Many condenser figure-8s are indeed synthesized from two back-to-back cardioids in opposite polarity--nearly all switchable-pattern dual-diaphragm microphones work that way (Neumann, AKG, etc.). Contrary to what someone posted, in principle the pattern is the same; in practice it's usually pretty nearly the same, except for a narrowing at high frequencies; dual-diaphragm microphones are most often large-diaphragm microphones, and those tend to get beamy toward the high end no matter what pattern you set them to.
Apart from that, however, the figure-8 setting usually has the most consistent pattern across the frequency range of any pattern that a multi-pattern mike will have. You can see that if you look at the polar diagram for a U 87, a C 414, a KM 56, etc.
And some stereo microphones do use separate cardioid elements for M/S. For example, the excellent Neumann RSM 190 and 191 stereo shotgun microphones actually have three capsules inside--one of which is used by the shotgun "M" microphone and a pair of back-to-back, side-facing cardioids whose signals are combined with a polarity inversion (i.e. subtracted) to form a figure-8 for the "S" signals. Neumann's KM 86 microphone actually had two complete KM 84-type (i.e. KK 84) capsule heads inside its own, larger capsule head, and that's how it got its figure-8 pattern. The extra spacing between the two capsules enhanced the low-frequency response of that pattern (though its high-frequency response off axis was rather peaky).
In practice you'd need to set the gains of the two cardioids very precisely (to within tenths of a dB) so as to match their sensitivities--and then having placed their membranes so that they're in the same vertical plane, you have to place a third (M) microphone as close as possible to the same point in space without, however, disturbing the acoustic symmetry of the pair of cardioids (which would distort the figure-8 pattern of the result), and without creating worse acoustic shadows or reflections than necessary.
All in all I think that for the most reliable results, a good small-diaphragm, single-diaphragm figure-8 with essentially flat frequency response is strongly preferable. But most of all, it's preferable to make a recording (if it's worth making) than not to make a recording. So don't wait for perfection, since I just heard that that's going to be delayed for yet another three weeks, no matter when you read this message.
--best regards