There are actually a number of reasons everyone's going to CMOS.
One is that yes, pound for pound, a CMOS sensor is cheaper to make. These days, there are other advantages. Largely because the CMOS processes are so well used and understood, the technology of the CMOS chip has been moving much faster.
Today, a CMOS sensor is quieter than a CCD at the same chip size. Part of this is the fact that the CCD is an analog chip -- each sensor site accumulates a charge, which is converted to a voltage, then bucket-brigaded off-chip.. the CCD is a big analog shift register. On a CMOS chip, each sensor site has its own charge to voltage conversion, and the ADC will be very close, possibly at each sensor site. Add to this recent innovations like backlit sensors (the sensor site is one side of the chip, all support circuity on the other side, and this is another great reason everyone's going CMOS.
Power is another issue on-camera. A CCD will consume as much as 100x the power of a similar CMOS chip. It's not just the power, either, but the heat from that power -- modern HDSLRs would be virtually impossible to create using CCDs. CMOS chips can also be made much faster, since any number of sensors can be processed in parallel. It's even possible to use a full frame shutter with a CMOS device, which would eliminate the "jello" effect, but no one's doing that right now, at least on affordable cameras.
Given the lower noise, a CMOS camera will show lower background noise for the same gain, and better color: CMOS chips almost universally produce a 14-bit output, while CCDs are usually limited to 12-bit. That's not to suggest the HMC40 or TM700 will produce a better low-light image than the HMC150 -- the HMC150's larger sensors win here. But compare the HMC150 to a modern CMOS 1/3" camcorder like the Canon XF300, and you'll see where CMOS can really work.
Another thing you don't get on CMOS sensors is vertical smear. Some higher end CCDs can reduce (though not eliminate) this as well. This is due to the bleeding of charge from cell to cell in the CCD, and can be minimized by using a dual charge bucket design -- one bucket gets exposed, the charge is coupled to another charge bucket, and that second one is where the charge to voltage and shift register operations occur... while the first is de-saturated. This takes twice the charge coupled array, so it adds cost. This is called an interline CCD.