For a mass mobilization army in the 70s that is fine. But the PLA is not that. They have limited manpower just like everyone else. And frankly modern clean-sheet IFVs arent affordable because they still need a weapon and fire control for that, plus comms, navs, networking, some sort of remote counter-mining system, APS, propulsion/engine, situational awareness enabler(s), and then actual materials for the chassis and armour. That's not taking into account whatever the PLA is looking at for their future IFVs but it's guaranteed to be expensive.
So I guess that you can make a modern BMP-2 if you have an endless stream of ultranationalistic brainwashed conscripts, but that's just going to make any mass armoured assault a human wave attack thats doomed to fail. The BMP-2 was so cheap because it was literally crap for its time ( active 1st gen NV, no FCS, reuse BMP-1 chassis which has no room for large growth, borderline 0 armour). Making that in the 21st century is just asking to be attritioned to death when any competent military would be fielding ICMs for their tubes that would clean up IFV companies like roomba through rubbish, or a FOGM spam from 50-100km away that your arty cant CBAT because the enemy is thrusting into your lines with something like Assault Breaker or massing SDB swarms while your C2 centers is a mess because the guys that are attacking are screaming for help. And then trying to make it modern runs into rising costs because the FCS cost so much ( you need thermals anyway), the hull cost so much ( cant really have an IFV that can be penned by GPMG fire or break apart after a 500g mine) or secured electronics and literally everything else.
You better spend your money on making attritable UGVs that can replace your MBTs and MICVs in the frontline, and put the human in a MBT-derived MICV because you need protection for the guys operating those radios, UGVs, UAVs, etc. Robotics warfare is attritional as Ukraine have shown, and robots can be procured en masse but not humans.