Hanmi Semiconductor
Capital-equipment OEM: designs and builds back-end assembly/packaging tools (flagship = TC Bonder for HBM die stacking, plus vision placement, EMI shield, laser/sawing systems) sold to memory and OSAT customers. Revenue is lumpy, order-driven, tied to customers' HBM capacity expansion cycles.
Earnings, margins, COGS & capex
FY2025 set a record for revenue (KRW 576.7B, +3.2% YoY) on the HBM3E ramp, though operating profit dipped 1.6% YoY to KRW 251.4B at a 43.6% operating margin - a peak-cycle level but just below the 2024 profit high. FY2026 opened with a steep air pocket: Q1 revenue fell 65.5% YoY and OP 87.9% YoY as customers paused TC-bonder orders during the HBM3E-to-HBM4 transition. Management and sell-side expect a H2 2026 recovery as HBM4 mass production ramps; a KRW 44.2B SK hynix HBM4 order (Jun 8 2026, ~15 TC Bonder 4.5 Griffin units) is the first concrete recovery signal. An earlier management sales target (KRW ~2T / ~$1.5B for 2026, floated mid-2024) looks highly aggressive against the Q1 trough and should be read as ambition, not consensus.
Income statement — where each revenue dollar goes
% of revenueOf every $1 of revenue, ~50¢ is cost of goods and ~0¢ operating expense, leaving ~50¢ of operating profit.
Revenue trend
Margins
peak-cycle high
compressed on volume collapse / fixed-cost deleverage
-1.6% YoY (revenue record; OP just below 2024 peak)
COGS structure
Not separately disclosed. Cost base is precision mechatronics - motion stages, machine-vision optics, bonding heads/thermal control, and skilled assembly labor; heavily fixed, so utilization swings drive the margin whipsaw seen between FY2025 and Q1 2026.
Capex
Elevated: a ~KRW 100B dedicated hybrid-bonder plant in Incheon's Juan complex (completion targeted H1 2027) plus a new Incheon TC-bonder facility to add capacity ahead of HBM4/HBM5.
Latest earnings
Sharp YoY decline; broadly in line with the pre-flagged HBM4-transition pause rather than a clean beat/miss
Management guides a meaningful earnings recovery weighted to H2 2026 as HBM4 volume ramps; the earlier KRW ~2T 2026 sales ambition looks out of reach given Q1
- Q1 2026 revenue
- KRW 50.9B
- Q1 2026 operating profit
- KRW 8.456B
- Global TC-bonder share (2025)
- ~71% (71.2%)
- SK hynix HBM4 order (Jun 8 2026)
- KRW 44.2B (~15 units)
Growth drivers
- HBM4 mass-production ramp at SK hynix / Micron / Samsung (H2 2026 onward) restarting TC-bonder orders
- More HBM layers per stack (12-Hi to 16-Hi and beyond) = more bonders and more bonding steps per wafer
- Molded-underfill (MUF) packaging suitability where Hanmi tools are well-fitted
- New products: TC Bonder 4.5 Griffin (HBM4) and Wide TC Bonder for HBM5/HBM6 (launch targeted H2 2026)
- Longer-term optionality on a Hanmi hybrid bonder (2nd-gen prototype targeted end-2026, dedicated plant H1 2027) to defend against BESI
Bull & bear
Hanmi is the toll-taker on HBM stacking during the biggest memory capex supercycle in history: it owns ~71% of TC bonders, prints 40%+ operating margins at peak, and the current earnings trough is a transition pause - not demand destruction - with HBM4 orders already turning up.
- Dominant ~71% share of a product every HBM cube must pass through; the Q1 collapse is order timing (HBM3E->HBM4 handoff), not lost demand
- HBM4 mass production ramps H2 2026 and beyond; the KRW 44.2B SK hynix order (Jun 8 2026, ~15 Griffin units) is the first concrete restart signal after a five-month gap
- Each HBM generation stacks more dies (12-Hi -> 16-Hi -> higher), raising bonders-per-fab and bonding-steps-per-wafer - content growth on top of unit growth
- New TC Bonder 4.5 Griffin and Wide TC Bonder (HBM5/HBM6) plus a hybrid bonder (plant H1 2027) give a roadmap to defend the platform
- Net-cash, dividend-paying, founder-aligned (chairman Kwak Dong-shin raised stake to 33.61% during the slump), 40%+ peak margins = high-quality compounder if the cycle re-accelerates
TC bonding may be a transitional technology whose share is already eroding: BESI's hybrid bonding is the likely endgame at high stack counts, ASMPT and BESI are actively displacing Hanmi at SK hynix and Micron, and the stock still trades near ~35-40x record-year sales into a 65%-down quarter.
- Hybrid bonding - where BESI leads and Hanmi has no production tool (only a 2nd-gen prototype targeted end-2026) - threatens to structurally shrink the TC-bonder TAM at >=16-Hi HBM
- Customers are diversifying away: SK hynix onboarded ASMPT for HBM4 (7 systems, reported late 2025) and dual-sourced Hanwha Semitech; Micron reportedly moving to BESI as sole HBM4 TCB vendor, replacing Hanmi and Shinkawa
- Hanmi's SK hynix share reportedly falling from near-exclusive supply toward 20-30% in 2026; sub-1-micron HBM4 accuracy reportedly favors ASMPT
- Brutal cyclicality: Q1 2026 revenue -65.5%, OP -87.9% YoY shows how fast the P&L deleverages on an order pause
- ~$15.7B market cap on ~$0.42B record-2025 revenue (~37x sales) into a trough year, plus a Hanwha Semitech patent overhang, leaves no margin for a delayed HBM4 recovery
What it is worth
Thematic HBM-capex multiple; trailing P/S on record FY2025 sales as a sanity anchor (no US listing).
Hybrid bonding advances faster than Hanmi's hybrid tool, ASMPT/BESI keep taking HBM4 share and Micron confirms BESI-only; the TC-bonder TAM plateaus, the trough persists into 2027, and the ~37x-sales premium compresses materially.
HBM4 recovery materializes but share settles at 20-30% at SK hynix as ASMPT/BESI/Hanwha take the rest; revenue recovers off the Q1 trough yet stays below a straight-line to management's ~KRW 2T ambition, and the multiple normalizes as growth is shared.
HBM4 orders ramp hard H2 2026 into 2027, Hanmi holds ~50%+ TC-bonder share and content grows with layer counts; earnings re-approach/exceed the FY2025 peak (~$183M OP) and the premium multiple re-rates on a durable supercycle, re-testing prior highs.
At ~$15.7B market cap (KRW 21.48T, Jul 6 2026) on FY2025 record revenue of ~$0.42B, Hanmi trades at roughly ~37x trailing sales into a trough year (Q1 revenue -65.5% YoY) - a momentum/thematic multiple wholly dependent on the HBM4 order recovery re-accelerating and on TC bonding staying relevant against hybrid bonding. Highly sensitive to HBM4 timing and to any further share loss at SK hynix/Micron. Not financial advice.
SWOT
Strengths
- ~71% global TC-bonder market share (71.2% in 2025); the incumbent standard for HBM die stacking
- Peak-cycle operating margins in the 40%+ range and a net-cash, dividend-paying balance sheet (record KRW 76B FY2025 dividend)
- Deep installed base and process know-how with the leading HBM makers, especially SK hynix
- Tools well-suited to molded-underfill (MUF) flows, which won Micron business (a 50-unit HBM3E order)
Weaknesses
- Extreme revenue lumpiness - a single quarter (Q1 2026) fell 65.5% YoY on an order pause
- Heavy customer concentration in a handful of HBM makers (SK hynix historically the largest)
- No production hybrid bonder yet — the technology most likely to succeed TC bonding at high stack counts (2nd-gen prototype only targeted end-2026)
- Sub-1-micron HBM4 placement-accuracy bar reportedly favors ASMPT, pressuring share
Opportunities
- HBM4 ramp restarting orders in H2 2026 across SK hynix, Micron and Samsung
- Rising bonder intensity per HBM cube as layer counts climb (16-Hi and beyond)
- Wide TC Bonder for HBM5/HBM6 and a hybrid bonder (plant H1 2027) to extend the platform
- New Incheon capacity to capture the next AI-memory capex wave
Threats
- Hybrid bonding (BESI's moat) displacing TC bonding at >=16-Hi, structurally shrinking Hanmi's core TAM
- Active customer diversification — SK hynix onboarding ASMPT and Hanwha Semitech; Micron reportedly moving to BESI as sole HBM4 TCB vendor
- Patent litigation with Hanwha Semitech over TC-bonder IP (mutual suits, trial underway 2026)
- Deep semiconductor-capex cyclicality and single-end-market (HBM) dependence
- Rich valuation leaving little room for a slower-than-hoped HBM4 recovery
Moats, dependencies & bottlenecks
Moats
TC-bonder market-share dominance (~71%) and reference-status installed base at leading HBM makers durable within TC bonding but exposed to the hybrid-bonding transition Share is a snapshot of an incumbent standard, not a structural lock; already eroding at SK hynix/Micron.
placement accuracy, MUF fit) Moderate-strong Real engineering depth, but ASMPT/BESI are matching or exceeding it on sub-1-micron HBM4 accuracy.
Switching / qualification cost (bonders are qualified into a customer's HBM process flow) Slows displacement but does not prevent it - customers are dual-sourcing deliberately.
Founder alignment + net-cash balance sheet funding the next platform Financial resilience to invest through the trough (new Incheon plant, ~KRW 100B hybrid bonder).
Dependencies
Customer concentration Historically the largest customer; now the epicenter of Hanmi's share erosion as ASMPT and Hanwha Semitech are onboarded.
End-market demand Order flow tracks HBM capacity expansion; a pause or digestion hits revenue immediately (see Q1 2026).
Near-term catalyst The entire H2 2026 recovery thesis rests on HBM4 volume arriving on schedule.
Technology roadmap If hybrid bonding wins high-stack HBM sooner, Hanmi's core TAM shrinks before its hybrid tool (plant H1 2027) ships.
Motion stages, optics, thermal and bonding-head components; not disclosed as single-sourced.
Advantages
- Incumbent ~71% TC-bonder share and reference status at the top HBM makers
- 40%+ peak operating margins and a net-cash, dividend-paying balance sheet
- Product breadth for HBM (TC Bonder 4.5 Griffin, Wide TC Bonder) plus adjacent back-end tools
- Founder-led, well-capitalized to invest through the down-leg of the cycle
Weaknesses
- Single-product, single-end-market (HBM TC bonders) concentration
- Severe order-driven earnings volatility (Q1 2026 -65.5% revenue)
- Behind on hybrid bonding, the likely successor technology
- Actively losing share at both SK hynix and Micron to ASMPT and BESI
Bottlenecks
- No production-grade hybrid bonder yet — 2nd-gen prototype targeted end-2026, dedicated plant H1 2027, a roadmap gap versus BESI
- Meeting HBM4 sub-1-micron placement-accuracy specs to retain share
- Ramping new Incheon capacity in time for the next HBM wave
- Resolving the Hanwha Semitech patent dispute without a licensing/injunction overhang
Top signals & trends
Top signals
First concrete sign the HBM4 order cycle is restarting after the Q1 pause; first Griffin/HBM4 order from SK hynix.
Confirms deliberate diversification away from Hanmi at its largest customer.
Would reverse the temporary Micron win and cede a growth customer (Hanmi + Shinkawa displaced).
Insider accumulation / alignment signal amid the earnings trough (additional ~KRW 5B purchase).
De-rating prices in the trough; sets up asymmetric outcomes on HBM4 timing.
Roadmap extension to keep the TC platform relevant into future HBM generations.
Trends
Structural multi-year growth in HBM units and layer counts expands bonder demand.
More dies per stack means more bonding steps and more tools per fab.
BESI-led technology transition risks displacing TC bonding - Hanmi's core - over time.
SK hynix (ASMPT, Hanwha Semitech) and Micron (BESI) adding vendors structurally caps Hanmi's share.
Causes near-term order air pockets (Q1 2026) but a larger installed base on the other side.
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.
Precision motion-stage / linear-motor vendors Enable sub-micron placement accuracy; not disclosed as single-sourced.
Alignment and inspection subsystems for bonding heads.
Thermal-control and bonding-head component makers Core to thermocompression process performance.
Historically largest customer and HBM leader; source of both the June 2026 order and the ongoing share erosion.
Recent HBM3/3E win on MUF fit (50-unit order), but reportedly shifting to BESI for HBM4.
HBM maker; historically leans on BESI for some bonding, a diversified account for Hanmi.
Advanced-packaging equipment leader; won HBM4 TC-bonder orders at SK hynix (7 systems) and is Hanmi's most direct share-taker.
Hybrid-bonding leader and leading hybrid die-bonder supplier; reportedly Micron's sole HBM4 TCB choice and the long-term structural threat.
Korean challenger in TC/hybrid bonders, dual-sourced by SK hynix; locked in a mutual patent dispute with Hanmi.
Bonding-equipment veteran pursuing advanced-packaging/TCB and hybrid-bonding roadmaps.
Japanese bonder maker historically at Micron; being displaced alongside Hanmi in some HBM4 sockets.