From what I've seen the power consumption figures listed by Intel seem a bit optimisitc. On the other hand the claim that it uses less power seems defensible, since the claimed power consumption of hard drives are optimistic as well.
Classic MHDDs use around 450-500mA of power while they idle, and more during I/O operations.
However, many MHDDs draw nearly 2A of power during spin-up, which is pretty hefty. You can read a white paper on MHDD power analysis if you want. Voltage-wise, MHDDs use both the 5V and 12V pins on their SATA power connector (or on their 4-pin Molex if the drive is an older model or PATA), while SSDs tend to just use the 5V line.
Regarding the SSD, since you're using it as a comparative model: there's no spin-up (no spindle motor, no platters), so drive power is more or less consistent from the moment of power-on. However, you're forgetting something: a SATA-to-USB3.0 adapter would require a SATA-to-USB conversion ASIC (basically a processor). Are you aware of how much power that ASIC can draw? Chances are that is what's responsible for most of the power draw.
Let's also not forget that with the introduction of USB 3.0, the default unit load is 150mA (that's a 50% increase over USB 2.0), so that's going to skew your comparisons as well.
Finally, regarding drive warmth: yes, this is normal. All SSDs get warm, regardless of who makes them (OCZ, Intel, Crucial, Corsair, etc.). This is due to the underlying controllers used; the term "warm" here is basically vague. When I touch a MHDD, "warm" would mean it's running at about 32C or so. SSDs run much cooler than that on average, but they are not cold. If you thought SSDs would be cold at all times, you thought wrong. :-)
I broke down and measured the current. I only measured 15 mA! That was during reads. I read a little less during idle. Since the drive is for a windows computer and the easiest way for me to measure the actual current was on my DeskTop Mac, I could not measure the current during writes.