The solution
Intel's products and processes became exponentially more complex, and design choices cascade into which parts are built where and how much inventory sits in the chain.
The traditional approach lets product architects optimize the design and lets supply-chain planners later scramble to make it work, and the mismatch is expensive.
The move was a corporate-wide capability that generates design alternatives with genetic algorithms and device-physics simulation, then couples them with large-scale supply-chain planning using decomposition and mixed-integer programming.
Across most of Intel's portfolio it added an average of $1.9 billion a year in revenue and cut $1.5 billion a year in cost, a total benefit of $25.4 billion since 2009, while supporting sustainability. It won the 2020 Edelman Award.
What it achieved
Optimize design choices and supply-chain plans as one
Why it worked
Architecture choices determine months of inventory and billions in capacity spend.
Optimizing the design without the supply chain understates its true cost.
The joint model could explore far more business scenarios in far less time.
A faster, better product-to-market response showed up as revenue.
What can be applied
When a design decision ripples into inventory and cost, do not optimize the design and the plan separately. Couple them so the architecture is chosen knowing its supply-chain consequence.
Aftermath
Intel rolled the capability across most of its product portfolio. INFORMS recognized it with the 2020 Franz Edelman Award, and Intel credited it with $25.4 billion of cumulative benefit since 2009.
FOLLOW THE EVIDENCE