Yea, though I walk

I had a funny realization today about the state of hardware technology startups.  I was reading through this nice summarization of DRAM technology and came across a link to "quadbandmemory.com".  The technology purportedly interleaves DRAM memory banks transparently to the memory controller. I looked at the date next to the link... which harkened back to the year 2002.  I thought, what are the odds that this link works at all?  I clicked and a website with asian characters comes up.

This might in fact be the best case for such a hardware technology company, since they were at least able to sell the domain before going under.  Researching a bit more, it looks like they may have gotten some licensing revenue before going under, even more good news.  Internet Wayback Machine shows they lost the domain some time between April and October 2005, and their last news update was about something in 2004.

Most surprising to me was that I was 99% sure that the company would not still have its domain 11 years later, and I was correct.  In fact, it did not even own its domain 3 years later.  I'm naturally an optimistic person, but what's more startling to me than my doomsaying, and the level of certainty that I have in it, is how accurate it is :-(

Chip Startups are Dead

Turns out that chip startups can't get funding.  Tilera is trying to turn into a company making fancy NICs. Adapteva had to turn to Kickstarter to raise a measly $900k (fortunately they were aiming for 750k so got to keep it) which has relegated their design to the education market.  That market focus was tried by SiCortex and led it to become the flaming pile of dead hope we all really needed to pick up our spirits during the first year of the great recession.

If you're like me, and still fighting to build a high performance computer chip, your best bet might be one of the countries trying to build its own Silicon Valley.