The seed may appear to be innert, but it's not. The embryo is carrying on a very slow metabolism within the seed. Eventually, if not planted in the proper environment, it will die.
The animal, including human, embryo, draws nutrients from the environment. The human embryo is never dormant - germination is immediate, from fertilization. If he or she is not in the proper environment, he or she will die, as we all would. The environment doesn't make the species.
Your analogy fails on the fact that the technicians who carry out IVF purposefully create the environment to enhance the growth of the embryo - at least for as long as they desire, in practical terms for at least 5 days. The limit is the wrong:
The responsibility of the ones who bring the oocyte and spermatocyte together in order to cause fertilization is to ensure the best conditions and environment possible for the nascent human being, without a limit, and certainly without a planned death.
My impression was that embryos have been frozen for years and later thawed and found to be viable. Yes, the environment is artificial, but I don't think an embryo frozen in liquid N2 is going to be metabolizing much of anything.
BTW, if I might use an analogy which will severely date me and very few people will understand, does the code stored in the $B600 ROM or on the sector of the first track on an Apple ][+ floppy constitute part of the operating system?
On startup, a normally-configured Apple ][+ will execute code at $C600; this code will create a data translation table in RAM, move the drive head to track 0, and read the first sector of the floppy into address $800-$8FF. The code in this sector will in turn read in code from the next few sectors (the boot-ROM code is so tight there's no room to even include a sector loop; the first-sector code can take advantage of the fact that the boot ROM set up the data transation tables and homed the drive head, so it can afford the loop to read the rest of the code.
Once the Apple ][+ has loaded the first sector, the code at $C600 will never be used again until the system is rebooted. Once the OS as a whole is loaded, the first-sector code will not be executed again until the system is rebooted. So do these bits of code constitute part of the "operating system" of a running system?
I would posit that the multi-step process of gestation is in some ways like the multi-step OS load on the Apple. A "bootstrapper" is prepared (roughly equivalent to the amniotic sac and placenta), then the bootstrapper prepares the rest of the OS, sharing memory with it (roughly equivalent to the way the amniotic sac and placenta protect and feed the developing fetus with which they share the womb), and finally the bootstrapper is discarded (as are the placenta and amniotic sac after birth).