[UMC] UMC: Q3 2026 earnings preview on utilization, pricing and gross margin

United Microelectronics Corporation (UMC) is a Hsinchu, Taiwan-based pure-play foundry that manufactures mature and specialty-node wafers to customer designs, and it will hold an online investor conference on 2026-10-28 at 17:00 Taipei time to report its Q3 2026 results, covering July through September [1]. The latest full quarter on record is the second quarter of 2026: revenue was NT$68.73 billion, up 12.6% sequentially and 17.0% year over year, gross margin was 32.5%, operating margin was 21.8%, capacity utilization was 85%, and net income attributable to the parent was NT$42.26 billion, or NT$3.39 per share and US$0.537 per ADS [2]. Third-quarter revenue is already known from monthly sales reports: July, August and September came in at NT$23.84 billion, NT$25.04 billion and NT$25.35 billion [3][4][5], a total of NT$74.24 billion, up 8.0% sequentially and 25.6% year over year. On July 29 the company guided third-quarter wafer shipments up by a high-single-digit percentage, ASP in US dollars to remain firm, gross margin in the mid-30% range, utilization above 90%, and a 2026 capital expenditure budget of US$2.0 billion [6]. ChartRow lists a third-quarter EPS estimate of US$0.23 per ADS against US$0.20 a year earlier, but that figure comes from a single analyst estimate, so the sample is very thin [7].

With revenue already disclosed, this UMC Q3 2026 earnings preview turns on profit structure, and three things matter most. The first is gross margin: as utilization rises from 85% in the second quarter to above 90%, roughly 70% of manufacturing costs that are fixed or semi-fixed get spread over more wafers, but second-quarter depreciation was already up 12.6% year over year [8], so where gross margin lands between 34% and 36% shows how much of the utilization gain depreciation absorbs. The second is pricing: blended ASP rose by a low-single-digit percentage sequentially in the second quarter [9], while third-quarter revenue growth of 8.0% sits close to the high-single-digit shipment guidance, so how the company describes third-quarter ASP will separate a genuine end to price erosion from a 22/28nm mix effect. The third is cash and the expansion schedule: in July the board approved a phased expansion of about NT$148.68 billion [10], the company then issued domestic convertible bonds and US$1.8 billion of zero-coupon overseas convertible bonds [11], yet first-half cash spent on property, plant and equipment was roughly flat year over year [12]; whether third-quarter spending accelerates and whether operating cash flow keeps covering it will set the starting point of the depreciation curve after 2027.

Company Background and Business Structure

UMC is a manufacturing-only foundry that sells no chips under its own brand, and almost all of its revenue comes from wafers. The company was founded in Hsinchu in 1980, listed on the Taiwan Stock Exchange in 1985 under code 2303, and has traded on the New York Stock Exchange as ADSs since 2000; its main customers include integrated device manufacturers such as Texas Instruments and Intel and fabless design houses such as MediaTek, Realtek and Novatek [13]. In the second quarter of 2026, wafer revenue was NT$66.44 billion, or 96.7% of consolidated revenue of NT$68.73 billion, with the remaining NT$2.29 billion coming from masks and other revenue [14]. As a foreign private issuer, UMC files 20-F and 6-K reports with the US Securities and Exchange Commission and publishes consolidated revenue every month under Taiwan rules, which is why its third-quarter revenue was public before the investor conference.

UMC's product mix is concentrated in 28nm and older mature nodes, and it does not compete in advanced logic below 14nm. By process technology in 2025, 28nm accounted for 36.8% of wafer sales, 40nm for 16.1%, 65nm for 17.0% and 90nm for 7.6%, with the rest on 8-inch processes at 0.11 micron and above; 14nm and below was 0 [15]. By the second quarter of 2026, 22/28nm made up 37% of revenue, with 22nm alone at 17.5%, both record highs [16]. Its fabs include the 8-inch cluster in Hsinchu and Fab 12A in Tainan, Fab 12i in Singapore, Fab 8N in Suzhou, Fab 12X in Xiamen and Fab 12M in Mie, Japan; in 2025 estimated total capacity was 5,163 thousand 12-inch equivalent wafers, actual output was 3,885 thousand and average utilization was 75.2% [17].

UMC's customer and regional exposure is fairly concentrated, and it also owns a group of investment companies that move net income heavily. In 2025 the top ten customers accounted for 57.0% of revenue and the largest for 11.8% [18]; by customer headquarters, Taiwan was 38.6%, the US 22.0%, China including Hong Kong 15.8%, Korea 10.7%, Europe 8.5% and Japan 4.4% [13]. Among equity-method investments, as of June 30, 2026 UMC held 36.49% of the unlisted Hsun Chieh Investment and 26.78% of Yann Yuan Investment, along with listed stakes such as 18.08% of Silicon Integrated Systems and 12.85% of Unimicron, and the combined carrying amount of equity-method investments was NT$96.47 billion, nearly double the NT$48.64 billion at the end of 2025 [19]. These investment companies mainly hold listed shares, so their gains and losses flow into UMC's non-operating income in line with the Taiwan stock market rather than with the foundry business.

Financial History and Current Position

In 2024 and 2025 UMC was in a falling-price phase, with revenue edging up while profit fell clearly. Revenue in 2025 was NT$237.55 billion, up 2.3% from NT$232.30 billion in 2024, because shipments grew from 3,446 thousand to 3,870 thousand 12-inch equivalent wafers while a 5.4% decline in average selling price and a 2.9% appreciation of the New Taiwan dollar against the US dollar offset most of the gain [20]. Over the same period operating costs rose 7.7% to NT$168.65 billion, gross profit fell from NT$75.65 billion to NT$68.91 billion, gross margin fell from 32.6% to 29.0%, operating income fell from NT$51.61 billion to NT$43.95 billion and operating margin fell from 22.2% to 18.5% [20].

UMC's cash flow stayed steady during the profit decline because non-cash costs such as depreciation were rising quickly. Operating cash flow was NT$86.00 billion, NT$93.87 billion and NT$99.86 billion in 2023, 2024 and 2025, even as pre-tax income fell from NT$68.45 billion to NT$57.05 billion and then to NT$48.08 billion; the depreciation and amortization add-back reached NT$59.26 billion in 2025 [21]. Total depreciation expense rose from NT$37.76 billion to NT$45.47 billion and then to NT$56.43 billion over those three years [22]. Cash spent on property, plant and equipment was NT$47.75 billion in 2025, far below NT$91.47 billion in 2023 and NT$88.54 billion in 2024 [21], and cash and cash equivalents stood at NT$110.66 billion at year-end [23].

In the first half of 2026, recovering utilization lifted gross margin quarter by quarter, but net income was carried mainly by investment gains. First-quarter revenue was NT$61.04 billion; second-quarter revenue was NT$68.73 billion, up 12.6% sequentially and 17.0% year over year, with gross profit of NT$22.32 billion, a 32.5% gross margin and operating income of NT$14.95 billion [2][24]. Second-quarter non-operating income totaled NT$30.24 billion, of which the share of profit of equity-method associates was NT$23.61 billion versus only NT$0.45 billion a year earlier, so quarterly net income reached NT$42.22 billion, nearly three times operating income [24]. First-half revenue was NT$129.77 billion, up 11.3%, with a 30.9% gross margin and EPS of NT$4.68 [9]; at the end of the second quarter cash on hand was about NT$124.7 billion and total equity NT$443.9 billion [9].

On the cash side, first-half operating cash flow grew while capital spending had not yet scaled up. Operating cash flow in the first half of 2026 was NT$55.68 billion, up 21.2% from NT$45.92 billion a year earlier; cash spent on property, plant and equipment was NT$21.35 billion, roughly flat against NT$21.70 billion; and first-half depreciation was NT$30.81 billion, above NT$27.23 billion a year earlier [12]. In the second half, July-to-September revenue totaled NT$74.24 billion and nine-month cumulative revenue reached NT$204.01 billion, up 16.08% [5], but third-quarter gross profit, non-operating income and cash flow will not be published until October 28.

Operating Model

UMC's revenue breaks down into three multipliers: shipments, US-dollar pricing and the exchange rate. Revenue roughly equals 12-inch equivalent wafer shipments times blended US-dollar ASP times the NT dollar exchange rate, plus about 3% from masks and other services [14]; shipments depend on customer demand and available capacity, while ASP depends on quoted prices and on the node and wafer-size mix, with the company pricing per wafer or per die and reviewing prices every quarter [25]. In 2025 these multipliers pulled in opposite directions: shipments grew 12.3%, ASP fell 5.4%, NT dollar appreciation added 2.9% of translation pressure, and revenue grew only 2.3% [20]. Seasonality also moves shipments, and the company says the first half is usually a demand low because customers correct inventory [26], so a second-half shipment rebound on its own does not mean demand has changed structurally.

Gross margin moves mainly with utilization and pricing because most costs are fixed. In 2023, 2024 and 2025 about 63.9%, 69.6% and 70.8% of manufacturing costs consisted of depreciation, part of indirect materials, license-fee amortization, indirect labor and utilities, so higher utilization lowers the cost allocated to each wafer [25]; this link works with almost no lag, as a change in utilization shows up in the same quarter's gross margin. Average utilization was 75.2% with a 29.0% gross margin in 2025 [17][20], and 85% with a 32.5% gross margin in the second quarter of 2026 [9]. The opposing force is depreciation: once new capacity and equipment are placed in service, depreciation goes straight into operating costs, and on the second-quarter call the CFO said the higher third-quarter gross margin guidance mainly reflects higher utilization, but that over the next two or three years gross margin "will come along with the depreciation curve" [27]. Below gross profit, operating income also bears operating expenses, which were NT$26.95 billion in 2025, up 6.2% [20], and second-quarter operating margin was 21.8% [2].

Cash flow follows a different logic from profit, because capital spending leaves as cash first and turns into depreciation and capacity years later. Operating cash flow roughly equals pre-tax income plus depreciation and amortization minus working-capital changes, and of NT$99.86 billion in 2025 operating cash flow, NT$59.26 billion came from the depreciation and amortization add-back [21]. The main use of cash is buying equipment and buildings; the company says its expansion is supported by customers that have signed multi-year supply agreements, and the first phase of the new Fab 12i P3 in Singapore has a design capacity of 30,000 wafers per month on 22/28nm and was scheduled to begin production in the second half of 2026 [28]. The raised capital budget this year therefore depresses free cash flow first, lifts depreciation after equipment enters service from 2027 onward, and adds shipments only if demand keeps pace.

Industry and Competitive Position

UMC ranks among the largest mature and specialty-node foundries, but its competitive pressure comes precisely from capacity growth in that market. In its 20-F the company says foundry competition is especially intense during periods of overcapacity and inventory correction, that its rivals include other dedicated foundries, integrated device manufacturers with in-house capacity and newly operating foundries, and that "new entrants in the foundry business are likely to initiate a trend of competitive pricing and create potential overcapacity in legacy technology" [29]. That is the backdrop to ASP declines of 5.0% in 2024 and 5.4% in 2025, which the company describes as nominal price erosion [25].

UMC's differentiation comes from specialty processes and a multi-country footprint, but the available comparison cannot quantify its edge over peers. Management says specialty processes now make up about 50% of revenue, that 2026 AI-related revenue from power management, connectivity, FPGA, advanced packaging and silicon photonics should be about US$300 million, and that this exposure should exceed US$1 billion within three years [30]; in July the company also announced its first mass-production delivery of 12-inch silicon photonics chips to a customer and plans to open its silicon photonics platform to general customers in 2027 [16]. Management acknowledged that consumer demand in handsets, PCs and notebooks will decline year over year, but said UMC's 2026 wafer shipments will still grow on its own foundry share gains and its customers' share gains [31]. The available material contains no peer pricing or utilization data by node and by quarter, so how much of UMC's share gain comes from technical differentiation versus orders shifting from rivals rests for now on the company's own account.

Core Debates

With utilization heading above 90%, can UMC's gross margin get back to the mid-30s?

This question matters because UMC is a classic asset-heavy foundry, where a change in utilization alters gross margin almost within the same quarter. In 2025 about 70% of manufacturing costs were fixed or semi-fixed, utilization was only 75.2% and ASP fell 5.4%, and gross margin dropped from 32.6% to 29.0% [25][20]. In the second quarter of 2026 utilization rose to 85% and gross margin recovered to 32.5%, and the company guided third-quarter utilization above 90% and gross margin in the mid-30% range [2][6]. This is UMC's first approach to full loading in several years, but depreciation is accelerating at the same time, so this quarter tests how much of the utilization gain actually survives.

Current evidence shows the shipment rebound has already happened, but the offsetting force from depreciation is just as clear. Second-quarter shipments were 1.13 million 12-inch equivalent wafers, up 10.6% sequentially, utilization rose from 79% to 85%, and gross margin was 32.5%, above 28.7% a year earlier [9][16]. The CEO said the 8-inch portfolio is seeing "a strong rebound," with demand for power management ICs, sensors and microcontrollers expected to lift 8-inch utilization significantly in the third quarter [16], and on the call he added that 8-inch loading should reach about 85% while 12-inch runs above the corporate average [32]; third-quarter revenue growth of 8.0% broadly matches guidance for high-single-digit shipment growth with firm US-dollar ASP [5]. The alternative reading is that second-quarter depreciation had already risen 12.6% year over year to NT$15.51 billion [8], and the CFO said new cleanrooms and fab shells will keep depreciation rising [27]; if utilization hits target while gross margin stalls around 33%, the gain is mainly being absorbed by depreciation rather than held back by weak loading.

The numeric baselines and transmission path for this debate are clear. The baselines are second-quarter utilization of 85% (79% in the first quarter and 75.2% for 2025), shipments of 1.13 million wafers, gross margin of 32.5% (29.2% in the first quarter and 28.7% a year earlier), and quarterly depreciation of NT$15.51 billion (NT$13.77 billion a year earlier) [9][2][8]. The chain runs from recovering 8-inch orders for power management, sensors and microcontrollers, to a high-single-digit sequential rise in shipments, to utilization moving from 85% to above 90%, to roughly 70% of fixed manufacturing cost being spread over more wafers, and finally to a higher gross margin; the opposing force is depreciation growth from new capacity and equipment entering service, which goes directly into operating costs [17][22].

What remains unresolved is the relative size of the two forces and whether third-quarter shipments can last. Readers can watch whether third-quarter utilization reaches above 90% and what the 8-inch level is, where gross margin lands between 34% and 36%, how the company splits the effects of utilization, ASP, exchange rates and depreciation, whether quarterly depreciation grows faster than revenue, and whether fourth-quarter shipment and utilization guidance shows a seasonal pullback [26]. If utilization hits target but gross margin stays below 34%, rising depreciation has eaten the utilization gain; if third-quarter shipments turn out to be customer restocking and fourth-quarter utilization falls noticeably, this quarter's margin improvement cannot be treated as a new starting point [31].

After two straight years of price cuts, has UMC's pricing really stopped falling?

ASP is the most sensitive variable in UMC's profit because, with most costs fixed, any price change falls almost entirely to gross profit. ASP fell 5.0% in 2024 and 5.4% in 2025, the main reason 2025 revenue grew only 2.3% while gross margin dropped 3.6 percentage points [25][20]. In 2026 the company's tone shifted clearly: blended ASP rose by a low-single-digit percentage sequentially in the second quarter, third-quarter guidance calls for firm ASP, and management said the annual pricing trend is better than it had anticipated and that it sees a more meaningful price uplift in 2027 [9][6][33]. If the end of price erosion is real, UMC's revenue growth no longer depends on shipments alone.

Existing evidence supports an improving mix but does not yet prove that quoted prices themselves are rising. Second-quarter wafer revenue was NT$66.44 billion, up 17.6% year over year [14]; 22/28nm made up 37% of revenue and 22nm alone 17.5%, both records [16], compared with 36.8% of foundry revenue from 28nm and below for full-year 2025 [15]. The CEO said the company is "not trying to maximize the short-term pricing based on the market cycle" and instead prices on its differentiated technology, and described current price discussions with customers as constructive [33]. The alternative reading is that the higher ASP comes mainly from the 22/28nm and 8-inch mix rather than from price increases: second-quarter shipments rose 10.6% sequentially and revenue 12.6%, a small gap, and third-quarter revenue growth of 8.0% is also close to high-single-digit shipment guidance [2][3][4], so the price contribution remains limited for now.

The baselines for this debate are mainly range language and node shares, because the company does not disclose absolute ASP. They are a low-single-digit sequential rise in blended US-dollar ASP in the second quarter (versus a 5.4% decline for full-year 2025), a 37% share for 22/28nm, a 17.5% share for 22nm, and second-quarter wafer revenue of NT$66.44 billion (NT$56.49 billion a year earlier) [9][14]. The chain runs from improving demand and tighter capacity to selective price increases and new orders and new capacity priced at new levels, which keeps blended US-dollar ASP firm or rising; at the same time a rising 22/28nm share lifts the blended ASP, and with fixed costs unchanged, the ASP gain flows almost entirely into gross profit [25][29].

What remains unresolved is whether the price improvement can move from a mix effect to a quoted-price effect, which will only become clear when the company states the third-quarter ASP direction. Readers can watch whether third-quarter blended US-dollar ASP held, rose or fell, whether 22/28nm and 22nm shares set new records, and what management says specifically about pricing in the fourth quarter and in 2027 [33]. If third-quarter ASP falls and revenue growth relies on shipments alone, or the company withdraws its talk of a 2027 increase, the view that prices have stopped falling does not hold.

With capex raised and convertible bonds issued, can UMC's cash flow fund the expansion?

This question sets the starting point for UMC's gross margin after 2027, because capital spending leaves as cash first and becomes depreciation and capacity years later. Over the past three years UMC held capital spending down, at about NT$94.09 billion in 2023 and NT$91.05 billion in 2024 before dropping to NT$50.56 billion in 2025 [28], while 2025 operating cash flow of NT$99.86 billion covered it comfortably [21]. In 2026 the direction reversed: the full-year budget was raised to US$2 billion [6], and the board separately approved a phased expansion of about NT$148.68 billion for a cleanroom build-out at Singapore P4 and a new fab shell in Tainan, mainly for silicon photonics and advanced packaging [10]; on the call management said this roughly US$5 billion would be spent across 2026 and 2027 [32].

Current evidence shows the higher budget has not yet reached cash flow, while the financing is already in place. Operating cash flow in the first half of 2026 was NT$55.68 billion, up 21.2%, and cash spent on property, plant and equipment was NT$21.35 billion, roughly flat year over year [12]. Singapore P3 has a design capacity of 30,000 wafers per month on 22/28nm and was scheduled to begin production in the second half of 2026 [28], and the CFO said Singapore quarterly capacity would rise to 192,000 12-inch equivalent wafers in the third quarter [9]. On funding, the company set a conversion price of NT$146.0 per share for its first domestic unsecured convertible bond at the end of July [34], the board resolved in August to issue up to US$1.8 billion of overseas convertible bonds to purchase machinery, equipment and facilities [35], and the bonds were issued on October 13 with a 0% coupon [11]. The alternative reading is that the company has ample cash (about NT$124.7 billion at the end of the second quarter) [9] and that zero-coupon convertibles are low-cost financing taken while the share price is high rather than a sign of weak operating cash flow; telling the two apart requires the actual gap between second-half capital spending and operating cash flow.

The baselines for this debate are the budget, capacity and cumulative figures for the two cash flows. They include the 2026 capital budget of US$2.0 billion (raised from US$1.5 billion on July 29), Fab 12i's 2025 capacity of 684 thousand wafers (about 171 thousand per quarter), first-half cash spent on property, plant and equipment of NT$21.35 billion (NT$21.70 billion a year earlier and NT$47.75 billion for 2025), and first-half operating cash flow of NT$55.68 billion (NT$45.92 billion a year earlier and NT$99.86 billion for 2025) [6][17][12][21]. The chain runs from a higher capital budget to more cash spent on property, plant and equipment in investing cash flow; once Singapore P3 and P4 enter service, capacity and depreciation rise together, shipments grow if demand keeps up, and depreciation weighs on gross margin if utilization falls short, with any funding gap filled by operating cash flow and the zero-coupon convertibles [23][30].

What remains unresolved is whether the timing of expansion will match the timing of demand, and whether the financing will turn into equity dilution. Readers can watch whether Singapore capacity reaches 192,000 wafers in the third quarter and whether P3 starts production, whether nine-month cash spent on property, plant and equipment accelerates noticeably, whether operating cash flow continues to cover capital spending, and what management says about 2027 capital spending and the depreciation curve [32][27]. If expansion runs ahead of demand and utilization falls when new capacity starts up, or convertible bond conversion visibly dilutes EPS, the view that cash flow can fund the expansion has to be discounted.

Risks and Falsifiers

The first risk is that net income depends heavily on investment gains and is detached from the foundry business. In the second quarter of 2026 operating income was NT$14.95 billion, but non-operating income totaled NT$30.24 billion, of which equity-method associates, mainly investment companies holding listed shares, contributed NT$23.61 billion, lifting quarterly net income to NT$42.22 billion [24][19]. The exposure sits in net income and EPS: if Taiwan stocks fall, non-operating gains could shrink sharply or turn into losses, leaving EPS far below the second quarter's NT$3.39. If third-quarter non-operating income falls back toward the 2025 full-year level of NT$4.13 billion while operating income keeps rising and becomes the main source of net income, this risk would be resolved [20].

The second risk is customer and regional concentration. In 2025 the top ten customers accounted for 57.0% of revenue and the largest for 11.8%, and by customer headquarters Taiwan was 38.6%, the US 22.0% and China including Hong Kong 15.8% [18][13]. Inventory adjustments or design moves by a few customers would directly change shipments and utilization and therefore revenue and gross profit; second-quarter revenue from China-based customers rose 41.0% year over year, which also makes UMC more sensitive to local mature-node competition and trade policy [14]. If the next annual report shows lower shares for the top ten customers and the largest customer, this risk would ease.

The third risk is elevated customer inventory, meaning part of third-quarter shipment growth may reflect early stocking. The CEO said customer days of inventory rose slightly in the first quarter, smartphone and consumer inventories rose together, and automotive and industrial demand was stable but inventory remained above historical averages [31]. If fourth-quarter orders fall and utilization drops from above 90% back below 85%, the fixed-cost leverage reverses and gross profit and gross margin fall together. If fourth-quarter shipment and utilization guidance is flat or higher and the company says customer inventory has returned to normal, this concern does not hold [26].

The fourth risk is mature-node overcapacity. The 20-F says new entrants and capacity additions by peers could create overcapacity in mature nodes and trigger price competition, which is the backdrop to the 2024-2025 ASP declines [29]. If peers release new capacity while demand slows, UMC could be forced to cut prices or accept lower utilization, putting revenue and gross profit under pressure at once. If the company keeps ASP firm or rising in the third and fourth quarters and the 22/28nm share keeps climbing, this risk has not materialized in the near term [25].

The fifth risk is a mismatch between expansion and demand. Lead time for new cleanrooms and fab shells exceeds 20 months, pushing large-scale silicon photonics and advanced packaging capacity into 2028-2029 [32], while UMC's AI-related revenue in 2026 is only about US$300 million [30]. If demand has not caught up when new capacity enters service, depreciation rises while utilization falls and gross profit comes under pressure; higher cash spending on property, plant and equipment also squeezes free cash flow. If later reports disclose customer commitments or long-term contracts for the new capacity, and AI-related revenue grows by more than US$300 million a year, this risk would fall markedly [10].

What to Watch Next

• Utilization and gross margin: watch whether third-quarter utilization reaches above 90% (85% in the second quarter, 79% in the first) and 8-inch reaches about 85%; hitting target with gross margin at 34%-36% confirms the case, while hitting target with gross margin below 34% falsifies it.
• Depreciation: second-quarter depreciation was NT$15.51 billion, up 12.6%; depreciation growing persistently faster than revenue would show it absorbing the utilization gain.
• Fourth-quarter guidance: third-quarter revenue was NT$74.24 billion, up 8.0% sequentially; flat or higher shipment and utilization guidance confirms the case, while a clear decline points to restocking in the third quarter.
• Pricing: blended US-dollar ASP rose by a low-single-digit percentage in the second quarter; firm or rising third-quarter ASP confirms the end of erosion, while a decline with revenue driven only by shipments falsifies it.
• Node mix and 2027 pricing: 22/28nm was 37% and 22nm 17.5%; a falling share would show a fading mix benefit, and a withdrawal of the 2027 price-increase language would falsify the pricing case.
• Expansion and cash: Singapore capacity was about 171 thousand wafers per quarter in 2025; watch for 192,000 and the P3 start-up, and whether first-half PP&E spending of NT$21.35 billion accelerates while operating cash flow of NT$55.68 billion keeps covering it.
• Net income structure: non-operating income was NT$30.24 billion in the second quarter versus NT$4.13 billion for all of 2025; operating income becoming the main driver of net income would show earnings led by the core business.

Conclusion

UMC's business is driven by three variables: utilization sets how far roughly 70% of fixed costs are spread, US-dollar ASP sets how much revenue converts directly into gross profit, and capital spending sets depreciation and capacity several years out. In the second quarter of 2026 utilization was 85%, gross margin 32.5% and operating income NT$14.95 billion; third-quarter revenue is already known at NT$74.24 billion, and the company has guided utilization above 90% and gross margin in the mid-30% range [2][5][6]. The central unresolved relationship is how much of the utilization gain stays in gross margin while quarterly depreciation grows 12.6% year over year and the expansion budget and convertible financing advance together, and the second quarter's net income came mainly from investment gains rather than the core business [8][24].

After the second-quarter results, two Commercial Times (Taiwan) reports cited the views of two US brokers, which agree on direction but stress different forces. One US broker cited by Commercial Times on July 30 argued that although consumer electronics demand remains uncertain, better utilization and a more positive pricing trend should support UMC's earnings growth, raised its 2026-2028 earnings estimates on that basis, and noted that higher depreciation from new capacity will limit long-term gross margin improvement [36]; this view acknowledges both the utilization benefit and the depreciation drag, matching the two forces in the first debate, but does not quantify when or by how much depreciation will offset the gain. Another US broker cited by Commercial Times on August 6 argued that the better-than-expected sequential rise in third-quarter shipments came from UMC's share gains and a growing contribution from AI-related business rather than a broad recovery in end demand, and that the higher capex also helps ease market concern about weak mature-node demand [37]; if that reading holds, the risk of a fourth-quarter utilization pullback is smaller, but it offers no testable figures on pricing or depreciation. Both are outside interpretations and cannot substitute for the breakdown in the third-quarter report itself.

From here, if third-quarter utilization reaches above 90%, gross margin enters the 34%-36% range, the company confirms firm or rising US-dollar ASP, fourth-quarter guidance shows no clear pullback, and operating cash flow keeps covering accelerating capital spending, the view that utilization and pricing are improving together would be clearly strengthened. Conversely, if utilization hits target but gross margin stalls around 33%, ASP falls or the 2027 price-increase language is withdrawn, fourth-quarter shipments decline, or operating income cannot carry net income once non-operating gains recede, the current understanding of UMC's earnings recovery would be clearly weakened.

Sources

[1] UMC 6-K 2026-10-13 — Q3 2026 investor conference notice · 2026-10-13 · 6-K · https://www.sec.gov/Archives/edgar/data/1033767/000119312526419602/

[2] UMC 6-K 2026-07-29 — second quarter 2026 results · 2026-07-29 · 6-K · https://www.sec.gov/Archives/edgar/data/1033767/000119312526322027/umc-ex99.htm

[3] UMC 6-K 2026-08-06 — July 2026 revenue report · 2026-08-06 · 6-K · https://www.sec.gov/Archives/edgar/data/1033767/000119312526336498/

[4] UMC 6-K 2026-09-04 — August 2026 revenue report · 2026-09-04 · 6-K · https://www.sec.gov/Archives/edgar/data/1033767/000119312526382689/

[5] UMC 6-K 2026-10-06 — September 2026 revenue report · 2026-10-06 · 6-K · https://www.sec.gov/Archives/edgar/data/1033767/000119312526414888/

[6] UMC 6-K 2026-07-29 — third quarter 2026 guidance · 2026-07-29 · 6-K · https://www.sec.gov/Archives/edgar/data/1033767/000119312526322027/umc-ex99.htm

[7] ChartRow — UMC next-quarter EPS estimate (page as of 2026-10-19) · 2026-10-19 · ChartRow · https://chartrow.com/quote/umc/earnings

[8] UMC 6-K 2026-07-29 — interim report, quarterly depreciation · 2026-07-29 · 6-K · https://www.sec.gov/Archives/edgar/data/1033767/000119312526322030/umc-ex99_1.htm

[9] UMC Q2 2026 earnings call — utilization, shipments, ASP and node mix · 2026-07-29 · earnings-call · https://www.umc.com/en/IR/events

[10] UMC 6-K 2026-07-29 — Singapore P4 and Tainan fab expansion approval · 2026-07-29 · 6-K · https://www.sec.gov/Archives/edgar/data/1033767/000119312526322027/umc-ex99.htm

[11] UMC 6-K 2026-10-05 — pricing of the 7th overseas convertible bonds · 2026-10-05 · 6-K · https://www.sec.gov/Archives/edgar/data/1033767/000119312526413371/

[12] UMC 6-K 2026-07-29 — interim report, first-half cash flow · 2026-07-29 · 6-K · https://www.sec.gov/Archives/edgar/data/1033767/000119312526322030/umc-ex99_1.htm

[13] UMC 20-F FY2025 — customers and revenue by region · 2026-04-30 · 20-F · https://www.sec.gov/Archives/edgar/data/1033767/000119312526193757/d91630d20f.htm

[14] UMC 6-K 2026-07-29 — interim report, revenue by product and geography · 2026-07-29 · 6-K · https://www.sec.gov/Archives/edgar/data/1033767/000119312526322030/umc-ex99_1.htm

[15] UMC 20-F FY2025 — wafer sales by process technology · 2026-04-30 · 20-F · https://www.sec.gov/Archives/edgar/data/1033767/000119312526193757/d91630d20f.htm

[16] UMC 6-K 2026-07-29 — CEO statement on second-quarter shipments and third-quarter demand · 2026-07-29 · 6-K · https://www.sec.gov/Archives/edgar/data/1033767/000119312526322027/umc-ex99.htm

[17] UMC 20-F FY2025 — fab capacity, output and utilization · 2026-04-30 · 20-F · https://www.sec.gov/Archives/edgar/data/1033767/000119312526193757/d91630d20f.htm

[18] UMC 20-F FY2025 — customer concentration risk · 2026-04-30 · 20-F · https://www.sec.gov/Archives/edgar/data/1033767/000119312526193757/d91630d20f.htm

[19] UMC 6-K 2026-07-29 — interim report, equity-method investees · 2026-07-29 · 6-K · https://www.sec.gov/Archives/edgar/data/1033767/000119312526322030/umc-ex99_1.htm

[20] UMC 20-F FY2025 — operating revenues, gross margin and operating margin · 2026-04-30 · 20-F · https://www.sec.gov/Archives/edgar/data/1033767/000119312526193757/d91630d20f.htm

[21] UMC 20-F FY2025 — operating and investing cash flow · 2026-04-30 · 20-F · https://www.sec.gov/Archives/edgar/data/1033767/000119312526193757/d91630d20f.htm

[22] UMC 20-F FY2025 — depreciation expense · 2026-04-30 · 20-F · https://www.sec.gov/Archives/edgar/data/1033767/000119312526193757/d91630d20f.htm

[23] UMC 20-F FY2025 — cash and liquidity · 2026-04-30 · 20-F · https://www.sec.gov/Archives/edgar/data/1033767/000119312526193757/d91630d20f.htm

[24] UMC 6-K 2026-07-29 — interim report, consolidated statements of comprehensive income · 2026-07-29 · 6-K · https://www.sec.gov/Archives/edgar/data/1033767/000119312526322030/umc-ex99_1.htm

[25] UMC 20-F FY2025 — pricing and capacity utilization · 2026-04-30 · 20-F · https://www.sec.gov/Archives/edgar/data/1033767/000119312526193757/d91630d20f.htm

[26] UMC 20-F FY2025 — seasonality · 2026-04-30 · 20-F · https://www.sec.gov/Archives/edgar/data/1033767/000119312526193757/d91630d20f.htm

[27] UMC Q2 2026 earnings call — gross margin drivers and depreciation curve · 2026-07-29 · earnings-call · https://www.umc.com/en/IR/events

[28] UMC 20-F FY2025 — capital expenditures and Fab 12i P3 · 2026-04-30 · 20-F · https://www.sec.gov/Archives/edgar/data/1033767/000119312526193757/d91630d20f.htm

[29] UMC 20-F FY2025 — competition and legacy-node overcapacity · 2026-04-30 · 20-F · https://www.sec.gov/Archives/edgar/data/1033767/000119312526193757/d91630d20f.htm

[30] UMC Q2 2026 earnings call — AI-related revenue and specialty share · 2026-07-29 · earnings-call · https://www.umc.com/en/IR/events

[31] UMC Q2 2026 earnings call — consumer weakness, share gain and customer inventory · 2026-07-29 · earnings-call · https://www.umc.com/en/IR/events

[32] UMC Q2 2026 earnings call — capex plan, lead times and 8-inch loading · 2026-07-29 · earnings-call · https://www.umc.com/en/IR/events

[33] UMC Q2 2026 earnings call — pricing environment and 2027 uplift · 2026-07-29 · earnings-call · https://www.umc.com/en/IR/events

[34] UMC 6-K 2026-07-30 — conversion price of the 1st domestic unsecured convertible bond · 2026-07-30 · 6-K · https://www.sec.gov/Archives/edgar/data/1033767/000119312526324785/umc-ex99.htm

[35] UMC 6-K 2026-08-26 — board resolution on the 7th overseas convertible bonds · 2026-08-26 · 6-K · https://www.sec.gov/Archives/edgar/data/1033767/000119312526365826/

[36] 工商時報 2026-07-30 —《半導體》聯電強彈攻頂 外資按讚升價 · 2026-07-30 · 工商時報(引述一家美系外资) · https://www.ctee.com.tw/news/20260730701200-430201

[37] 工商時報 2026-08-06 —《半導體》聯電7月營收近45月高 攜前7月齊寫同期次佳 · 2026-08-06 · 工商時報(引述另一家美系外资) · https://www.ctee.com.tw/news/20260806701466-430201