It's a good bet your company doesn't have an FPGA strategy. One of the most challenging of electronic components, a field programmable gate array is the logical equivalent of a huge box of Lego bricks.
Add configuration data and it can become any sort of digital circuit, including a custom processor designed for one task only. It's a bit like having your own chip fabrication plant bolted to a circuit board — all you have to do is programme it, and it becomes your very own hardware.
Could it do business IT better than plain old Intel hardware? Yes. Has anyone worked out how to make that happen? Not yet.
But up in Scotland, co-operation between SMEs and academia at the Edinburgh Parallel Computing Centre (EPCC) has produced an FPGA supercomputer showing the sort of figures that Intel or IBM would be proud to call their own. It's run some traditional high-performance computing applications. It's probably not ready to take over your server farm — but if SMEs can afford to develop and produce exciting new technology, other companies can afford to spend a little time checking it out for competitive advantage.
Or take the way Google evolved its data centres. There's a standard way to build very big, high-performance server farms: you go to the companies that sell very big, high-performance servers and put very big, high-performance software on them. Google didn't do that. Google found the cheapest PCs it could, wrote special software to mesh them together and cope with the inevitable failures. Could you do that? Probably not — but if you don't think about it, you may be missing out.
Every part of mainstream IT was once new, different, untested. Every part had to fight for acceptance against established, well-funded antecedents which were "the way everyone does it" — and the people who refused to be limited by conventional wisdom got the greatest rewards. Not everything new works well, of course: going with novelty for the sake of it is for toddlers, not IT managers. But a company that always takes the word of its existing suppliers, that doesn't make the effort to find out what's happening at the edges, is a company that will miss out. Invention need not be expensive and the rewards can be substantial: if you don't look, you'll never know.