Analysts said that partly reflects how difficult it is to predict which skills – and which people – will prove most valuable as AI develops.
“People don’t know how AI will play out, so the search is broader,” said Guo from Hutong Research.
“You don’t know who you are looking for.”
For companies, that uncertainty creates an incentive to cast the net wider and earlier.
“We are still in the trial-and-fail cycle, and it’s more a fear of missing out,” Guo added.
“You want to hire the talents, but you do not know what they can do yet.”
That does not mean everyone with AI skills is commanding extraordinary salaries.
Wang from the Eurasia Group, said a rare researcher with exceptional mathematics and algorithm skills occupies a very different market from a graduate who applies existing models.
“It’s really just a handful of talents we’re talking about, and that’s probably why there is this very expensive investment in signalling,” she said.
EIU’s Su similarly described the outsized pay packages as “isolated cases”.
But when companies cannot yet know where the next generation of talent will emerge, there is little incentive to wait until it is fully formed.
“It’s better to access those talents as young as possible,” Guo said.
“It’s worse if the talents go to the US, stay there, and never come back to China.”