop in projected sales, technology companies need to invest in the future. We live in a cyclical business, and cycles begin and end.

Intel is better positioned than most semiconductor companies. It has the largest R&D budget—it’s hard to think of IBM as a chip company, even though it has a larger R&D budget. It also has been sitting on a pile of cash, and it knows exactly what’s needed to push down to the next process nodes because it’s the only processor maker that still relies exclusively on its own fabs to develop a digital manufacturing process.

Intel also knows that if it doesn’t jump on the next process node—it’s already doing core research at 10 nanometers—then someone else might steal its market. That’s Intel-think for paranoia. The company doesn’t really have any rivals left in its core markets. But it does have plenty of new market it’s looking to tap—everything from replacing microcontrollers in automotive and industrial applications to embedded processors, where it will plant future, extremely low power versions of its Atom processor.

But Intel is the start of the cycle. As the largest chip company, when it cuts back on spending the industry feels the pinch. It’s not just a matter of a cutback in consumer spending. People still need computers, whether they’re in data centers or smart phones. Ultimately the companies that produce chips at advanced process nodes will need capital equipment to build them. And they will generate sales throughout the electronics ecosystem that will drive another period of boom and bust.

Some booms will be larger than others. Some downturns will be deeper, like the current one. But if history is any indication, once the machinery gets going again it will run for years. And companies that don’t invest along the way in new tools, new equipment and new technologies could find themselves sitting on the sidelines when the next boom cycle begins.

Convergence and complexity at the design level are creating communications problems, something that became particularly evident during a couple panels at DesignCon this week.

In one panel, which was focused on design for manufacturing, the crux of the problem was trust and expertise. On the trust side, the issues that have been talked about for years of foundries sharing data with fabless companies, and vice versa, continues to persist. The information exchange has gotten better, but as Intel will tell you with great smugness, having a fab is a competitive advantage because you can tweak designs and processes simultaneously without these kinds of issues. Not owning one—and the cost is prohibitive for most companies—means you have to trust someone else with your secret sauce.

The foundries don’t want to release too much information because many of their larger customers use the services of more than one foundry. That distrust carries over from the fabless companies, too, when they don’t think they’re getting enough information. At times, there simply isn’t enough because processes are too untested at the leading edge. But when exactly does that information become available, and is everything shared along the way? There are even questions about whether it should be.

In another panel, focused on globalization, there was confusion about what is outsourcing, offshoring and globalization. That’s the easy stuff, too. Trying to communicate across cultural and language barriers is an enormous challenge, and it will continue to become even harder as new countries begin entering the market. China and India were first, but Eastern Europe, Vietnam, the Philippines, parts of northern Africa and portions of Latin America will join this fray soon enough.

Perhaps even more daunting on the system-level side is getting hardware and software engineers in the same room and working on the same design team. Things that seem logical to one group seem impossible to the other.

We are at the beginning of this transition across language issues. It will only get better, or worse, from here. At the very least, you can count on it getting more interesting.

–Ed Sperling