You should expect this trend to continue. Semiconductors have become a hideously expensive business, aggravated by fewer fabs and more draconian fab/IP acquisition costs. It's a well known that nobody on Sand Hill road will touch semiconductor startups at all, leading to a stagnant collection of semiconductor companies founded 20+ years ago. There was recently announced a semi-only incubator here in SV[1] because nobody can raise any traditional VC money, although time will tell how many successful exits it will have.

[1] http://www.eetimes.com/document.asp?doc_id=1326501

It's far easier to build a start-up doing software than pretty much any hardware, so current giants have a huge advantage over anyone trying to break in. Consider for a moment that a new fab with a recent process is probably a many billion dollar investment, and that even fabless semiconductor companies (a la LSI, recently) there are huge amounts of IP, capital, and brain power tied up; everything also really needs to be really correct up front, and tooling is expensive (you wouldn't want to waste millions of dollars building a chip that's really unproven and undesired from a market perspective). Even technologies with less expensive (though not at all cheap) tooling, like hard disk drives, the IP is so tied up with the current giants (WD/HGST/Seagate), and the firmware needed to make the drives work at current capacities is so complex, that it is essentially impossible to break into the market.

The only notable recent entrant I can think of is Spreadtrum from China. But I don't know that it counts as a startup when it is backed and partially owned by the Chinese government.

Capitalism at work. Collusion, consolidation and no competition.