Red,
First, I presume the measurement your customer needs to make is a low-side (shunt is referenced to ground). The choice of the gain is a trade-off with the size of the shunt resistor used unless they also try to minimize the power dissipated in the shunt; in order words, if they pick 10x larger shunt (5mohm resistor instead of 0.5mohm, they will need to lower the gain by 10x, for the same output range.
Second, the gain of 1000 requires the input resistors to be 200ohms instead of 266ohm.
Third, 2Mohm pull up resistor prevents the output from being pushed into negative rail under 0A shunt current so the low current levels can be measured.
With 2Mohm pullup maximum shunt current of ~5.6A can be measured before the output becomes non-linear closer than 50mV from either rail - see below.
In order to increase the Ishunt range, customer may choose a higher value of pull-up resistor, which will assure linear operation of the output stage under Ishunt=0A while allowing higher shunt current measurement above 6A - see below.
Also, there is no need to buffer the output with another OPA376 since a single stage is able to do the same thing in terms of output swing and driving capabilities.
