Explain It in 30 Seconds
An exposure machine prints a pattern; an inspection system helps determine whether the resulting structures contain defects. At small dimensions, some problems require electron-beam techniques rather than optical inspection alone.
HMI brought electron optics, inspection algorithms and customer application knowledge. ASML wanted to connect that information with modeling, measurement and scanner correction.
The acquisition was therefore a move into better feedback about manufacturing results, not the purchase of another scanner company.
What the Independent Company Built
Electron-beam inspection
HMI was founded in 1998 and developed its first eScan300 system by 2003. It began research in Silicon Valley and later established its headquarters and manufacturing activities in Taiwan. [S32]
Its tools scanned wafer structures with electrons and interpreted the resulting signals. That enabled detailed investigation of defects and pattern behavior that could be difficult to characterize optically.
Applications and analysis
The value did not reside only in the electron source. Imaging, motion control, signal interpretation and software had to identify useful manufacturing problems without overwhelming engineers with irrelevant detections.
HMI's product families served research, process development and production applications. Its expertise was therefore a combination of equipment and knowledge about what customers needed to inspect.
Where HMI Sat in the Industry
HMI sold into semiconductor process control. Foundries and memory manufacturers needed to find problems early enough to protect yield or accelerate a new process.
A process-development team might accept slower, highly detailed inspection while investigating a specific defect. A production team needed enough sampling speed to monitor a running line. Those are related but different requirements.
This explains the industry's sensitivity-throughput trade-off. A tool that sees very small defects may be valuable without inspecting every feature on every production wafer. The correct deployment depends on the manufacturing question.
The Problem ASML Needed to Solve
ASML's exposure, computational and optical-measurement products could influence patterning, but customers also needed detailed information about what had actually formed on a wafer.
A model can predict a shape and an optical measurement can describe alignment, yet neither necessarily explains an individual defect or an electrical-contrast anomaly. Adding electron-beam capability offered another observation channel.
The 2016 announcement described using detailed patterning information to improve models and, in turn, using those models to guide measurement toward relevant features. [S30]
Why ASML Bought HMI
The companies had already collaborated before the transaction. The stated rationale was to integrate complementary technologies more quickly and improve process control at advanced nodes. [S30]
The intended mechanism was a feedback loop. Better inspection could reveal process behavior; models could help decide what to inspect; corrected information could support scanner settings and manufacturing choices.
That is different from assuming that every inspection purchase must be bundled with an ASML scanner. Customer adoption still depended on independent tool performance, throughput, application value and integration with existing fab systems.
The market was also competitive. Owning HMI did not give ASML the same position in all inspection markets that it held in production EUV.
Deal Timeline
June 16, 2016 — Acquisition agreement announced
The announcement valued the cash transaction at approximately TWD 100 billion, then about €2.75 billion. The stated offer was TWD 1,410 per HMI share. [S30]
November 22, 2016 — Acquisition completed
ASML and HMI announced completion of the transaction to acquire all outstanding HMI shares. [S31]
Subsequent development — Integration and multibeam work
The HMI identity remains present in ASML's e-beam portfolio, including inspection and metrology products. The later roadmap includes multibeam systems intended to improve inspection productivity. [S18][S54]
The approximate euro announcement value should not be silently substituted for the accounting consideration measured at closing. Exchange rates and transaction accounting use different dates and conventions.
What Changed After the Acquisition
Inspection became part of a broader patterning portfolio
The acquired tools could be developed alongside scanner, computational and optical-metrology capabilities. That widened ASML's ability to address manufacturing variation.
A measurement could inform an action
The strategic aim was not merely collecting more images. Useful information could support model calibration, process investigation or scanner correction. The commercial value depended on improving decisions, not on generating a larger data set.
Development gained a larger industrial context
HMI could draw on ASML's customer relationships and technical resources, while ASML gained specialist electron-beam knowledge. The combination still needed to solve inspection-specific problems; common ownership did not remove the throughput challenge.
What Capability Survives Today
HMI remains associated with e-beam inspection and measurement in ASML's product portfolio. These products complement optical methods by investigating selected structures and defects at high resolution. [S18]
Multibeam development seeks to inspect more area or more sites in a practical time. It is not equivalent to replacing every optical inspection tool or scanning every atom of every chip.
The acquisition's continuing importance is therefore its role in learning and control: understanding defects well enough to improve the manufacturing process.
Limits and Integration Risks
An inspection system creates value only if its information is relevant and timely. A highly sensitive system can still be uneconomic if it produces too many nuisance detections or cannot sample enough of the process.
Integration must also respect customer workflows and confidential information. A fab may operate equipment from several suppliers and need data to move through an existing process-control environment.
Finally, early acquisition expectations are not the same as realized returns. The actual test is qualified customer use and useful contribution to process learning, rather than the size of the purchase price.
What Most People Misunderstand
HMI makes exposure systems
Its relevant activity is inspection and measurement, not projecting the chip pattern onto the wafer.
Electron-beam inspection replaces optical metrology
The methods provide different balances of speed, resolution and information. They are complementary in a manufacturing control strategy.
Beam count alone measures customer value
Usable throughput also depends on data processing, inspection recipes, defect relevance and the rest of the system.
Common Questions
Why did ASML need a closer view of defects?
Smaller and more complex structures reduced the margin for error. Detailed information could improve models and guide corrective action.
Does HMI's historic business still matter?
Yes. Its specialist knowledge and product lineage help explain why inspection became part of ASML's holistic approach rather than an unrelated diversification.