That's a classic purchasing dilemma I face all the time! When dealing with high-cost, long-lead-time critical parts like those $15,000 servo drives, I use a multi-layered approach. First, I conduct a criticality analysis - I look at which equipment absolutely cannot go down without crippling production. For those mission-critical machines, I'll keep at least one spare on hand despite the cost.
Then I use ABC analysis combined with VED (Vital, Essential, Desirable) classification. Your servo drives would likely be 'A' items (high cost) and 'V' items (vital to operations). For these, I calculate the true cost of downtime - not just lost production, but also rush shipping, overtime labor, and potential customer penalties. If a 6-hour downtime costs $50,000, that $15,000 spare starts looking like insurance rather than inventory.
I also explore alternatives like repair programs, consignment stock with suppliers, or shared inventory pools with other plants. Sometimes I'll negotiate with suppliers for advance purchase discounts or guaranteed delivery windows. The key is treating these decisions as risk management, not just inventory management - you're buying operational resilience, not just parts.