Corporate Services (CS), the facilities organization at Intel, spent nearly two years refining a TCO model designed specifically for the corporation and its internal customers. Having a model that takes the guesswork out of planning for the future is especially important to a large corporation like Intel, which has more than 300 worldwide facilities, over 100,000 employees, and annual revenue in excess of $38 billion. CS employs 4,000 people, manages 54 million sf of space, and has an annual expense budget of approximately $2.2 billion.

“With its successive generations of microprocessors, Intel is getting horrendously complicated. We are facing decisions from the manufacturing organization which constantly wants to introduce new products, new gasses, new chemicals, or new processes,” says Brendan Keogh, program manager for the TCO model at Intel Ireland. “More facility systems are being added, the manufacturing process is getting more complicated, and costs are increasing. We needed to come up with a methodology that looked at the real costs associated with those decisions.”

The efforts of CS to develop a model capable of forecasting future operating and maintenance expenses resulted in the creation of TCO. The model is being celebrated by Intel as an industry-changing tool that can bridge a fragmented organization, help a customer define its value proposition, provide a vision for the future, move facilities maintenance from a support position to a partner role, and create cost savings.

Defining TCO

Total Cost of Ownership gives a comprehensive and objective analysis that can be used for decision-making purposes regarding capital expenditures, the purchase or maintenance of equipment, and process design. CS defines TCO in terms of a so-called facilitization factor, or making it applicable to a particular facility and its future requirements.

A primary component of the model involves expressing the costs of systems, services, and utilities in the customer’s terms.

“Speaking the customer’s language is the first key learning we got from this process,” says Kris Goranson, CS strategic initiatives manager. “In our particular case, when we say customer, we mean perhaps the manufacturing organization or the data center for which we are providing the facility. We now express our costs in terms of units that make sense to the customer.”

Expresses Utility Costs in Customer’s Terms

Intel produces its microprocessors and other chips on circular silicon discs, known as wafers, which range in size from six to 12 inches in diameter. At the very highest level of the model, total facility cost is provided relevant to the customer’s output. For example, a wafer facility that costs $100 million per year to operate from a facilities standpoint and produces one million wafers would have an annual facility cost of $100 per wafer.

“While simple in concept, it allows the facility organization to position facilities as a key input to the manufacturing process. Further refinement of the model allows us to take this dollar per unit of production all the way down to the individual utility or facility system,” says Goranson. “This allows for cost reduction exercises to be undertaken in a focused manner.”

TCO also models the facility with regard to what the customer consumes, models the dependencies between systems, and combines construction costs and operating expenses to provide a Total Life Cycle Cost. Typical facility systems, involving a multiplicity of factors such as heating, lighting, air conditioning, access, and fire protection, are often more complex than customers understand.

Models the Facility in Terms of What Customer Consumes

Chip manufacturing, in particular, deals with hundreds of inter-dependent systems from power to chemicals, gasses, waste, air, and space. In the case of Intel, the customer has a requirement for at least 120 discrete systems, many of which are straightforward gasses or chemicals supplied to manufacturing tools. Other physical spaces, such as the cleanroom, are more complex and must be maintained to very narrow ranges of temperature, humidity, and cleanliness.

The customer can express his requirements, but it is unlikely he will specify the requirements as they relate to support systems. TCO addresses this through the use of primary and secondary systems. Primary systems are outputs to the customer, while secondary systems support primary systems.

For instance, process cooling water is used to cool the tools that produce the wafers. The customer understands process cooling water and has specifications from his tool vendors to talk to about how many gallons of process cooling water he needs. The water system requires chilled water systems to provide the cooling capability and it also requires power and space in the utility buildings.

In this scheme of things, chilled water becomes a secondary system, supplying capacity to the air-conditioning system which, in turn, supplies capacity to the primary cleanroom system.

“What the customer cares about is how much the process cooling water will cost. A simple example shows it costs $10 million to buy the chilled water system or to build it, and it costs $1 million a year to maintain it. We look at depreciation associated with the capital spend. Typically, we use a 10-year depreciation for chilled water systems,” explains Keogh.  “That $10 million of capital investment is one million a year in depreciation and another million a year in operations. It is costing about $2 million a year for the chilled water system. That is how we model it.”

Models the Dependencies Between Systems

The TCO model provides burdened costs for primary systems based on detailed engineering analysis that calculates the capacity demanded by each system from every other system, modeling the relationships throughout the entire facility. These capacity relationships determine the burden or allocation from one system to another. For example, if the chilled water sends 50 percent of its capacity to the cleanroom air conditioning and 50 percent to the process cooling water system, then each of these systems will carry half the total cost of the chilled water system. Therefore, all costs associated with a facility are ultimately reflected in the primary systems and no costs are left unallocated.

Combines Construction Capital with Operating Expenses

Intel’s model combines the capital cost of construction, expressed in depreciation terms, with the annual operating costs for a system.

“This provides a total annual cost, which we usually run out over a 20-year horizon. The relevant period tends to be a lot shorter since most factories go through major refit cycles every few years,” says Keogh. “The model also allows us to add capital investments at different points in time to reflect system upgrades.”

TCO Drivers

CS’s decision to create Total Cost of Ownership was prompted by the increasing complexity of the manufacturing process, the need for facilities to influence upstream, and the goal of better predictability.

“The complexity of the manufacturing process, driven by the product roadmap, was creating a demand for more and more facility systems, thereby causing cost pressures for the facility organization,” says Goranson.

CS also introduced TCO because there is a need for facilities to influence upstream decisions. Manufacturing tool selection is driven by cost and performance considerations. Having a TCO model that produces a fully burdened cost for each utility allows the facility organization to offer input into the tool selection process.

The TCO model identifies a dollar per unit of measure for each utility or facility, including physical space, which is used as part of the tool selection process.

“We have discovered that vendors aren’t particularly sensitive to the real cost of utilities,” says Goranson. “Thanks to the whole TCO process, we now understand the real cost of each utility and we are feeding that back to the vendors. Boiler-plate utility requirements are being re-examined and so far, we have seen a number of key reductions on tools by manufacturers revisiting their calculations.”

The model developed by CS provides a methodology for understanding the utility systems required by customers, as well as the necessary support systems. Total facility requirements can be modeled by projecting the required capacities for future technology. This means a greater deal of accuracy when predicting future facilities and this data is used to estimate construction and operating costs.

CS also has a model that links utility capacity with total cost, allowing facilities management to scale based on changing capacity information and project costs that can be used in high-level decision support. CS is now working to use TCO analysis as part of its business-as-usual model for Capital Project Authorization to link capital spending to long-term affordability.

What Made TCO Possible

CS recognized there was a need to impact facility decisions driven specifically by manufacturing process changes. Senior management offered its support in the quest to develop a model geared toward accurately predicting the long-term affordability and fiscal responsibility of projects.

A team was assembled to reflect representation from both non-technical and technical disciplines, including facility engineering, operations, construction, finance, systems, and software development.

“More importantly, individuals were selected for their passion for change and drive for innovation, folks who were prepared not only to try something different, but also to be flexible in their approach and be prepared to fail,” says Keogh. “Much of what the TCO program has been about is a change process, change in how we understand our facility systems and how we drive change as a result of our findings. In order for TCO to be effective, there must be appropriate time to gather data, draw conclusions, and influence outcomes. It will take at least two or three years to embed this type of modeling into your organization.”

The modeling is geared toward providing codified knowledge through the development of software applications to manage the complexity of the system-to-system relationship and to provide a central repository for cost and capacity data. When key information is scattered throughout an organization in an individual’s head, on a laptop, or in a spreadsheet, valuable time is lost trying to find the proper data upon which to base a decision.

TCO is rapidly spreading throughout CS as its value is realized by various groups. Trying to influence decision-makers and key stakeholders in such a large organization is a full-time job for CS. It is important to have people on board who have networking and relationship-building skills. Being flexible enough to change the model, when necessary, is key to its success.

“Increasingly, TCO is being asked to model different aspects of the business. We are expanding the modeling capability to assemble test facilities and data centers at the request of those organizations,” says Goranson. “We are continuing to sell the concepts and benefits internally, developing new opportunities, looking for new possibilities. This requires a constant marketing mindset within the team.”

Maintaining a flexible approach both in terms of the modeling capability and the application development allows CS to tailor the TCO program to meet the requirements of new customers, as well as the ever-changing needs of existing customers.

“It has been a very exciting and fun voyage,” says Keogh.

By Tracy Carbasho