
Amkor Technology
Contract manufacturer: outsourced semiconductor assembly, advanced packaging (flip-chip, wafer-level, SiP, 2.5D) and final test sold to fabless firms, IDMs and foundries. Capital-intensive, capacity-utilization-driven, low-margin volume services business; revenue concentrated in a handful of large accounts.
- 2026-08-04This market capitalisation previously read ~$17.3B (as of 2026-07-02). Restated to ~$12.63B on this refresh, roughly 27% lower.
- 2026-08-04This share price previously read $69.65 (as of 2026-07-02). Restated to $50.84 on this refresh, roughly 27% lower.
Sources — 10 figures with citations
- Q2 2026 net sales, gross profit, operating income, net income, EPSfiled2026-06-30Net sales $1,897.965M (+25.6% YoY); gross profit $318.590M (16.8%); operating income $199.854M (10.5%); net income attributable to Amkor $173.753M; diluted EPS $0.70; EBITDA $400Msec.gov — Consolidated Statements of Income and Quarterly Financial Results table, Form 8-K Ex-99.1 filed 2026-07-27. Q2 2025 comparatives: $1,511.392M sales, $181.897M gross profit (12.0%), $91.967M operating income (6.1%), $0.22 diluted EPS
- Gross-margin trajectory and cost-of-sales mixfiled2026-06-30Gross margin 12.0% (Q2 2025) -> 14.2% (Q1 2026) -> 16.8% (Q2 2026); Q3 2026 guided 18.5%-19.5%. Depreciation fell to 8.6% of sales from 9.6%, labor to 10.0% from 11.7%, other manufacturing to 12.0% from 13.8%; materials 52.6% vs 52.9%sec.gov — Selected Operating Data - Gross Margin Data table, plus the Business Outlook section for Q3 guidance
- H1 2026 cash flow, capex and free cash flowderived2026-06-30Operating cash flow $381.573M; payments for property, plant and equipment $688.425M; implied free cash flow -$306.852M (-8.6% of H1 revenue)sec.gov — Condensed Consolidated Statements of Cash Flows (both inputs filed). FCF = 381.573 - 688.425 = -306.852; margin = -306.852 / 3,582.666 = -8.57%. Capex intensity = 688.425 / 3,582.666 = 19.2%
- Balance sheet — cash, investments, debt, net cash, accrued capexderived2026-06-30Cash $1,551.890M + short-term investments $960.272M = $2,512.162M; total debt $150.781M + $2,334.950M = $2,485.731M; net cash ~$26.4M. Capital-expenditures payable $621.326M vs $243.543M at Dec 31, 2025. Total equity $4,686.828Msec.gov — All components filed on the Consolidated Balance Sheets. Net cash = 2,512.162 - 2,485.731 = 26.431. Debt at Dec 31, 2025 was 162.430 + 1,282.816 = 1,445.246, so total debt rose ~$1,040M in six months; the cash-flow statement shows $1,150,000K of long-term debt proceeds and $56,350K of capped-call payments
- End-market mix and customer concentration (Q2 2026)filed2026-06-30Communications 42%, Computing 22%, Automotive/industrial/other 22%, Consumer 14%. Top ten customers 66% of net sales (vs 72% in Q2 2025, 68% in Q1 2026). Advanced products $1,557M vs mainstream $341Msec.gov — Selected Operating Data - Net Sales Data and End Market Distribution Data tables
- Q3 2026 and FY2026 guidancefiled2026-07-27Q3 2026 net sales $1.95-$2.05B; gross margin 18.5%-19.5%; net income $180-$205M ($0.72-$0.82 diluted). FY2026 capital expenditures approximately $2.5-$3.0Bsec.gov — Business Outlook section of the Q2 2026 earnings release
- H1 2026 income statementfiled2026-06-30Net sales $3,582.666M (+26.5% vs $2,832.967M); gross profit $557.622M (15.6%); operating income $300.141M (8.4%); diluted EPS $1.03sec.gov — Six-months columns of the Consolidated Statements of Income; margin percentages derived from the filed dollar figures
- Price and market capitalizationmarket2026-08-03$50.84 close; market cap ~$12.63B on 248.45M shares outstanding; 52-wk range $21.87-$96.68stockanalysis.com — Aug 3, 2026 regular-session CLOSE (4:00 PM EDT), not an intraday level. EV ~$12.60B net of ~$26M net cash; ~7.9x annualized Q2 EBITDA ($400M x 4 = $1.6B); ~18.2x annualized Q2 diluted EPS ($0.70 x 4)
- Q2 2026 result vs consensus, and the post-earnings share reactionmarket2026-07-28Diluted EPS $0.70 vs ~$0.48 consensus (~46% beat); shares nonetheless fell sharply on Jul 28, 2026 as the Q3 revenue guide came in light and management flagged a high-single-digit sequential decline in Communicationsseekingalpha.com — Consensus and the share reaction are third-party aggregates, not filed figures. The reported EPS itself is from the 8-K
- Prior-year comparability adjustmentfiled2025-06-30Q2 2025 results included a $32M net benefit to operating income and EBITDA (and $16M / $0.07 per diluted share to net income) from a Nanium acquisition contingency paymentsec.gov — Footnote (2) to the Quarterly Financial Results table. This means the YoY operating-margin expansion is understated on a clean basis
Earnings, margins, COGS & capex
A ~$6.7B-revenue contract packaging/test business with structurally thin margins (gross ~14%, operating ~7%) and heavy, rising capital intensity. FY2025 grew only 6% and gross margin compressed on higher overhead/comp, but Q1 2026 inflected to a record $1.68B (+27% YoY) led by communications (Apple ramp) and advanced packaging. The company is spending aggressively ($2.5-3.0B FY2026 capex, up from ~$950M in FY2025) on the Arizona campus and 2.5D/advanced-packaging capacity, compressing near-term free cash flow in a bet on AI/onshoring demand.
Income statement — where each revenue dollar goes
% of revenueOf every $1 of revenue, ~83¢ is cost of goods and ~6¢ operating expense, leaving ~11¢ of operating profit (~6¢ net).
Revenue trend
Margins
compressed vs 14.8% FY2024; guided 14.5-15.5% Q2 2026
FY2025 flat/slightly up vs 6.9% FY2024
low, cyclical
capex consumes most of it
COGS structure
COGS is dominated by direct materials (substrates, leadframes, gold/copper wire, molding compound), depreciation of packaging/test equipment, and factory labor across Korea, Japan, China, Taiwan, Vietnam, Philippines, Malaysia and Portugal. High fixed-cost base makes gross margin acutely utilization-sensitive; FY2025 gross margin fell partly on higher employee compensation and overhead ahead of volume.
Capex
FY2026 guided $2.5B-$3.0B, a step-up from ~$950M in FY2025 (of which ~$200M was Arizona), mostly facilities expansion + advanced-packaging (2.5D/SiP) capacity and the Arizona campus. This is the swing factor for free cash flow and the central execution risk.
Latest earnings
Large beat on the quarter, but the stock fell hard on the outlook. Diluted EPS $0.70 vs ~$0.48 consensus (~46% beat) on record net sales of $1.90B (+26%) and a 480 bps gross-margin expansion to 16.8%. Shares nonetheless sold off sharply on Jul 28 because the Q3 revenue guide ($1.95-$2.05B) came in light and management flagged a high-single-digit sequential decline in Communications
Q3 2026: net sales $1.95-$2.05B; gross margin 18.5%-19.5% (a further ~200 bps step-up); net income $180-$205M or $0.72-$0.82 per diluted share. FY2026 capital expenditures of approximately $2.5-$3.0B (unchanged)
- Q2 2026 net sales
- $1,897.965M, record, +25.6% YoY
- Q2 2026 gross profit / margin
- $318.6M / 16.8% (vs $181.9M / 12.0% in Q2 2025)
- Q2 2026 operating income / margin
- $199.9M / 10.5% (vs $92.0M / 6.1%)
- Q2 2026 net income / diluted EPS
- $173.8M attributable to Amkor / $0.70 (vs $54.4M / $0.22)
- Q2 2026 EBITDA
- $400M (net income $175M + interest $17M + tax $28M + D&A $180M)
- Cost-of-sales mix improvement
- Depreciation 8.6% of sales (from 9.6%), labor 10.0% (from 11.7%), other manufacturing 12.0% (from 13.8%); materials roughly flat at 52.6%
- End markets (Q2 2026)
- Communications 42%, Computing 22%, Automotive/industrial 22%, Consumer 14%. Record revenue in Computing and Automotive & Industrial
- Customer concentration
- Top ten customers 66% of net sales, DOWN from 72% in Q2 2025 and 68% in Q1 2026
- Advanced vs mainstream
- Advanced products $1,557M (82% of sales) vs mainstream $341M
- Prior-year comparability
- Q2 2025 included a $32M net operating-income benefit (and $16M / $0.07 EPS) from a Nanium acquisition contingency payment, so the YoY margin expansion is understated on a clean basis
Growth drivers
- AI/HPC advanced packaging (2.5D, high-density SiP, wafer-level) demand pull-through
- Apple/communications ramp (smartphone SiP and RF modules) - Q1 2026 communications +42% YoY
- US onshoring — Peoria, Arizona advanced-packaging & test campus (first of its kind in US), aligned to CHIPS Act (up to $400M proposed funding)
- TSMC collaboration - packaging/test adjacent to TSMC's Arizona fabs
- Automotive & industrial content growth (Q1 2026 automotive +28% YoY; cyclically soft near-term)
Reported financials — SEC EDGAR
Audited GAAP figures pulled from SEC filings · latest filing 2026-02-20. The audited primary-source spine — not financial advice.
Revenue — annual (GAAP)
Margins & balance sheet — FY’25
Bull & bear
The margin thesis that OSAT bears said was structurally impossible is actually happening: gross margin went 12.0% -> 14.2% -> 16.8% in three quarters and is guided to 18.5-19.5% in Q3, operating margin nearly doubled to 10.5%, and the stock is now ~47% below its 52-week high at roughly 7.9x annualized EBITDA.
- Gross margin expanded 480 bps YoY to 16.8% and is guided to 18.5%-19.5% in Q3 2026 — that would be roughly 650-750 bps above Q2 2025 and far above the ~14% level the business was stuck at, driven by depreciation, labor and other-manufacturing all falling as a percentage of sales on higher utilization
- Operating margin 10.5% vs 6.1% a year ago, and the YoY comparison is understated: Q2 2025 included a $32M one-time operating benefit from the Nanium contingency
- Record net sales of $1,898.0M, +25.6% YoY, with record revenue in Computing and in Automotive & Industrial — the mix is diversifying away from smartphones rather than depending on them
- Customer concentration is actually improving: top-ten customers fell to 66% of sales from 72% a year ago and 68% last quarter — the single biggest structural bear point on Amkor is easing
- Valuation has reset hard: at $50.84 the stock trades ~7.9x annualized Q2 EBITDA ($400M x 4) and ~18x annualized Q2 EPS — back inside the historical OSAT range even as margins run at record levels
- Advanced products are $1,557M of $1,898M (82%) of sales, so the AI/HPC packaging mix shift is already the dominant revenue base, not a future option
The market punished the quarter for a reason: Q3 revenue guidance was light, Communications (42% of sales) is guided down high-single-digits sequentially, free cash flow is deeply negative at -$307M for the half, and the company has added ~$1.04B of debt in six months to fund a capex program that still has no revenue attached to it until ~2028.
- Free cash flow is negative -$306.9M in H1 2026 (operating cash flow $381.6M vs $688.4M of PP&E payments), and the true burn is worse than cash shows: capital-expenditures payable rose to $621.3M from $243.5M, meaning ~$378M of committed capex has not yet been paid
- The capex program is now debt-funded, not self-funded: total debt rose to ~$2,486M from ~$1,445M at Dec 31, 2025 ($1,150M of long-term debt issued in H1, with $56.4M of capped-call payments indicating a convertible), converting the prior ~$0.5B net-cash cushion into roughly zero
- Q3 guidance is the reason the stock fell despite a ~46% EPS beat: net sales of $1.95-$2.05B was light versus expectations, with a high-single-digit sequential decline flagged in Communications — 42% of Q2 revenue
- Communications at 42% of sales means the smartphone cycle, not AI, is still the swing factor; Consumer also fell to 14% of sales from 18% a year ago
- Peak-margin risk: 18.5-19.5% guided gross margin is a record for this business and depends on utilization staying high — the same operating leverage that took margins up 480 bps reverses just as fast, and the company itself flags 'difficulty achieving the relatively high-capacity utilization rates necessary to realize satisfactory gross margins given our high percentage of fixed costs'
- Amkor has no backlog — the company explicitly lists 'our absence of backlog and the short-term nature of our customers' commitments' as a risk factor — so there is no contracted visibility offsetting the ~$2.5-3.0B FY2026 capex commitment
What it is worth
Market multiples cross-checked vs OSAT peers and history. As of 2026-07-02 close: price $69.65 (after a -12.9% single-day drop), market cap ~$17.3B, EV ~$16.8B (net cash ~$0.5B) on ~$1.16B FY2025 EBITDA (~14-15x EV/EBITDA); P/E ~40x trailing on TTM EPS (~46x on FY2025 EPS of $1.50), ~32-36x forward on the Q2-run-rate guide.
If demand normalizes, Apple orders soften, or Arizona ramps slowly/at low utilization, margins stay ~14% and the multiple compresses toward the historical OSAT range (~6-9x EV/EBITDA), implying material downside from ~$70 given the ~240% run-up and beta ~2.26.
Steady ~high-single-digit revenue growth with gross margin range-bound ~14-15% and heavy capex keeps FCF thin; a ~10-12x EV/EBITDA (mid-cycle OSAT premium for onshoring) implies a valuation below the recent ~$70 print - i.e. the current multiple already discounts a good outcome.
If AI/HPC advanced packaging and Arizona drive mid-teens revenue growth with gross margin expanding toward 16-18% and FCF inflecting once capex normalizes, the current ~14-15x EV/EBITDA is defensible and EPS re-bases materially higher (>$2.50), supporting the ~$80-95 range seen in H1 2026.
Amkor is re-rated well above its historical OSAT range (typically high-single to low-double-digit P/E, ~6-9x EV/EBITDA) on the AI-packaging + onshoring narrative. The multiple embeds a successful Arizona ramp (production ~2028) and sustained advanced-packaging mix shift; at ~14-15x EV/EBITDA it is priced closer to a growth semi than a contract manufacturer. ASE (ASX) trades at a lower multiple on larger scale - a relative-value check that flags AMKR's premium. The -13% single-session move on 2026-07-02 underscores the de-rating sensitivity (beta ~2.26).
SWOT
Strengths
- World's #2 OSAT (~15% share) and the largest US-headquartered player - scale, breadth and a multi-country footprint
- Deep advanced-packaging capability (flip-chip, wafer-level, SiP, 2.5D) - Advanced Products were 82.8% of FY2025 sales (~$5.55B)
- Anchor relationship with Apple plus Nvidia/TSMC alignment on Arizona; long qualification cycles create switching friction
- Solid balance sheet: modest net cash (~$0.5B), ~$4.5B equity, self-funding the capex ramp
Weaknesses
- Structurally thin margins (gross ~14%, operating ~7%) - a price-taking, capital-intensive volume business
- Extreme customer concentration: Apple ~29.8% of FY2025 sales, top 10 ~72%
- High and rising capital intensity ($2.5-3.0B FY2026 capex vs ~$950M FY2025) compresses free cash flow and raises execution risk
- Cyclical demand tied to smartphone units and the broader semi cycle
Opportunities
- AI/HPC advanced-packaging supercycle (2.5D, chiplets, high-density SiP) lifting mix and pricing
- US onshoring premium — Peoria, Arizona campus + CHIPS support positions Amkor as the domestic advanced-packaging partner to TSMC/Apple/Nvidia
- Automotive/industrial content growth over the cycle
- Diversifying away from Apple as HPC/compute and communications customers scale
Threats
- Chinese OSATs (JCET, Tongfu, HT-Tech) gaining share with state support and domestic demand
- Foundry encroachment: TSMC's in-house CoWoS/advanced packaging both partners with and competes against OSATs
- ASE Technology's dominant scale (~2.5x Amkor's revenue) in pricing/capacity
- Semiconductor down-cycle or an Apple order/share shift would hit utilization and margins hard
- Valuation re-rating risk — P/E ~40x TTM (~46x on FY2025 EPS) after a ~240% run leaves little room for a demand or Arizona-execution stumble
Moats, dependencies & bottlenecks
Moats
#2 OSAT (~15% share) with fabs across 8+ countries; scale matters for capacity and cost but ASE is ~2.5x larger and Chinese players are scaling.
2.5D/SiP/wafer-level capability plus multi-year customer qualification cycles create switching friction; but TSMC and ASE hold equal-or-better advanced-packaging IP.
Apple/Nvidia anchoring the Arizona campus and long design-in cycles raise switching costs - but concentration is a double-edged moat.
Only US-headquartered scaled OSAT + first US advanced-packaging campus + up to $400M CHIPS funding - a policy-favored position that's hard to replicate quickly.
Dependencies
Customer concentration ~29.8% of FY2025 net sales; smartphone SiP/RF cycle drives communications segment - a share or unit shift is the single biggest earnings risk.
Partner + adjacent competitor Arizona packaging/test tied to TSMC's fabs and customer flow; TSMC's in-house CoWoS also competes for advanced-packaging volume.
Cyclical end-market Utilization-driven margins swing with smartphone units, HPC/AI capex and auto/industrial demand.
Arizona economics lean on up to $400M proposed CHIPS funding and anchor-customer commitments; policy or funding changes alter the return.
Advanced substrates (Ibiden, Unimicron, AT&S) are a historical bottleneck for advanced packaging.
Advantages
- Largest US-headquartered OSAT - geopolitically favored for onshoring and defense/US-supply-chain work
- Broad advanced-packaging portfolio already at 82.8% of sales
- Anchor customers (Apple, Nvidia) plus TSMC-adjacent Arizona campus
- Net-cash balance sheet funding the capex cycle without dilution
Weaknesses
- Structurally thin, utilization-sensitive margins (gross ~14%, operating ~7%)
- ~30% single-customer (Apple) revenue concentration
- Capex intensity ~35-43% of revenue compressing FCF
- Cyclicality and exposure to smartphone unit trends
- Rich valuation (~40x trailing P/E TTM, ~46x on FY2025 EPS) leaves no margin for error
Bottlenecks
- Advanced-substrate supply (flip-chip/2.5D packaging is substrate-constrained industry-wide)
- Arizona campus ramp — construction (first facility ~mid-2027, production ~2028), workforce, yield and utilization on a first-of-kind US facility
- Free cash flow squeezed while $2.5-3B capex runs ahead of revenue
- Capacity/lead-time on high-density 2.5D packaging to capture AI demand fast enough vs TSMC/ASE
Top signals & trends
Top signals
Demand inflection led by communications (+42%) and automotive (+28%) - Apple ramp plus advanced packaging.
Higher overhead/comp ahead of volume; margins remain structurally low.
Bullish on demand conviction / onshoring; bearish on near-term FCF and execution risk.
AI-packaging re-rating prices in perfect execution; high de-rating risk and volatility on any miss.
Hard-to-replicate onshoring asset if it ramps on time and at yield (production ~2028).
Trends
Value migrates from monolithic scaling to packaging - core Amkor addressable growth.
US/Europe fabs need local advanced-packaging - Amkor's Arizona/Portugal footprint fits.
State-backed, double-digit growth pressuring the incumbents' share and pricing.
Both a partner and a competitor for the highest-value advanced-packaging volume.
Communications (Apple SiP) is large but a mature, cyclical driver.
Ecosystem & competitor graph
Suppliers feed the company; customers pull from it. Line thickness shows the strength of each tie (supply-chain dependency, customer earnings contribution). Hover to isolate a tie.
Die-attach / hybrid-bonding assembly equipment for advanced packaging.
Wire/ball bonders and advanced-packaging assembly tools.
Assembly & packaging equipment (bonding, TCB).
IC substrate supplier (advanced-substrate constraint).
Flip-chip / 2.5D IC substrate supplier - a key industry bottleneck.
Flip-chip / advanced IC substrate supplier - a key industry bottleneck.
Largest customer, ~29.8% of FY2025 net sales (smartphone SiP/RF).
AI/HPC advanced packaging; Arizona campus anchor customer.
Mobile/RF SoC packaging & test.
Networking/compute packaging.
Automotive/industrial power & MCU packaging.
Automotive/industrial packaging & test.
Automotive/industrial/consumer packaging.
World's #1 OSAT (~$18-19B FY2024, ~2.5x Amkor) - larger scale, advanced-packaging leader; the primary Western-listed comp.
China's largest OSAT (~$5B, +19% YoY) - fast-growing, state-supported; context only, not a buy call.
Taiwan memory/logic packaging & test specialist; competes on cost and memory packaging.
China OSAT, AMD packaging partner, gaining share with government backing; context only.
In-house CoWoS/SoIC advanced packaging competes for the highest-value AI volume - simultaneously Amkor's Arizona-adjacent fab partner.