Optimizing Capital Investment Decisions at Intel Corporation

The Problem

To maintain a leadership position, Intel develops improved manufacturing process technologies with a two-year cadence. Over the years, this has resulted in transistors shrinking to nanoscale, enabling ever more capable products. Each of these process advances necessitates the purchase of a few billion dollars’ worth of progressively more sophisticated manufacturing equipment to outfit Intel’s factories.

The core business problem is one of timing and risk management. There is tension between Intel’s desire to delay ordering manufacturing equipment until demand forecasts are more reliable (one to two quarters ahead of sales) and the need to place orders at the lead times required by suppliers (four to six quarters in advance). The cost of over-ordering is a few hundred million dollars of equipment depreciating while sitting idle. The cost of under-ordering may be a few billion dollars in missed revenue or worse—lost market share. Historically, this asymmetry has enticed Intel’s decision makers to repeatedly overcapitalize by several hundred million dollars.

The Analytics Solution

Intel’s most recent solution to this problem was the development of a dual-mode multi-period procurement methodology with each supplier. The “base” mode supports ordering capacity at a normal lead time (four to six quarters) to cover the high-confidence, lower bound demand forecast. The “flex” mode is realized through option contracts that include reduced lead time (two quarters) for covering the capacity gap to the low-confidence, upper bound demand forecast. Balancing the two modes enables agility and flexibility in Intel’s capacity supply chain while guaranteeing risk sharing with suppliers, thus contributing to a healthier ecosystem.

The Value

Initial success with one supplier led to the opportunity to expand and engage many suppliers. More recently, adoption of this new decision method for capital investments has been driven by the documented benefits. When the initial demand forecast turned out to be too high (+10% to +15%), Intel realized approximately $300 million in benefit through abandonment of the options, avoiding the purchase of unnecessary equipment. When the initial forecast turned out to be too low (-10% to -15%), the options were exercised early to build capacity more quickly than was previously possible, resulting in the capture of a $2 billion revenue upside.

The work has transformed the way Intel makes capital investment decisions, resulting in repetitive cost savings or revenue upside at each manufacturing process transition. The approach has fostered a win–win relationship with equipment suppliers by providing a practical risk-sharing mechanism. Customers worldwide benefit through reduced cost for ever more innovative and capable products. The method has broad applicability because any organization making capital investments based on demand forecasts can benefit from its implementation.