China watchers were stunned this week by CSET's new data on Chinese civil-military fusion—the U.S. export control sanction regime has targeted only 8 percent of the PLA AI equipment suppliers. https://cset.georgetown.edu/publication/harnessed-lightning/

The figure is concerning and will likely lead to greater pressure on White House to buff up the sanction regime. But I've heard an alternative perspective on this data that I thought is worth sharing. The thinking goes: No matter how much we expand the U.S. entity list, PLA firms will be able to secure advanced hardwares anyways from alternative sources. The firms can either look to European or Asian tigers markets for equipment with equivalent capacities, or obtain U.S.-build equipment from under-the-table deals. An expanded sanction list will just send the hardware-purchasing activity underground, inflicting only a temporary dent in the civil-military fusion process as Chinese firms find alternative supply chains.

Therefore, my friend says, instead of toughening up the sanction regime, we should just let the vast majority of PLA suppliers continue to acquire U.S. hardware. That way, we can keep track of what *sort* of equipments the PLA suppliers are purchasing, and extrapolate from that data the current state of PLA AI development. If we send the equipment-purchasing activity underground, we lose this crucial source of intel. Keeping the PLA dependent on the U.S. supply chain will also leave China vulnerable to U.S. sanction regimes if they do something particularly egregious, like invade Taiwan.

I would add to my friend's thought and say that, whatever the sanction regime might be, we need more transparency on transactions involving the sale of U.S. technology abroad. That way, we can name-and-shame particularly egregious equipment sales (like of surveillance-related technology). Perhaps CFIUS reform to make the process more transparent to the public can help?