Good framework. I'll add some numbers from our experience buying 3 servo presses over the past 6 years that might help people building their own models.
The hidden ROI killer that nobody puts in the spreadsheet: die life extension. We run progressive dies for automotive connectors — small parts, high volume, tight tolerances. On our mechanical press we were resharpening every 180K hits. Same die on the servo press with optimized velocity profile (slow through the cutting zone, fast everywhere else) gets us to 280K hits. That's 55% more life per sharpen cycle. At $1,200 per sharpen and 12 sharpens per die life, that adds up fast across 40+ active dies.
On the energy side — the real savings aren't just kWh. It's peak demand charges. Our mechanical press line had a 450 kW peak demand that we were paying $18/kW/month for regardless of usage. The servo presses with regenerative drives and staggered stroke timing brought peak demand down to 280 kW. That's $3,060/month in demand charge reduction alone, separate from actual energy consumption.
One thing I'd push back on: the "reduced setup time" benefit is real but overstated in most ROI models. Yes, you program instead of adjusting cams. But the first time you set up a new job on a servo press, you're still spending 2-4 hours dialing in the motion profile, testing at low speed, adjusting BDC, etc. The savings come on repeat jobs where you just recall the recipe. If you run 200+ different part numbers like we do, the recipe library becomes incredibly valuable. If you only run 10 jobs, the setup time savings are marginal.
My rule of thumb: if your payback calculation depends on energy savings alone, don't buy the servo press. If scrap reduction + die life + flexibility together get you under 3 years, it's a solid investment.