Theses›Semiconductor manufacturing›Deposition materials and process

Deposition materials and process

Deposition value accrues to the molecule (precursors, inhibitors), never the tool

  • Bear
  • Closing
  • $44B by 2035

The callChips go 3D and selective molecules replace patterning steps, but the incumbents already ship the easy version; the opening is a molecule, and nobody has shown up in it.

D1D2R1R2R3C1foragainst D1D2R1R2R3C1foragainst
Fig. 1The balance. Drivers and catalysts against restraints, each block sized by its weight; catalysts hang dashed because they have not happened yet, sized by their odds. It tips toward the restraints.
  1. D1Chips are built in 3D now, gate-all-around, backside power and stacked devices, and lithography plus etch is running out of room at 2nm and below.
  2. D2A selective molecule is consumed on every wafer and sold to chemical majors that buy what they did not invent, so the material content per wafer outgrows wafer count.
  3. R1The weak version is already consensus and incumbent-owned: ASM runs molybdenum ALD and area-selective deposition in volume at 2nm.
  4. R2Few materials deposit selectively enough and any imperfection is a killer defect, so the strong self-aligned-via version stays out of volume this decade.
  5. R3No venture-backed selective-deposition or precursor company exists, qualification takes 3-5 years, and a listed winner is worth about $1B.
  6. C1A named foundry qualifies an area-selective inhibitor for self-aligned vias at 2nm, turning a lithography and mask step into a molecule.end-2028
Rests onthe beliefs underneath
Supported by · confidence
What would prove itdated tests, open and resolved
20272028202920302031today123456789predictionswatch 20272028202920302031today123456789predictionswatch
Fig. 2Each ring is a dated test: dashed while open, struck through when overdue, ticked or crossed once resolved.
  1. 1By end-2027 at least one venture-backed pure-play area-selective-deposition or specialty-precursor company raises a round >$15M OR reaches foundry production qualification.Dec 2027 · open
  2. 2A selective-deposition or precursor company raising more than $15M or reaching foundry qualification; more than $20M by end-2028.Dec 2027 · open
  3. 3Whether Entegris revenue outgrows wafer count, the public-markets check on materials content per wafer.Dec 2027 · open
  4. 4No venture-backed pure-play area-selective-deposition company (inhibitor/SAM/process IP) raises a Series A >$20M or reaches foundry production qualification by end-2028 — ASD value stays captured by Merck/EMD + AMAT/Lam/ASM + imec.Dec 2028 · open
  5. 5ALD/CVD precursors stay >55% top-5-concentrated through 2029, and the durable seed-scale wins in deposition materials are novel-precursor/specialty-molecule companies that EXIT to a materials major rather than build independent franchises.Dec 2029 · open
  6. 6Meta-call: Europe's deposition sovereignty advantage is in PRECURSORS (Merck/EMD, Air Liquide) and COMPOUND-SEMI EPI (Aixtron ~70% MOCVD), NOT in deposition tools or CMP — no European deposition-TOOL seed that raised >€3M reaches HVM front-end qualification by end-2029 without >€50M follow-on or a specialty pivot.Dec 2029 · open
  7. 7No fully-self-aligned-via (FSAV) ASD scheme reaches broad HVM by end-2030 — selective caps/liner/fill spread, but FSAV-by-dielectric-on-dielectric stays R&D on the missing low-κ material.Dec 2030 · open
  8. 8Direct-write deposition (iCVD/direct-ALD/LCVD) produces NO >$50M-revenue HVM-patterning company by end-2030; the category stays specialty/prototyping/coatings.Dec 2030 · open
  9. 9A foundry disclosing a self-aligned via scheme in a production node, or Merck, Air Liquide or Entegris buying a precursor startup, would change the view.Dec 2030 · open
Expressionslisted, private, technologies
Private · 50 private companies, 14 of them in the US.

Strem Chemicals (Ascensus Specialties) · Singulus Technologies · Sifco · Shincron · SFA Engineering · Semicore Equipment · Rasirc · PVD Products · Plasma-Therm · NuFlare Technology · Mustang Vacuum Systems · Merck KGaA (Electronics division) · Magnetron Sputtering Solutions (MagSol) · Lotus Applied Technology · Kolzer · Ionautics · Intevac · Gencoa · Forge Nano · Encapsulix · Canon Anelva · Bühler Leybold Optics · Angstrom Engineering · Air Liquide Advanced Materials (Voltaix) · Adnano Technologies · AlixLabs · DNF Co. · Evatec · Forge Nano Europe / ALD nanosolutions · GVD Corporation · Hansol Chemical · Hauzer Techno Coating · Jusung Engineering · Kaufman & Robinson · Kokusai Electric · Kurt J. Lesker Company · Leybold · Oerlikon Balzers · Oxford Instruments · Picosun · scia Systems · Soulbrain · Telemark · Tokyo Electron (TEL) · ULVAC · Von Ardenne · Wonik IPS · ATLANT 3D · conformally · epiray

4 more private companies tracked, not shown

Technologies
Who has already bet on itprivate rounds, from public sources

$31M raised by 4 companies, $15M of it in the last three years. Wave building: Money is arriving: recent rounds and tier-1 funds are in, but no runaway leader yet.

  • $30MATLANT 3D
    • Founded 2018
     Series A+, Mar 2025 $15M (West Hill Capital) · Series A, Sep 2022 $15M (West Hill Capital)
  • $1Mconformally
    • Founded 2025
     Pre-Seed $1M
  • ·epiray
    • Founded 2021
     Rounds not itemised
  • ·Encapsulix Series A
What changedthe record, newest first
  • 3 Sep 2026An independent-sounding expert call on a deposition-tool startup produced no new numbers on the two things that decide whether such a tool can ever reach a production fab: how fast it deposits, and how evenly it coats a full-size wafer. The expert did confirm that even coating is still an open problem — which is exactly what stopped the last serious attempt at this tool class, a machine that has been sitting in customer evaluation for nearly three years. The absence of a number after six months of asking is itself the information.
  • 3 Sep 2026Applied Materials is the largest deposition-equipment vendor at roughly 30% of a market around $23B a year, with Lam and Tokyo Electron next and ASM leading atomic layer deposition. An earlier lower share on this page came from a much wider, non-semiconductor definition of deposition and has been retired.
  • 2 Sep 2026A Dutch deposition-tool startup put a real number on the table for the first time: 70C wafer temperature over a five-hour run with no cooling, on a machine whose whole difficulty is keeping a high-power plasma source from destroying itself within minutes. Low-temperature deposition matters because the industry is moving to stacking chips vertically, and heat is what damages the layers underneath. The number is genuinely good. The open question is not physics but time — the incumbent shipped a comparable class of tool to a memory customer in late 2023 and is still in evaluation nearly three years later. The startup says it will be in a fab by early 2028. That is the claim worth watching.
4 earlier
  • 31 Jul 2026A useful calibration on how long new deposition hardware takes to get into a fab: the incumbent shipped a 300mm ion-beam system to a tier-one memory customer in late 2023, and nearly three years later two systems are still in evaluation, with the evaluation extended again. The films are genuinely better (about 20% lower resistivity than conventional sputtering, on exactly the metals the industry is switching to). Better films are not the hard part. Meanwhile the atomic-layer-deposition leader confirmed that selective deposition and molybdenum both entered volume production at the 2nm node — delivered by the incumbent, which is where that value was always going to land.
  • 9 Jul 2026A European direct-write-deposition company walked its own resolution target back from 80nm to a few hundred nanometers on its third call, and still could not name who buys the flagship product or in what volume — which is what "specialty tool, not a manufacturing revolution" looks like in real time. The one thing worth watching: a genuinely clever throughput argument (activate the chemistry in nanosecond pulses, let it grow on its own while the laser moves on) that, if they ever measure it, would be the first real crack in the "direct-write is always too slow for volume" consensus the incumbents hold.
  • 23 Jun 2026The seed-relevant number in deposition is not the headline equipment market, it is the ~$1.7B precursor/inhibitor consumable layer, and a fresh triangulation across three independent reads now brackets it within about 20%. Still no venture-backed pure-play area-selective-deposition or precursor startup raising a Series A; new chemistry keeps exiting to the materials majors rather than spinning out a franchise.
  • 17 Jun 2026An incumbent deposition-equipment view on maskless direct-write: it lacks the throughput for high-volume front-end manufacturing, but ~80nm direct-write could find a real home in advanced packaging, where throughput requirements are lower and the bottleneck is actually binding. The application, not the tool, is the question.