Direct Answer

Semiconductor fab utilization is the percentage of a chipmaker's production capacity actually being used. It's relevant primarily to companies that own and operate their own fabrication plants, integrated device manufacturers (IDMs) and dedicated foundries, rather than fabless companies that outsource manufacturing. High utilization, often cited in the 90%+ range, generally supports strong gross margins because fixed manufacturing costs are spread across more units produced. When utilization falls during industry downturns, those same fixed costs are spread across fewer units, which pressures margins.

Key Takeaways

  • Utilization is capacity in use, not capacity that exists. A fab can be large and modern while running well below its practical maximum if demand is soft.
  • High utilization (often cited in the 90%+ range) generally supports strong gross margins. Fixed manufacturing costs, depreciation, facility overhead, staffing, are spread over more units produced.
  • Falling utilization during downturns pressures margins because the same fixed costs are spread over fewer units, even if selling prices hold up.
  • The metric is most relevant to companies that own fabs. Fabless companies don't report a utilization rate of their own since they don't operate manufacturing plants.
  • Utilization is cyclical. It tends to move with broader semiconductor demand cycles and can lag or lead margin commentary depending on how quickly a company adjusts output.
  • Definitions vary by company. There is no single standardized disclosure format, so utilization figures are best read within one company's own historical trend rather than compared precisely across firms.

How Fab Utilization Works

A semiconductor fabrication plant, or fab, is built to a certain maximum production capacity, the volume of wafers or units it can process in a given period when running continuously. Utilization is the share of that capacity actually being used:

Utilization Rate = (Production Actually Running) ÷ (Total Available Production Capacity) × 100

Building and operating a fab is capital-intensive: the equipment, cleanroom facilities, and specialized staffing represent largely fixed costs that don't scale down proportionally when fewer units are produced. That's the core mechanic behind why utilization affects margins. When a fab runs near its capacity, the fixed cost per unit produced is low, because the total fixed cost is divided across a large number of units. When demand softens and the fab runs at a lower percentage of capacity, the same fixed cost total is divided across fewer units, raising the fixed cost embedded in each one.

This is why utilization shows up as a recurring theme in semiconductor company earnings commentary, particularly for integrated device manufacturers (IDMs) that own their fabs, companies such as Intel, Texas Instruments, and Samsung's semiconductor division, and dedicated foundries such as TSMC and GlobalFoundries, which manufacture chips on behalf of other companies. Fabless companies, such as Nvidia, AMD, and Qualcomm, design chips but contract manufacturing out to foundries, so fab utilization isn't a metric they report about their own operations. It still matters to them indirectly, since foundry capacity and pricing are shaped in part by the foundry's own utilization levels.

Worked Example: Utilization and Fixed Cost Absorption

Hypothetical example, for education only. These figures are illustrative and do not represent any real company's actual capacity, output, or costs.

  1. Set up the fab's capacity and fixed costs. Assume a fab has practical capacity of 100,000 wafer starts per month, and fixed manufacturing costs (equipment depreciation, facility overhead, base staffing) of $500 million per month, regardless of how many wafers are actually processed.
  2. Calculate fixed cost per wafer at high utilization. If the fab runs at 95% utilization, it processes 95,000 wafers that month. Fixed cost per wafer = $500,000,000 ÷ 95,000 ≈ $5,263.
  3. Calculate fixed cost per wafer at reduced utilization. If demand softens and utilization drops to 70%, the fab processes 70,000 wafers. Fixed cost per wafer = $500,000,000 ÷ 70,000 ≈ $7,143.
  4. Compare the two scenarios. The drop from 95% to 70% utilization raises the fixed cost embedded in each wafer by roughly 36%, even though nothing about the fab's efficiency changed, only the number of units available to absorb the same fixed cost base. If selling prices per wafer stay constant, that higher embedded cost flows directly into a lower gross margin.
  5. Read the direction, not a universal number. The specific dollar figures above are illustrative; the takeaway is the relationship, utilization down, fixed cost per unit up, margin pressure up, all else equal, which is the pattern investors watch for in real company disclosures and earnings commentary.

Limitations and Common Mistakes

Treating utilization as universally comparable across companies

Because there's no single standardized disclosure format, how one company defines and measures utilization can differ from another's. A given percentage from one manufacturer isn't necessarily measuring the same thing as the same percentage from a different one. Utilization is most reliably used to track a single company's trend over time rather than to rank companies precisely against each other on this figure alone.

Detailed close-up of a microprocessor circuit board showcasing intricate circuitry and components.
Photo by ed br via Pexels

Applying the metric to fabless companies

Fab utilization is relevant primarily to companies that own and operate their own fabs. Fabless companies don't have a utilization rate of their own to report, since manufacturing is outsourced. Looking for this metric directly in a fabless company's disclosures, or assuming it drives their margins the same way, misapplies the concept.

Assuming utilization alone explains a margin move

Utilization is one input to gross margin, not the only one. Product mix (higher-margin advanced nodes versus lower-margin legacy nodes), pricing, yield rates (the percentage of chips on a wafer that meet quality standards), and input costs all affect margins as well. A margin change can reflect some combination of these factors, so utilization commentary is best read alongside the rest of a company's earnings disclosure rather than in isolation.

Ignoring the cause behind a utilization change

A lower utilization figure can reflect a broad industry downturn, a company-specific demand issue, planned maintenance, or newly added capacity that hasn't yet been filled with demand. These have different implications for how long the pressure might last, so the reason behind a utilization change matters as much as the direction.

FAQ

What counts as high fab utilization in semiconductors?

There is no single official threshold, but utilization in the 90%+ range is commonly cited as the point at which a fab is running near its practical maximum and fixed manufacturing costs are being spread over the largest possible number of units. What counts as "high" varies by company, process node, and product mix, so a specific percentage is best read in the context of that company's own historical range rather than as a universal cutoff.

Why does fab utilization affect gross margins?

Operating a fabrication plant involves large fixed costs, depreciation on equipment, facility overhead, and staffing, that don't change much whether the fab runs near capacity or well below it. When utilization is high, those fixed costs are divided across more units produced, lowering the fixed cost embedded in each unit and generally supporting stronger gross margins. When utilization falls, typically during industry downturns, the same fixed costs are spread over fewer units, which pressures margins even if selling prices hold steady.

Which companies report fab utilization?

Fab utilization is relevant primarily to companies that own and operate their own fabrication plants, integrated device manufacturers (IDMs) and dedicated foundries. Commentary on utilization trends most commonly appears in these companies' earnings calls and investor materials rather than as a single standardized line item, and disclosure practices and definitions vary by company, so figures are not always directly comparable across firms.

Does fab utilization apply to fabless chip companies?

Not directly. Fabless semiconductor companies design chips but outsource manufacturing to foundries, so they don't operate fabs and don't report a utilization rate of their own. Fab utilization still matters to fabless companies indirectly, however, since foundry capacity and pricing, which are influenced by the foundry's own utilization, affect the fabless company's manufacturing costs and supply availability.

How does fab utilization differ from capacity expansion?

Capacity is the maximum volume a fab (or network of fabs) is built to produce; utilization is the percentage of that existing capacity actually being used at a given time. A company can be expanding capacity, building new fabs or adding tools, while utilization of its current capacity is falling if demand hasn't kept pace, or vice versa. The two move independently and both matter: capacity decisions are long-term investment choices, while utilization reflects near-term demand relative to the capacity already in place.

Is falling fab utilization always a bad sign for a chipmaker?

Falling utilization generally pressures gross margins because fixed manufacturing costs are spread over fewer units, but the cause matters. A broad industry downturn affecting demand across the sector is different from a company-specific issue, and utilization can also dip temporarily during planned maintenance, a transition to a new process node, or newly added capacity that hasn't yet been filled with demand. Reading utilization alongside revenue, order trends, and management commentary gives a fuller picture than the utilization figure alone.

Why can utilization at leading-edge nodes diverge from mature nodes?

Leading-edge and mature capacity serve different products and different customers, so they can move independently. A period of strong demand for advanced processors can keep the newest lines full while older lines producing analogue, power, or microcontroller parts sit underused, or the reverse. A single company-wide utilization figure blends the two and can mask a meaningful split. Where a company discloses by node or by segment, the detail explains margin behavior that the aggregate figure does not.

How do underutilization charges appear in cost of goods sold?

When a plant runs well below normal capacity, accounting rules generally prevent the unabsorbed fixed overhead from being capitalized into inventory, so it is expensed in the period instead. Companies sometimes disclose the amount as an underutilization or idle capacity charge. The effect is that margin falls faster than volume during a downturn, and the disclosed charge indicates how much of a margin decline came from running the plants below normal rather than from price.

Why is fab utilization hard to compare with published manufacturing capacity utilization statistics?

National statistical agencies publish capacity utilization for broad manufacturing categories using their own survey definitions and seasonal adjustment. Chipmakers define capacity by their own standards, sometimes by wafer starts and sometimes by equipment availability, and disclosure is voluntary and inconsistent. The two series answer related questions from incompatible measurements. Using an official statistic as a proxy for what an individual company is running at introduces more error than it resolves.

References

  • SEC EDGAR: company 10-K and 10-Q filings, where semiconductor manufacturers commonly discuss capacity, utilization, and manufacturing cost trends.
  • SEC EDGAR: Company 10-K Filings: search interface for individual semiconductor manufacturers' annual reports.

Disclaimer

This article is for educational and informational purposes only and does not constitute personalized investment, financial, or legal advice. Utilization figures, disclosure practices, and margin dynamics vary by company and change over time. Always verify current data from primary sources such as company filings. Trading involves risk, including the possible loss of principal.