Dan Lyons has a sobering outlook on the future of the American IT industry over at Newsweek.  The article argues that the Government’s focus on bailing out weaker industries is occurring at the expense of relatively strong industries such as IT.  Dan argues that a lack of investment in maintaining domestic IT talent could have serious implications in the future.  Dan writes:

“…unless we boost government spending on science, technology, engineering and math—STEM, in industry jargon—we will be unable to keep up with countries like China and India. At some point, companies such as Apple, Cisco, HP, IBM, Microsoft and Oracle could be eclipsed by foreign rivals, just as Ford, General Motors and Chrysler have been.”

Could American IT vendors face the same fate as American auto makers?  I’m not sure that the analogy holds.  American auto makers all but ignored the competitive threat from foreign rivals.  On the other hand, American IT vendors have made significant investments in growth geographies in order to establish a foothold with customers in these markets and take advantage of local skills.   One could argue that the investments in India and China by large American IT vendors will help create future competition to American IT vendors.  By that logic, Google is creating its future competition by hiring and creating teams of  bright programmers.  The history of employees leaving IT firms to start competitive or adjacent ventures is long and storied.

To the broader question about training STEM students, I absolutely think it’s important.  However, can America expect to graduate more STEM students than India or China in the long run?  So the real question is how can American IT vendors best leverage local and international skills? And in doing so, can America’s IT vendors retain their leadership position?  To the first question, I’d argue that this is exactly what American IT vendors are doing today with their hiring in growth geographies.  The answer to the second question remains much more open.

What do you think?

ComputerWorld has a very interesting article that even as Lehman Brothers was heading towards bankruptcy it invested $309M on technology and communications in the quarter ended August 31.  This figure represents a 9.5% year to year increase, which is down form the 18% increase, to $1.145 billion, in IT costs for the fully year 2007 vs. 2006.

It seems that Lehman was involved in selling some of its technology, specifically around a high-speed trading platform for equities called Baikal, to a number of investment banks and brokers.

Ralph Silva, a senior analyst at financial services advisory firm Tower Group Inc. is quoted:

“The units of Lehman currently selling technology to other banks are likely to be sold off

All banks are concerned about the ramifications of losing technology, and want more technology to be in-house,” he said. “So in this kind of situation, they tend to negotiate to get hold of the code or the entire systems. In Lehman’s case, the phone calls are probably already happening.”

It’s interesting to ponder whether the third parties using Baikal would be better off if the technology was open sourced?  Probably, but who, amongst the 100s of customers, would continue its development?

Does any customer want to take on the effort of developing a trading system in-house?  There is very little competitive benefit gained from the core of a trading system.

Next, consider the scenario of an independent company taking the open source code and offering support and future development.  Or worse, if multiple companies began to fork the code and offer support and future development.  As a current Lehman technology customer, I would be weary of having to select between 2, 4, or x number of forks.

On the other hand, there is also a risk if Lehman sells it proprietary technology to a company that then decides to go in a different direction than Lehman’s technology customers had expected of Lehman. But this is what source code escrow is for I guess.

I’m not sure that there is an easy answer here.  Is there ever?