TBM-940 pilot reported the aircraft lost airspeed and engine torque as the autothrottle began to reduce power. The pilot stated the throttle was attempting to prevent the engine from overheating after the anti-ice system was activated during climb.

Narrative

The plane was climbing under autopilot control; and ATC cleared me to 24;000 feet. I was about to enter clouds so I turned on the Inertial Separator. The automatic de-icing system turned on as we entered the clouds. About the time the inertial separator was fully engaged the high temp engine warning came on and peaked at 840 degrees. I glanced at the temperature gauge as the air speed audible warning sounded (plane had slowed still in the climb to 95 knots).I tried to push the yoke forward to encourage the autopilot to pitch down; and it didn't work; so I turned off the autopilot; and pitched down to gain airspeed.Once I was pitched down I checked my instruments and the torque had gone to zero. I assumed since the NG and temperatures were still live that my pump was not delivering enough fuel; so I used the manual override as my throttle.In retrospect the autothrottle likely shut the torque to zero to avoid over heating the engine; and when I turned off the autopilot to 'fly the plane'; the throttle was in idle position; so that is where it remained. This all happened in 10 to 15 seconds. Once I had control of the plane and I was flying using the manual override I [requested priority handling] thinking my fuel pump was going out. In hindsight it was overly conservative; but having the engine spool to zero torque in IMC; in the mountains; with icing conditions; with my wife and kids on board my priority was to get the plane on the ground.The root cause as I see it now is: the autothrottle allowed the engine to get too hot when the inertial separator came on; and then it pulled back once the temps hit a critical point (840). As the autothrottle pulled back the plane did not immediately pitch down which caused the speed to drop to 95 knots.By happenstance I turned off the autopilot to pitch down as the torque was at zero. If it was at any other number I probably would have just pushed the throttle forward instead of immediately jumping to the conclusion that I had a fuel pump issue. In our simulated training we are taught that if the torque goes to zero and the temperature and NG are still high that we should immediately use the manual override; which I did. My mistake was assuming the torque at zero meant the engine was failing (ie. pump was not delivering enough fuel). If I would have pushed the throttle forward I would have realized that the fuel system was functioning; and likely not declared an emergency. As a side note; I have never seen the autothrottle take the torque to zero except when we are on a very aggressive approach such as ZZZ.This entire event was less than one minute; so please forgive me if I am not remembering the facts correctly. I will amend as the flight data clarifies the sequence of events.