Done both for B-pillar and door ring work and I'd push back gently on the framing of the question. "Hot vs cold on servo" usually gets debated as if servo is the variable. It isn't ??the press is the easy part. The hard part is the line around it.
For 22MnB5 hot stamping you need:
A furnace that holds austenitizing temp (900-950?C) with line speed matched to your SPM. Drop a blank below ~830?C at the die and you don't get a full martensitic transformation; you get inconsistent hardness across the part and the warranty engineer calls you.
Transfer time under 8 seconds blank-out-to-die-close, including any centering. Robotic transfer with insulated grippers or a gravity drop chute, not human hands. Cycle time on hot work is fundamentally transfer-limited, not press-limited.
Water-cooled dies with proper flow balance. Quench rate above 27?C/sec is what gives you the 1500 MPa final strength. We had a die that ran 1450 MPa for six months because one cooling channel had a 30% flow restriction nobody caught ??the press tonnage looked perfectly normal the whole time.
For cold stamping high-strength (DP980, MS1300) on the same servo you want:
Longer BDC dwell (200-400ms typical) for springback control, which mechanical presses can't really do. This is where servo earns its money on cold work.
Die material upgrade ??Vanadis 4 Extra or similar PM steel. Conventional D2 will galling-fail in 50K hits on DP980.
Nitrogen springs sized for the higher stripping forces. Blanking-grade gas pressure, not coining pressure.
Our shop runs both on the same servo line by changing the front-end (furnace in or out) and the cooling water plumbing. The press itself doesn't change much. People who think hot stamping is "just stamp it hot" lose money for two years before figuring this out.