Builds the bonding and inspection machines that glue memory chips into AI hardware stacks.
What stands out
- Valued far above the size of its business
At a glance
Builds the bonding and inspection machines that glue memory chips into AI hardware stacks.
What this company is and how it runs — written from structure, not news.
Hanmi Semiconductor builds the bonding and inspection machines that stack the memory dies inside HBM chips — the high-bandwidth memory that AI training hardware depends on. The machines work because Samsung and SK Hynix feed Hanmi real thermal stress data from their data centers, which Hanmi compiles into bonding parameter databases stored inside the equipment itself, telling each machine exactly how to handle a specific manufacturer's dies under the heat patterns of actual AI workloads. Switching to a competing bonder means discarding those accumulated parameters and restarting a six-to-twelve month requalification process, which memory manufacturers are reluctant to do precisely when HBM supply is tightest and every lost week of yield matters. The same arrangement that makes the equipment hard to displace is also what makes the business fragile: if Samsung or SK Hynix stops supplying that thermal data — by bringing equipment development in-house or qualifying a rival — Hanmi's parameter databases stop updating, and the next HBM generation arrives without the calibration advantage that made the current one sticky.
How does this company make money?
Each HBM packaging system sells for between $2 million and $8 million per unit. On top of those one-time sales, the company collects ongoing revenue from spare parts, consumables, and annual service contracts tied to machines already in the field — that installed-base revenue accounts for 25 to 30 percent of total income.
What makes this company hard to replace?
Swapping to a different equipment supplier requires a 6 to 12 month requalification process before the new machines can run in production — and during that window, yield suffers at exactly the moment when HBM supply is under the most pressure. The bonding parameter databases that have been built up over thousands of hours of testing cannot be transferred to a competing platform; a new supplier starts from zero. Non-Korean suppliers face an additional barrier because technical documentation and on-site service are structured around Korean language and local presence.
What limits this company?
Before this company's equipment can run in full production at Samsung or SK Hynix, it must pass a qualification process that takes 18 to 24 months — and that clock only starts once the memory manufacturer releases the specs for their next chip generation. If Samsung or SK Hynix changes direction mid-cycle, any equipment being tested against the old specs has to be re-engineered, which shrinks the window in which that generation's machines can actually be sold.
What does this company depend on?
The company cannot operate without Samsung and SK Hynix, who supply both the roadmap specifications that define what the equipment must do and the thermal stress data that makes the machines work better than rivals. TSMC and ASE Group are the primary customers who actually buy the equipment. Japanese precision component suppliers provide the bonding head assemblies the machines are built from. The Korean government controls export licenses required to sell advanced packaging equipment to customers in China.
Who depends on this company?
Samsung and SK Hynix rely on this company's 6-side inspection systems to maintain the packaging yield on their HBM production lines — without them, defect rates rise. NVIDIA and AMD depend on OSAT providers like TSMC and ASE Group to attach HBM to their AI chips with precision; if those providers lose access to this equipment, chip supply slows. More broadly, if export restrictions cut off technology transfer, the Korean memory industry's packaging capabilities could fall behind those of Chinese manufacturers.
How does this company scale?
Once a process recipe or equipment design is finished for a given HBM generation, it can be copied across many machines at relatively low cost. What cannot be copied is the human support work: every memory manufacturer requires a dedicated on-site engineering team to tune bonding parameters for their specific chips, and that team cannot serve two customers at once. Growth in unit sales does not reduce the need for that customer-by-customer engineering work.
What external forces can significantly affect this company?
US-China semiconductor export controls already restrict how much of this equipment can be sold to Chinese memory manufacturers, cutting into the potential customer base. When the Korean won strengthens against the US dollar, the equipment becomes more expensive for international buyers, squeezing pricing. Chinese government subsidies are funding domestic competitors in packaging equipment, creating state-backed rivals who do not need to turn a profit in the near term.
Where is this company structurally vulnerable?
If Samsung or SK Hynix decided to stop sharing next-generation thermal stress data — whether by developing equipment in-house, qualifying a Japanese supplier instead, or simply cutting off access — the protocol databases inside these machines would stop being updated. The thermal advantage the equipment holds today does not carry forward to the next HBM architecture automatically; it has to be rebuilt from new data. Without that cooperation, this company would have no faster path to the next generation than any well-funded competitor starting from scratch.
Price is read as structure — trend, levels, range, peak and volatility drawn on the chart. It does not predict where price goes next.
Sign in to view price data.
Sign in3 interpretations currently present — each is a set of fired observations whose alignment reads as one structural pattern. Click an observation to see the numbers behind it.
Screen for these patternsHow is this stock behaving?
Three observations have aligned in the up direction: the Ichimoku-cloud composite is firing on its up-side configuration, the trend-strength composite is in the upper portion of its mapped range, and the volume-weighted-returns sum over the 60-week lookback is net positive.
Three observations describe the present configuration: a high share of the trailing three years' weekly closes were higher than the prior week, the company has reported positive net income in each of the last five annual periods, and the book-value-increase-consistency composite over the trailing 5 years is elevated.
Three observations describe the present configuration: a high share of the trailing year's weekly closes were higher than the prior week, the company has reported positive net income in each of the last three annual periods, and the industry-benchmarked TTM operating cash flow margin is in the upper peer range.
An interpretation is present only while every observation it reads stays fired (score ≥ 70). It describes what the aligned readings show — never a verdict, never a prediction.
What the company actually pays, and whether its own cash supports it.
The reported statements, read against the company's own industry.
8 interpretations currently present — each is a set of fired observations whose alignment reads as one structural pattern. Click an observation to see the numbers behind it.
Screen for these patternsIs this company financially stable?
Three observations have aligned: most-recent-quarter total cash is in the upper portion of its mapped range against most-recent-quarter total debt, EBITDA-to-total-liabilities is in the upper portion of its mapped range, and FCF-to-total-liabilities is in the upper portion of its mapped range.
Three liquidity ratios co-occur in their elevated ranges: current ratio (industry-benchmarked), quick ratio, and cash ratio. The simultaneous firing means coverage is elevated through progressively more liquid asset layers, not concentrated in inventory or receivables.
Three balance-sheet observations co-occur: industry-benchmarked current ratio elevated, industry-benchmarked equity ratio elevated, and total cash at MRQ at least equal to total debt. The configuration describes equity-heavy capital structure with cash covering total debt.
How does this company use capital?
Three observations describe the configuration: operating income margin is elevated, capex intensity (capex / operating cash flow, industry-benchmarked) is high, and EBIT-to-EBITDA is high (small D&A gap). This pattern is consistent with a growing asset base, an asset-light operating profile, or current-period cost capitalization.
Three margin observations have aligned: industry-benchmarked gross profit margin is in the upper peer range, operating income margin is in the upper portion of its mapped range, and industry-benchmarked TTM operating cash flow margin is in the upper peer range.
Three margin observations have aligned: industry-benchmarked gross profit margin is in the upper peer range, operating income margin is in the upper portion of its mapped range, and industry-benchmarked net profit margin is in the upper peer range.
How is this stock valued?
Retained earnings are a large share of total assets; net income was positive in each of the last 5 fiscal years; shareholders' equity is in the upper part of its industry's equity-to-assets range.
Where is this company structurally exposed?
Three price-behavior observations have aligned: the ulcer index (drawdown depth and duration composite) is elevated, current drawdown from peak is significant, and 20-week annualized volatility is in the upper portion of its mapped range.
An interpretation is present only while every observation it reads stays fired (score ≥ 70). It describes what the aligned readings show — never a verdict, never a prediction.
Shared structure with peers — never a ranking.
Structural observations derived from financial data, industry benchmarks, and supply chain position.
Companies that share the same coordination system — how they create, deliver, or capture value.
Companies that share active interpretations — structural patterns currently present in both stocks.
CompanyGraph