According to Wikipedia:
The use of gigabyte may be ambiguous. Hard disk capacities as
described and marketed by drive manufacturers use the standard metric
definition of the gigabyte. Therefore, one gigabyte is 1 000 000 000
This is true even for phone manufacturers, in fact just as rjknight pointed out, The RAM marketing strategy is similar to the "hard drive maker conspiracy" story is driven by the fact that manufacturers have no real incentive to switch to binary prefixes, because that would make their drives (RAM in this case) look "smaller".
AFAICT, using the mebibyte (MiB), gibibyte (GiB) and so on, is a more accurate way of presenting the volume of data stored in these quantities.
The gibibyte or binary gigabyte is a multiple of the unit byte for
digital information . The binary prefix gibi means 2^30 , therefore
one gibibyte is equal to 1 073 741 824 bytes = 1024 mebibytes .
These other binary prefixes (kibi, mebi, tebi, etc.) were introduced in an attempt to reduce such confusion, but these prefixes have yet to see widespread adoption.
In PC 2GB of ram refers to gibibytes, not gigabytes. RAM comes in allocation in powers of 2 due to architecture. Memory addressing is done in binary form, this implies it is going to be a power of two.
So, since writing memory addresses in binary always results in a binary number, RAM developers have (almost) always stuck to creating RAM in units of powers of 2, combining individual chips which are (almost) always combining to powers of 2.
Here is another explanation:
RAM is addressable by an index called an address . The most efficient
way to construct this address is as a binary number which corresponds
to the physical area on the chip where the byte or word of memory is
accessed from. As a matter of efficiency its expensive to check this
address for validity (does the address really correspond to a real
address?) for every access. The only way to restrict the address
without having to check the value is to allow for every possible
combination of bits, but set the exact number of valid bits that are
supported for the RAM chip. But that means your range has to be a
power of 2.
RAM that is used in smartphones is technically DRAM, with the D
standing for dynamic. The structure of DRAM is such that each
capacitor on the RAM board stores a bit, and the capacitors leak
charge and require constant “refreshing”; thus the “dynamic” nature of
the RAM. It also means that the contents of the DRAM module can be
changed quickly and easily to store different files.
And as expected, the exact number of bytes in a DRAM module is always an integral power of two.
This is is more like a business idea, that takes advantage of confusion regarding the common usage of "gigabyte" instead of "gibibytes" much like ISPs advertise 8Mbps (= 1MBps) as 8MBps, so it needs awareness from end users or customers.
It seems most manufacturers hide under metric system or SI units to "advertise" to the unsuspecting consumers, although in real terms the say what they don't mean.