Atum Works logo

Atum Works

Gigascale Nanomanufacturing

Spring 2025active2025Website
Hard Tech3D PrintingSemiconductors
Sponsored
Documenso logo

Documenso

Open source e-signing

The open source DocuSign alternative. Beautiful, modern, and built for developers.

Learn more →
?

Your Company Here

Sponsor slot available

Want to be listed as a sponsor? Reach thousands of founders and developers.

Report from 15 days ago

What do they actually do

Atum Works has built a nanoscale 3D printer that fabricates semiconductor and microelectronic parts by depositing and patterning materials in three dimensions. Today they operate this as an in‑house service: customers share designs, Atum prints the parts on its system, and ships prototypes or small batches for testing and iteration YC launch, YC company page.

The company has publicly shown a working machine and states they are shipping parts to early partners, with at least one strategic engagement (NVIDIA) mentioned in their launch materials and profile YC launch, YC company page.

Independent coverage suggests practical resolution is around the ~100 nm range, which makes current use most compelling for advanced packaging features, photonics, sensors/MEMS, and 3D interconnect structures rather than cutting‑edge logic nodes Tom’s Hardware, Softonic.

Who are their target customer(s)

  • Semiconductor packaging houses / OSATs: They need denser vertical interconnects and complex 3D structures that are slow and costly to prototype in planar flows; short‑run iterations are bottlenecked by tooling and long lead times YC launch, Tom’s Hardware.
  • Photonics companies (PICs, micro‑optics): Integrating precise 3D waveguides and optics with electronics is expensive and alignment‑sensitive in standard fabs; they need fast, low‑volume iterations to validate designs Tom’s Hardware, Softonic.
  • Sensor and MEMS manufacturers: Custom nanoscale 3D geometries (e.g., microfluidics, MEMS) face long tapeout cycles and high NRE; small batches and rapid design changes are hard to justify in traditional processes YC launch.
  • AI accelerator and specialized chip teams: They want novel vertical stacking or interconnect patterns to test performance/power ideas, but non‑planar prototypes are costly and slow in standard flows; they also need low‑volume parts for experiments YC company page.
  • Corporate R&D and academic labs: They need quick, low‑cost access to nanoscale fabrication for proofs‑of‑concept without multi‑month foundry cycles or large minimum orders YC launch.

How would they acquire their first 10, 50, and 100 customers

  • First 10: Use YC and existing LOIs to run short, scoped pilots with a strategic partner (e.g., NVIDIA‑type), one OSAT, a photonics startup, a MEMS/sensor maker, and 1–2 labs; deliver working prototypes and a brief joint test report to convert to paid follow‑ons.
  • First 50: Standardize a fixed‑price prototype package and quick‑quote form, publish application notes/case studies with concrete cycle‑time and cost data, and convert conference/webinar leads with time‑boxed pilot slots and referral credits.
  • First 100: Layer in service tiers (Standard/Plus/Custom), add reseller/OSAT channel trials and regional demo days, publish throughput/yield metrics, and scale support (FAE + ops) plus a self‑service portal for vetted repeat customers.

What is the rough total addressable market

Top-down context:

Adjacent end markets today: advanced packaging ~$39.6B (2024), photonic ICs ~$14–15B (2024), MEMS ~$16–17B (2024); packaging equipment ~$9B, broader semi equipment ~$150B (2024) Advanced packaging, PICs, MEMS, Packaging equipment, Semi equipment.

Bottom-up calculation:

Combine the three most relevant end markets (~$70.8B) and assume 1–5% is spent on prototyping/small‑volume specialty work suitable for nanoscale 3D printing, yielding an illustrative near‑term SAM of ~$0.7–3.5B/year Advanced packaging, PICs, MEMS.

Assumptions:

  • Only 1–5% of end‑market revenue flows to prototyping, small‑volume production, and related services/tools.
  • Near‑term fit is constrained by ~100 nm‑class resolution and throughput, steering demand to packaging, photonics, and MEMS rather than cutting‑edge logic Tom’s Hardware.
  • Market figures overlap; the calculation is directional and avoids double‑counting by treating the 1–5% as an aggregate prototyping slice across segments.

Who are some of their notable competitors

  • Nanoscribe: Pioneer in two‑photon polymerization micro‑/nano‑printing used for micro‑optics, photonics packaging, and MEMS; strong overlap with Atum’s early use cases and a mature tool/service ecosystem.
  • UpNano: High‑throughput two‑photon printers and a NanoPro production service for scaling micro‑part batches; competes where customers want fast, small‑volume optical/mechanical parts or a service model.
  • RAITH: Electron‑beam and laser direct‑write systems used for nanoscale prototyping in labs and small device teams; an alternative for bespoke patterning without a new additive tool.
  • Thermo Fisher Scientific (FIB‑SEM): Focused‑ion‑beam + SEM platforms enable site‑specific milling/deposition and circuit edits; widely used for nanoprototyping, FA, and low‑volume edits that can substitute for dedicated 3D prints.
  • EV Group (EVG): Supplier of nanoimprint, wafer bonding, and packaging equipment for photonics and advanced packaging; offers replication/processing routes to similar outcomes vs. direct 3D printing.