Invisix, an Eindhoven chip-metrology startup spun out of ASML, has closed an oversubscribed €20 million (about $21 million) seed round to bring soft x-ray measurement tools to the world’s most advanced chip lines. The backers include Hitachi Ventures, imec.xpand, Doosan Investment, Transition Ventures and an unnamed tier-one semiconductor manufacturer, all betting that a kind of light no optical tool can produce will let manufacturers see structures buried deep inside the next generation of three-dimensional chips.
For two decades, looking inside a finished device layer meant optical metrology, a business KLA has dominated. The features that now decide whether an advanced chip works sit too small and too deep for visible light to resolve. That gap is the opening Invisix was built to exploit, and the reason a seed-stage hardware company just pulled money from three corporate strategics.
Why Optical Tools Stopped Seeing Inside the Chip
Modern chipmaking is a stacking problem. Each layer of a transistor or memory cell has to be measured and confirmed before the next one is deposited on top, because a defect buried three layers down cannot be reworked later. Metrology, the science of measuring those structures, is the quality gate that keeps yield from collapsing.
Conventional optical scatterometry works by bouncing visible or ultraviolet light off the wafer and reading the reflected pattern. It is fast and non-destructive, which is why it became the workhorse. But as logic moved to gate-all-around (GAA, a transistor design that wraps the gate fully around the current channel) and memory grew into towering vertical stacks, many of the critical features shrank below what that light can cleanly resolve.
The squeeze matters now because the chips driving artificial-intelligence accelerators and high-performance computing are exactly the ones with the most buried, hardest-to-read structures. Slow or blind measurement means slower ramps and lower yields on the most expensive silicon ever made.
- 2 nanometers: the leading-edge logic node entering production, built on gate-all-around transistors that hide the channel from view.
- 200-plus layers: the vertical depth of today’s advanced 3D NAND memory, each layer needing verification before the next.
- 58%: optical’s share of metrology spending in 2026, the segment Invisix intends to chip away at.
When the thing you need to measure is smaller than the wavelength you are measuring it with, you eventually run out of physics. That is the wall optical inspection is now hitting, and it is the same wall lithography hit a generation ago.
Soft X-rays and a Nobel-Winning Light Trick
Invisix borrows the lesson that reshaped lithography: when device dimensions shrink, the wavelength you use to work on them has to shrink too. Lithography answered with extreme ultraviolet (EUV) light. Invisix is applying the same logic to measurement, using soft x-rays whose far shorter wavelength can reach buried nanoscale features that visible light slides right past.
The clever part is how it makes those x-rays. The platform relies on high harmonic generation (HHG, a process where intense laser pulses convert a gas into a source of much higher-frequency light), the same effect at the heart of the 2023 Nobel Prize in Physics for attosecond light pulses. Anne L’Huillier first observed those harmonics in noble gases in 1987; Invisix has turned the phenomenon into a measurement engine.
- A short-pulsed laser is fired into a chamber of noble gas, driving the atoms into a high-energy state.
- The excited atoms emit soft x-rays across many wavelengths at once, not a single line.
- That multi-wavelength beam produces a richer three-dimensional signal than a conventional single-wavelength laser system.
- Proprietary reconstruction algorithms and machine learning turn the scattered signal into a detailed 3D image of the device interior.
Because the beam reads the structure rather than slicing it, the wafer survives the measurement intact. The architecture has also been designed for the throughput that high-volume manufacturing (HVM, production at full factory scale) demands, which is the difference between a lab curiosity and a tool a fab will actually buy.
A Decade of De-Risking Inside ASML
What makes the bet unusual is its starting line. Invisix’s soft x-ray measurement platform is not a fresh napkin sketch; it carries technology that sat in a corporate research program for years before becoming a company.
We are entering the market with technology that has been incubated inside ASML for more than a decade, a level of technical de-risking that is unusual for a seed-stage hardware company and gives our customers a faster path to deployment.
That line came from Christina Porter, PhD, co-founder and chief executive of Invisix, on the financing. It is the spine of the company’s pitch.
What the Lithography Giant Licensed Out
Invisix launched with licensed patents, a portfolio the company puts at more than 100 issued and pending filings and over 1,500 publication citations behind the core science. That is an unusually deep technical moat for a firm at the seed stage.
It also came with the people who built it. Porter and Sietse van der Post, a PhD physicist and Invisix’s co-founder and chief technology officer, both spent years leading soft x-ray scatterometry research at their former employer before spinning the work out.
The Case for Cutting It Loose
The parent company has spent recent years pouring capital into EUV lithography and advanced packaging, its core franchises. A standalone metrology tool, however promising, was adjacent to that mission rather than central to it.
Spinning the program out lets a focused team chase the metrology market directly while the inventors keep equity and momentum. The technology has already been validated through collaborations with Intel and the research institute imec, and the company is running customer demonstrations from a new cleanroom in Eindhoven. The funding pays to grow the team and finish the first commercial system.
Who Wrote the Cheques, and What They Signal
The investor list reads less like a financial syndicate and less like a typical seed round than a who’s-who of strategic chip money.
- Hitachi Ventures, the corporate venture arm of Japan’s Hitachi, which itself sits among the world’s metrology and inspection suppliers. (Hitachi’s corporate venture fund closed a $400 million vehicle last year.)
- imec.xpand, the deep-tech venture fund tied to the Belgian research institute imec, a partner in Invisix’s early validation.
- Doosan Investment, the investment arm of South Korea’s Doosan Group, plus Transition Ventures.
- An unnamed tier-one semiconductor manufacturer taking a strategic stake.
Korean outlets reported the unnamed chipmaker to be Samsung, which has been fighting to lift yields on its 2-nanometer gate-all-around process; Invisix has not confirmed the name. Either way, when a fab buys equity in its measurement supplier, it is signalling that the tool solves a problem the buyer cannot solve with what is already on the market.
The Metrology Incumbents Invisix Has to Crack
The prize is large and concentrated. The global semiconductor metrology and inspection market was worth USD 13.03 billion in 2025 and is forecast to reach roughly $17 billion by 2030, according to the semiconductor metrology and inspection equipment market report from Mordor Intelligence. Process control is dominated by a short list of names, with KLA holding an estimated 40-to-45 percent of metrology equipment alone.
| Company | Core metrology approach | Position |
|---|---|---|
| KLA Corporation | Optical critical-dimension and overlay, e-beam | Process-control leader |
| Applied Materials | E-beam and optical critical-dimension | Diversified challenger |
| Onto Innovation | Optical and packaging-focused metrology | Advanced-packaging specialist |
| Nova | Optical critical-dimension and materials metrology | Dimensional and materials specialist |
| Invisix | Soft x-ray scatterometry, non-destructive 3D | Seed-stage entrant |
These incumbents are not standing still. Onto and Applied have pushed into hybrid bonding and packaging metrology, and KLA keeps extending optical and e-beam tools further down the node roadmap. The moat is real, and it is defended by deep service relationships inside every major fab.
But the vulnerability is specific. Optical’s edge fades exactly where the hardest 3D structures live, and that is the slice Invisix is aiming at rather than the whole market. A challenger does not need to beat KLA everywhere; it needs to own the measurements optical tools can no longer make.
Winning even a thin wedge of buried-feature metrology on leading-edge nodes would be a serious business, because those are the wafers fabs care most about getting right.
What Production Scale Will Demand
The open question is throughput. A measurement that takes minutes in a lab has to take seconds on a production line, repeatably, for years, across thousands of wafers. HHG-based x-ray sources are powerful but complex, and proving they can run at fab speed and fab reliability is the test no slide deck can pass for the company.
Invisix says the system architecture was designed for high-volume manufacturing from the start, and the decade of prior engineering shrinks the gap between demo and deployment. The near-term plan is concrete: finish the first commercial tool, keep running customer demonstrations from the Eindhoven cleanroom, and convert the strategic backers into reference accounts.
If that first system clears the throughput bar inside a real fab, the optical metrology moat develops its first genuine breach at the leading edge and the strategic money looks early. If it stalls in qualification, Invisix becomes another brilliant physics program that could not make the jump to a production tool, and the incumbents keep the gate.
