I suspect the battery itself is unchanged, and that Seidio just did their own independent tests on its capacity and decided to report the value they came up with rather than relying on the battery's nominal mAH value.
The problem is that a battery's capacity in mAH depends upon BOTH its total energy-storing capacity AND "how hard" you're drawing power from it. For example, a given battery might be able to supply 250mA for 4 hours, but the same battery might be able to supply 100mA for 16 hours. If you simply multiply current draw times hours it can be sustained, you might conclude that the first battery has a capacity of 1000mAH and the second has a capacity of 1600mAH, but in fact they could easily be the exact same battery... and that battery might come from the FACTORY with a nominal rating of 2400mAH, because THEY tested it with a ridiculously-low current load, like 10mA.
Just to give a recent example, one popular site rated the capacity of Seidio's extended battery for the Evo (nominally 3500mAH) at ~2800mAH, because that's what it appeared to be when he tested it with a constant load of 250mA. Part of the problem is that right now, there really ISN'T any good metric for objectively describing the "true" capacity and quality of a battery, and the life of any given battery will vary WILDLY depending upon how it's used. Someone whose phone spends most of the day sleeping, with an occasional 5 minute phone call, might see an effective life of 2500-3400mAH from a battery. Another user who uses the phone as a pocket laptop that's always online via 4G and tracking its location via GPS might see an effective life of 1800mAH from the same battery.
Of course, it's also possible that Seidio scaled back the battery for one that's a little bit thinner, but my guess is that they just decided to re-evaluate its capacity and publish a battery capacity that's more in line with what users are likely to see in real-world use. Even though their Evo battery is by leaps and bounds the best and longest-capacity of all known extended batteries, it still made them look kind of bad that they were advertising a capacity of 3500mAH for a battery that seemed to test out at 2800mAH (though, in reality, they themselves were basically going by what the manufacturer of the battery itself said, and couldn't really do much more than try to find a manufacturer of quality battery cells and hope that its rating bore some resemblance to reality).
The best analogy I can give is amplifier power vs THD and speaker impedance. If you state ONLY maximum power, without qualifying it by THD and/or RMS, you can claim just about anything. Even a small junky amp costing $15 can output a kilowatt for a quarter of a second into a 1-ohm speaker load before going up in smoke. What makes the wattage-figure meaningful is when you say "RMS, at .1% THD into an 8-ohm load". Right now, there's no equivalent battery rating. We have "mAH", but without the context of load and max temperature, it's a statistic that's almost devoid of meaning on its own.
There's also the matter of choosing meaningful parameters. An Epic or Evo tethering 4G via wi-fi will drain its battery while drawing power from what it believes to be a 500mA-max USB port. An Epic or Evo throttled down to 400MHz with the display off polling for incoming phone calls might draw ~100mAH. I'm kind of inclined to say that 250mAH is the most realistic draw given the way most of the people who BUY extended batteries actually USE their phones, but it's kind of an awkward metric any way you look at it.