I've been on holiday for a week, during which I read another book on the rising economic power of what's become known as Chindia. China and India no doubt have the potential to become economic superpowers, but for very different reasons--the former primarily sets its sights on low- and mid-end manufacturing, and the latter focuses on mid- to high-end services.

This is a reminder of why India may not be such a hot place for a semiconductor plant, despite a recently announced policy to sweeten the pot for manufacturers such as Intel Corp.

If you're involved in fab operations, when you think of India you think of infrastructure--the glaring lack of it. A fab requires a complex web of infrastructure support, including chemicals, gases, clean water and reliable electricity, as well as nearby test and packaging support.

"They are a lot further behind than they think in getting that infrastructure up and running. It will be an uphill battle," said Shahin Sharifzadeh, executive vice president of manufacturing and technology at Cypress Semiconductor Corp., during a recent conversation. Cypress has a huge design operation in India.

It's certainly possible to get a fab going if you throw in enough cash and other government incentives. But does India really need Intel? Policy planners in India should consider return on investment, and in that formula, small and medium-size fabless companies usually win.

Even some observers in China privately wonder if its drive for fabs is worth the cost. Those concerns have tempered interest in at least one 300-mm wafer fab--though I suspect many more will be built here. In China, big is beautiful, and spending on fixed infrastructure accounts for about half of its GDP growth.

Moreover, India's strength across many industries is design engineering, not large-scale manufacturing. Sharifzadeh even argues that India's advanced-engineering resources are more plentiful than China's, making it much easier to find a competent design engineer. That has led to just about every major chip company (and many minor ones) having a design outfit there, especially for embedded software.

China has long been known as a country that can mobilize huge amounts of capital and cheap labor. That helped it flourish as a maker of low-cost, low-tech products. This equation is slowly changing as high tech takes hold.

India, however, started out by targeting high-end services and products, such as advanced chip design and software development. This is why India, in the long run, may be a greater threat to the U.S. economy than China is.

India should stick with the game plan: Accelerate the financing of fabless startups, and leave the fabs to somebody else.