Atomically Precise Manufacturing: Comprehensive Research Report

Table of Contents

1. Executive Summary
2. Current State of the Art
3. Key Researchers and Labs
4. The Drexler Vision
5. Mechanosynthesis
6. Self-Replicating Assemblers
7. Key Milestones Achieved (2020-2026)
8. Expert Timeline Assessments
9. Funding Landscape
10. Companies Working on APM
11. The Path from Current Nanotech to Matter Compilers
12. Risks and Governance
13. Conclusions

1. Executive Summary

Atomically Precise Manufacturing (APM), the assembly of materials, structures, devices, and products with every atom placed at an exactly specified location, remains one of the most ambitious goals in science and engineering. As of early 2026, APM development is in its early stages, with practical applications confined to specialized domains (silicon quantum devices, DNA nanotechnology, atomically precise catalytic clusters) rather than constituting a unified, general-purpose manufacturing discipline.

The field is advancing along several parallel tracks:

• Scanning probe-based manufacturing (STM/AFM) has matured significantly, with Zyvex Labs achieving sub-nanometer lithography (7.7 nm pitch) surpassing ASML’s best EUV tools, and Silicon Quantum Computing demonstrating an 11-qubit processor with 99.99% gate fidelity built atom-by-atom. SQC is now selling commercial products (Quantum Twins, Watermelon) to real customers including Telstra and the Australian Department of Defence. In November 2025, SQC patterned 250,000 qubit registers in 8 hours, demonstrating industrial-scale throughput for atom-precision fabrication.
• Covalent mechanosynthesis demonstrated for the first time (December 2025): An inverted-mode STM achieved a 96.4% success rate for hydrogen abstraction from silicon, the first experimental demonstration of deterministic covalent mechanosynthesis. The work involved a team of 54 researchers including Ralph Merkle (arXiv:2512.24431). While this is a landmark result, it is limited to hydrogen removal from silicon, not carbon placement on diamond.
• DNA origami has become a reliable platform for atomically precise nanoscale construction, with demonstrated applications in drug delivery, biosensing, and templating for metal oxide nanofabrication.
• Atomically precise metal nanoclusters are finding commercial traction in catalysis, with precisely defined compositions enabling tunable selectivity.
• Synthetic molecular machines (motors, rotors, switches) continue to advance, with light-activated artificial muscles demonstrated in 2025.
• Diamond mechanosynthesis remains primarily theoretical, with computational studies validating specific tooltip reactions but no experimental demonstrations of positional diamond bond formation.
• General-purpose molecular assemblers and matter compilers remain theoretical constructs with no experimental prototype.

The gap between current capabilities and the Drexler/Freitas/Merkle vision of nanofactories producing arbitrary macroscale products atom-by-atom remains vast but is being methodically narrowed from multiple directions.

2. Current State of the Art

2.1 Scanning Probe-Based Atom Manipulation

The most direct approach to APM uses scanning tunneling microscopes (STMs) and atomic force microscopes (AFMs) to position individual atoms on surfaces.

What can be done now:

• Single-atom placement and imaging: STMs can reliably pick up, move, and place individual atoms on crystal surfaces. This was first demonstrated by IBM (Don Eigler) in 1989 with xenon atoms on nickel, and has been refined dramatically since.
• Hydrogen depassivation lithography (HDL): Using STM tips to selectively remove hydrogen atoms from H-passivated silicon surfaces, creating patterns for dopant incorporation. This is the foundation of silicon quantum device fabrication at atomic precision.
• Atomically precise silicon electronics: Phosphorus atoms can be placed at exact lattice sites in silicon to create single-atom transistors, quantum dots, and qubit registers. The February 2024 review in ACS Nano documents the maturity of this approach.
• MEMS-integrated STM: Batch-fabricated MEMS STM nanopositioners have achieved atomic-resolution imaging at higher scan speeds than conventional STMs, opening a path toward parallelization.
• Machine learning integration: ML-assisted techniques now enable real-time prediction of dopant number at qubit sites during STM patterning (published 2024), improving yield and throughput.

Key limitation: STM manipulation is inherently serial and slow. Patterning even a small device can take hours. This is the central bottleneck for scaling APM.

2.2 DNA Origami and Structural DNA Nanotechnology

DNA origami uses the programmable base-pairing of DNA to fold long strands into arbitrary 2D and 3D nanostructures with approximately 3-6 nm resolution, and sub-nanometer precision for specific attachment points.

Current capabilities:

• Arbitrary 2D and 3D shapes at the ~100 nm scale, with precise placement of functional groups, nanoparticles, proteins, and other cargo
• Atomically precise spatial organization of drug payloads for targeted delivery
• Smart nanofluidic systems capable of navigating biological barriers and responding to intracellular cues
• 3D nanofabrication via atomic layer deposition on DNA origami crystal templates (2025), enabling functional metal oxide nanostructures with precisely defined topologies
• Self-assembling and disassembling swarm molecular robots directed by DNA molecular controllers (Tohoku/Kyoto, 2024)
• Integration with synthetic biology for intercellular communication engineering

Key limitation: DNA origami operates in aqueous solution, is limited to ~100 nm feature sizes, and the structures are soft and not mechanically robust. It is atomically precise in the sense of defined connectivity but not in the sense of rigid 3D atomic coordinates.

2.3 Atomically Precise Metal Nanoclusters

A rapidly growing area where clusters of metal atoms (typically gold, silver, copper, platinum) with exact, known compositions (e.g., Au25, Au38, Au144) are synthesized with crystallographically determined structures.

Current capabilities:

• Atomically precise gold, silver, copper, and bimetallic nanoclusters with defined formulas
• Catalytic applications: selective hydrogenation, electrocatalysis, photocatalysis, enzyme-mimicking catalysis
• Low-nuclearity cluster catalysts (2-10 metal atoms) achieving maximum atom efficiency
• Structure-activity relationships can be studied with single-atom resolution
• Beginning to find commercial applications in water decontamination and chemical processing

2.4 Synthetic Molecular Machines

Following the 2016 Nobel Prize (Sauvage, Stoddart, Feringa), the field of synthetic molecular machines has matured considerably.

Current state:

• Molecular motors achieving 10 million rotations per second (Feringa group)
• Light-activated photoactuating artificial muscles from self-assembled molecular switches (Nature Communications, 2025)
• Molecular assemblers that produce specific polymer sequences (demonstrated in Nature Communications, 2020, Leigh group)
• Autonomous chemically-fueled synthetic molecular motors (Leigh group, 2016)
• Two-photon near-infrared sensitized rotary motors
• Programmable molecular robotics systems emerging as forerunners of a new technological era

2.5 Protein Design and Engineering

The 2024 Nobel Prize in Chemistry (David Baker for computational protein design; Hassabis and Jumper for AlphaFold) represents a major milestone for atomically precise structures. Designed proteins are atomically precise molecular machines assembled from amino acid building blocks.

Current capabilities:

• De novo protein design from scratch using AI (“family-wide hallucination”)
• AlphaFold3 predicting structures of protein-nucleic acid-small molecule complexes
• Predictions scoring above 90/100 on precision scales (effectively atomically precise)
• Custom-designed enzymes, binding proteins, self-assembling nanocages
• This represents arguably the most mature form of atomically precise manufacturing, albeit limited to protein-based structures

2.6 Atomically Precise Membranes

Christian Schafmeister (Temple University) has developed spiroligomer-based macrocycles, chemically synthesized, atomically precise membranes one molecule thick. Triangular macrocycles of tunable sizes assemble into membranes with atomically precise pores exhibiting molecular sieving. Applications include portable dialysis devices for blood purification. This work won a Foresight Feynman Prize.

3. Key Researchers and Labs

3.1 Foundational Theorists

3.2 Experimental Leaders

3.3 Key Laboratories and Institutions

4. The Drexler Vision

4.1 Core Concepts from Nanosystems (1992)

Eric Drexler’s vision, laid out rigorously in his MIT doctoral thesis and subsequent book Nanosystems: Molecular Machinery, Manufacturing, and Computation, proposes:

1. Molecular assemblers: Nanoscale devices that can position reactive molecules with atomic precision to build structures atom-by-atom or molecule-by-molecule
2. Mechanosynthesis: Using mechanical force to drive chemical reactions at specific positions, rather than relying on thermally-driven solution-phase chemistry
3. Nanofactories: Hierarchical systems of molecular assemblers that can manufacture macroscale products by coordinating vast numbers of nanoscale operations
4. Diamondoid structures: Using diamond-like carbon (sp3 bonded) as the primary structural material due to its extraordinary strength, stiffness, and chemical stability
5. Mechanical computing: Nanomechanical logic using interlocking rods rather than electronic transistors
6. Self-replication: Assemblers capable of making copies of themselves (later de-emphasized by Drexler himself due to safety concerns)

4.2 What Has Been Validated

• Positional control at atomic scale: STM atom manipulation proves that individual atoms can be reliably placed. This validates the basic premise of positional assembly.
• Molecular machines exist: Both biological (ribosomes, motor proteins) and synthetic (Feringa motors, Leigh assemblers) molecular machines demonstrate that nanoscale mechanical operations are physically possible.
• Computational analysis of mechanosynthesis: Extensive DFT (density functional theory) studies have shown that specific tooltip-mediated reactions on diamond surfaces are energetically favorable. The DCB6Ge tooltip successfully placing C2 dimers on C(110) diamond at room temperature has been computationally validated.
• Megahertz-frequency synthesis: Drexler’s analysis of positional control enabling synthetic steps at megahertz frequencies “has withstood a decade of scientific scrutiny.”
• Stiff molecular structures: Diamondoid molecules have been isolated from petroleum and synthesized, confirming the existence and properties of the proposed structural materials.
• Protein-based molecular machines: Baker’s designed proteins and biological molecular machines (ATP synthase, the ribosome) validate that atomically precise molecular machines can perform complex functions.

4.3 What Remains Theoretical or Undemonstrated

• No experimental diamond mechanosynthesis: No one has experimentally demonstrated positional placement of carbon atoms to build diamondoid structures using a tooltip.
• No molecular assembler prototype: No device exists that can autonomously perform a sequence of mechanosynthetic reactions to build a specified product.
• No nanofactory: The hierarchical manufacturing system remains entirely conceptual.
• Tooltip fabrication: The proposed tooltips (e.g., DCB6Ge) have not been synthesized and mounted on positioning systems.
• Error correction at scale: The question of how to handle defects in billion-atom assemblies remains open.
• Feedstock delivery: No demonstrated system for reliably transporting individual molecules to a mechanosynthetic worksite.

4.4 The Drexler-Smalley Debate

Nobel laureate Richard Smalley (co-discoverer of C60 buckminsterfullerene) challenged Drexler’s vision on two grounds:

• “Fat fingers” problem: Nanoscale manipulators would be too large to precisely control reactions in the confined space around an assembly site
• “Sticky fingers” problem: Atoms and molecules would adhere to the manipulator rather than being placed where desired

Drexler and supporters have argued these objections do not apply to the actual proposed mechanisms (which use precisely designed tooltip chemistry, not literal “fingers”), and that biological molecular machines (which demonstrably work) face and overcome the same challenges. The debate was never fully resolved before Smalley’s death in 2005.

4.5 MSEP.one, Drexler’s 2024 Software Release

In October 2024, Drexler released the Molecular Systems Engineering Platform (MSEP.one), free and open-source molecular design software built on the Godot game engine. Features include:

• Atomistic simulation of molecular machines
• Virtual motors, anchors, and springs for constraining nanomechanical motions
• Point-and-click molecular building
• Plans for quantum chemistry and multi-scale modeling integration
• MIT License, freely available

Drexler frames MSEP as a step toward “generative nanotechnologies”, extending generative AI into the physical world at the nanoscale. This represents the first practical software tool specifically designed for exploring the design space of molecular machines envisioned in Nanosystems.

5. Mechanosynthesis

5.1 Diamond Mechanosynthesis (DMS)

Diamond mechanosynthesis is the formation of covalent carbon bonds using precisely applied mechanical forces via positionally controlled tooltips, typically proposed to operate in ultra-high vacuum (UHV).

Computational progress:

• The Nanofactory Collaboration (Freitas, Merkle, and 23 researchers across 10 organizations in 4 countries) has computationally analyzed numerous tooltip designs and reaction pathways.
• The DCB6Ge tooltip motif has been extensively studied via DFT and shown to successfully place C2 carbon dimers on C(110) diamond surfaces at both 300K and 80K.
• A “minimal toolset” of 9 tooltip types for DMS was published in 2008 (the landmark paper), covering primary, auxiliary, and intermediate tools needed for a complete diamond-building cycle.
• Freitas published “Molecular Workstation Roadmap I” in May 2025, surveying key technologies needed for positionally controlled DMS using a molecular workstation.

Experimental status:

• No experimental demonstration of positional diamond mechanosynthesis exists as of early 2026.
• The closest experimental work involves STM-based manipulation of individual atoms on various surfaces, but not the specific tooltip chemistry proposed for DMS.
• Diamondoids (small diamond-like hydrocarbon cages) have been isolated, synthesized, and studied, confirming some properties assumed in DMS theory.

First experimental covalent mechanosynthesis (December 2025, arXiv:2512.24431):

In December 2025, a team of 54 researchers (including Ralph Merkle) demonstrated the first experimental covalent mechanosynthesis using an inverted-mode STM. Key results:

• 96.4% success rate for single hydrogen atom removal from H-passivated silicon surfaces
• 100% yield for sequential hydrogen abstractions to form dangling bond pairs
• Sub-angstrom positioning precision with zero-bias operation
• Critical innovation: functionalizing both sides of the tunnel junction (both the tip and the surface), rather than just the tip
• This constitutes the first experimental demonstration of deterministic, positionally controlled covalent bond breaking with near-unity yield

Important caveats: This result demonstrates hydrogen abstraction from silicon only, not the carbon-carbon bond formation on diamond surfaces that is central to the Drexler/Freitas/Merkle diamondoid mechanosynthesis vision. Extension to other elements and bond types “is expected” by the authors but remains undemonstrated. The gap between removing hydrogen from silicon and placing carbon on diamond is substantial.

5.2 Patents

CBN Nano Technologies holds 24 patents related to mechanosynthesis, including:

• US Patent 11,708,384 (2023): Systems and methods for mechanosynthesis
• US Patent 11,592,463 (2023): Methods involving bulk chemical preparation of tips
• US Patent 11,180,514 (2021): Various aspects of mechanosynthesis

Robert Freitas holds the first mechanosynthesis patent (US Patent 7,687,146, 2010) for a simple tool for positional diamond mechanosynthesis.

5.3 Near-Term Experimental Pathway

The 2025 Freitas “Molecular Workstation Roadmap” identifies several key technologies needed to bridge from current STM capabilities to a working mechanosynthesis demonstration:

1. Fabrication of atomically precise tooltips with known chemistry
2. UHV positional control at sub-angstrom precision
3. Feedstock presentation mechanisms
4. Reaction verification via in-situ spectroscopy
5. Automation of multi-step synthesis sequences

6. Self-Replicating Assemblers

6.1 Historical Context

The concept of self-replicating machines dates to John von Neumann’s theoretical work in the 1940s-50s. Drexler’s Engines of Creation (1986) popularized the idea of self-replicating molecular assemblers, leading to both excitement and the “grey goo” fear, that uncontrolled self-replicators could consume the biosphere.

Drexler himself later argued that self-replication is neither necessary nor desirable for molecular manufacturing, and that nanofactories (which produce products but not copies of themselves) are a more practical and safer design.

6.2 Current Status

DNA-based self-replication (demonstrated):

• Researchers at the University of Science and Technology of China (2024) achieved precise control over replication and catalytic assembly of DNA-functionalized colloids at room temperature, facilitating large-scale ordered nanomaterials.
• A 2024 study demonstrated rapid sigmoidal self-replication of DNA wireframe nanoassemblies via isothermal ligase chain reaction.
• DNA molecular controllers for autonomous assembly/disassembly of molecular robots were demonstrated (Tohoku/Kyoto, 2024).

General-purpose self-replicating assemblers:

• No experimental demonstration exists or is imminent.
• The concept remains theoretically plausible (biological cells are existence proofs of self-replicating molecular machinery) but extraordinarily difficult to engineer from scratch.
• The primary barrier is the lack of a bootstrapping path: you need an assembler to build an assembler.

6.3 Expert Assessment

Most researchers treat self-replicating assemblers as a long-term theoretical possibility rather than a near-term engineering goal. The field has largely shifted focus to more practical intermediate targets: non-replicating molecular machines, atomically precise fabrication of specific device types (quantum computers, catalysts, membranes), and incremental improvements in positional control.

7. Key Milestones Achieved (2020-2026)

8. Expert Timeline Assessments

8.1 The Uncertainty Problem

There is no expert consensus on APM timelines. Estimates range from “decades away” to “may never be fully achievable in the Drexler sense.” The 2024 arXiv analysis (Shvydun, Sato, Bristot) notes that the literature is “often dominated by older, speculative papers” without grounding in current practical limitations.

8.2 Probability Estimates

• 80,000 Hours (effective altruism research organization) estimates a 20-70% chance that advanced nanotechnology (in the Drexler sense) is possible in principle. This wide range reflects genuine expert disagreement.
• The same analysis rates APM as a “potential highest priority area” for risk research due to the transformative (and potentially catastrophic) implications if achieved.

8.3 Near-Term Milestones (achievable in 5-10 years)

Based on current trajectories:

• Atomically precise quantum processors with 50-100+ qubits (SQC pathway): likely by 2030
• Parallel STM-based patterning using MEMS arrays: demonstrated at lab scale by 2028-2030
• EUV-based scalable atom patterning: integration with semiconductor fabs by 2030-2035
• Atomically precise catalysts in commercial chemical processes: already beginning, widespread by 2030
• DNA origami drug delivery in clinical trials: underway, approvals possible by 2028-2030
• First experimental mechanosynthesis of a simple diamondoid structure: highly uncertain, possibly 2030-2040 if heavily funded

8.4 Medium-Term Goals (10-25 years)

• Molecular workstation: A programmable tool for performing sequences of mechanosynthetic reactions (Freitas’s 2025 roadmap target)
• Integrated nanosystems: Devices that transport feedstock, modify it, position it, and bond it to a growing structure with no defects (DOE workshop 5-year goal, originally set in 2015, now likely 2035+)
• Hybrid bio-synthetic molecular machines: Combining protein engineering with synthetic molecular machines for more capable assemblers

8.5 Long-Term Vision (25+ years)

• Nanofactories: Hierarchical systems producing macroscale products via molecular manufacturing
• General-purpose matter compilers: Devices that can fabricate arbitrary structures from feedstock materials
• These remain deeply uncertain; many researchers consider them possible in principle but unpredictable in timeline

9. Funding Landscape

9.1 U.S. Government Funding

National Nanotechnology Initiative (NNI):

• FY2025 budget request: $2.2 billion (all-time record)
• Cumulative investment through FY2025: $45 billion since NNI inception
• 12 federal agencies participate
• Largest contributors: NIH, NSF, DOE, DoD, NIST

Department of Energy (DOE):

• Advanced Materials and Manufacturing Technologies Office (AMMTO) funds APM-specific research
• FY2024: SBIR/STTR awards of ~$200,000 each to companies including Tiptek (atomically precise quantum devices) and Zyvex Labs (3D qubit architectures)
• APM funding peaked in 2018 at $10.5M SBIR matched with $8M Emerging Research Exploration funds
• Projects have included 2D/3D positional assembly, self-assembling graphene nanoribbons, and atomically precise catalysts

DARPA:

• Atoms to Product (A2P) program: Selected 10 performers to develop technologies for assembling nanometer-scale components into millimeter-scale systems
  • PARC (Micro-Assembly Printer): printing with smart nano-material “ink”
  • SRI (Levitated Microfactories): combining MEMS precision with pick-and-place robotics
  • Embody: reinforced collagen nanofibers mimicking natural ligaments
• Quantum Benchmarking Initiative (QBI): Awarded contracts to SQC (Stage A in 2025, advanced to Stage B)
• Historical: ~$10M to Zyvex-led Atomically Precise Manufacturing Consortium

NIST:

• Atom-scale devices program: engineering, metrology, and manufacturability of atomic-scale devices
• Atom manipulation with STM research program

9.2 International Government Funding

Australia:

• Significant investment in SQC through UNSW and direct government support
• ARC Centre of Excellence for Quantum Computation and Communication Technology

Canada:

• $40 million investment in CBN Nano Technologies announced by Minister of Innovation, Science and Economic Development
• CBN’s $220 million nanotechnology project aimed at first commercial-scale APM

European Union:

• Continued support for molecular machines research (Feringa group, Netherlands; Leigh group, UK)
• Paul Scherrer Institute (Switzerland) EUV hydrogen desorption research

9.3 Private Investment

Venture Capital:

• Prime Movers Lab (Jackson, WY): VC firm with explicit focus on breakthrough physical science including APM. Dan Slomski has published extensive analysis of APM investment opportunities.
• Global nanotechnology market: $6.59 billion (2024), projected $115.41 billion by 2034 (33.14% CAGR)
• Over 15,000 nanotechnology companies globally, including 667 startups

Corporate:

• CBN Nano Technologies: Backed by Canadian Bank Note Company ($30.6M raised + $40M government investment)
• SQC: Backed by Australian government, Telstra, Commonwealth Bank, and others
• Zyvex Labs: ~$28M from DARPA, Army Research Office, DOE

10. Companies Working on APM

10.1 Direct APM Companies

10.2 Adjacent / Enabling Companies

10.3 Notable Nanotechnology Startups (Adjacent)

• Group14 Technologies ($1.2B funding): Advanced battery materials
• Sila Nanotechnologies: Silicon-dominant battery anodes
• INBRAIN Neuroelectronics: Graphene-based neural interfaces
• SiTration (MIT spinout): Critical mineral recovery using silicon wafers

10.4 Research Nonprofits

11. The Path from Current Nanotech to Matter Compilers

11.1 The Foresight-Battelle Roadmap

The original roadmap (2007) identified a “vision gap” between current nanoscience and productive nanosystems. It proposed incremental steps grouped into two parallel pathways:

Pathway 1: Tip-Based Positional Assembly

1. Current: STM/AFM single-atom manipulation on surfaces (achieved)
2. Next: Reliable, automated multi-step mechanosynthesis sequences
3. Then: Parallel tip arrays for throughput scaling
4. Then: Tooltip recycling and feedstock delivery systems
5. Goal: Programmable molecular workstation

Pathway 2: Integrated Nanosystems Using Molecular Machine Components

1. Current: Synthetic molecular machines (motors, switches, rotors) (achieved)
2. Next: Integration of multiple molecular machine components into systems
3. Then: Self-assembling functional nanosystems
4. Then: Molecular machines that can fabricate other molecular machines
5. Goal: Nanofactory with hierarchical assembly

11.2 DOE Workshop 5-Year Goal (2015)

A realistic integrated nanosystem would:

1. Transport individual feedstock molecules to a workspace (actively or passively)
2. Modify or chemically activate the feedstock
3. Manipulate or transport the feedstock to the attachment point at a specified atomic position
4. Chemically bind the feedstock to a growing structure at that attachment point
5. Repeat with no defects

This goal has not been achieved as of 2026, and likely requires another 5-10 years of concentrated effort.

11.3 Key Intermediate Steps (Updated Assessment, 2026)

Step 1 (Current, 2026-2030): Specialized APM

• Atomically precise quantum devices (SQC, Zyvex), already commercial
• Atomically precise catalysts, entering commercial use
• DNA origami-based drug delivery, entering clinical trials
• Atomically precise membranes, approaching medical applications
• MEMS-integrated parallel STM arrays, lab demonstration

Step 2 (2030-2035): Scalable Atomic Patterning

• EUV-based hydrogen desorption for large-area atom patterning (PSI/UCL work)
• Integration of ML/AI for automated atomic-precision fabrication
• First experimental mechanosynthesis of simple diamondoid structures
• Programmable molecular machines performing multi-step chemical synthesis

Step 3 (2035-2045): Molecular Workstation

• Desktop-scale tool for performing programmed sequences of positional chemical reactions
• Tooltip libraries and automated tooltip recycling
• Demonstration of fabricating one molecular machine using another
• Hybrid biological-synthetic molecular manufacturing systems

Step 4 (2045-2060+): Productive Nanosystems

• Self-assembling arrays of molecular workstations
• Hierarchical manufacturing: molecular-scale operations producing mesoscale components
• Integration with conventional manufacturing for macroscale products
• This is the “matter compiler” horizon, deeply uncertain

11.4 Critical Enabling Technologies

1. Parallelization: Moving from single-tip to massively parallel tip arrays (MEMS-based or EUV-based)
2. Automation / AI: ML-guided fabrication, automated error detection and correction
3. Tooltip chemistry: Experimentally validated tooltip designs for specific bond-forming reactions
4. Feedstock handling: Reliable presentation of specific molecules to the reaction site
5. Metrology: Real-time atomic-resolution verification of fabrication progress
6. Materials science: Understanding of diamondoid surface chemistry under mechanosynthetic conditions
7. Software tools: Design software for molecular machines (MSEP.one is a first step)

12. Risks and Governance

12.1 Identified Risks (per 80,000 Hours)

• Weapons proliferation: APM could enable production of nuclear-equivalent weapons from cheap feedstocks
• Accelerated AI development: More powerful computing hardware could increase risks from transformative AI
• Economic disruption: Manufacturing that uses cheap inputs and fast design cycles could destabilize global economics
• Biological risks: Atomically precise fabrication of pathogens or toxins
• Power concentration: States or corporations with APM capability could gain decisive strategic advantages

12.2 Governance Gap

80,000 Hours rates APM-related risk research as “extremely valuable” given current uncertainties. There is no international governance framework for APM, and the technology is not yet mature enough for most policymakers to take seriously, creating a window of opportunity for proactive governance development.

12.3 The “Grey Goo” Question

The self-replicating nanomachine scenario has been largely de-emphasized by serious researchers. Drexler himself argued for non-replicating nanofactory designs. Current expert consensus is that accidental self-replicating nanomachines are extremely unlikely, but deliberate weaponization of molecular manufacturing technology is a genuine long-term concern.

13. Conclusions

What is Real Now

Atomically precise manufacturing is not science fiction, it is being done today in specific domains:

• Silicon quantum devices are being manufactured atom-by-atom and sold commercially (SQC)
• STM lithography achieves higher resolution than the best conventional semiconductor tools (Zyvex)
• DNA origami constructs are entering clinical applications
• Atomically precise catalytic clusters are being deployed in industrial processes
• AI-designed proteins represent mature atomically precise molecular engineering

What Remains Aspirational

The Drexler/Freitas/Merkle vision of general-purpose molecular manufacturing, nanofactories producing arbitrary macroscale products atom-by-atom, remains decades away at minimum. Diamond mechanosynthesis has not been experimentally demonstrated. No molecular assembler prototype exists. The path from current specialized APM to general-purpose matter compilers requires solving numerous unsolved engineering problems.

The Most Promising Near-Term Directions

1. Silicon quantum computing (SQC pathway): The most commercially advanced APM application, with revenue and government contracts
2. AI-accelerated molecular design: Combining AlphaFold/Baker-style protein design with molecular machine engineering
3. EUV-based scalable atom patterning: Could bridge the gap between STM precision and industrial-scale throughput
4. MSEP.one and simulation tools: Enabling wider exploration of the molecular machine design space
5. Atomically precise membranes and catalysts: Near-term commercial applications driving investment

The Key Bottleneck

The fundamental challenge remains throughput. An STM can manipulate perhaps one atom per second. A nanofactory would need to coordinate trillions of operations per second. Bridging this gap, through parallelization, self-assembly, or entirely new approaches, is the central unsolved problem of the field.

Sources

Research Papers and Reviews

• A Comprehensive Analysis of the Future of Atomically Precise Manufacturing (arXiv, Sept 2024)
• Atomically Precise Manufacturing of Silicon Electronics (ACS Nano, Feb 2024)
• EUV-induced hydrogen desorption for silicon quantum device patterning (Nature Comms, 2024)
• An 11-qubit atom processor in silicon (Nature, 2025)
• Machine Learning-Assisted Precision Manufacturing of Atom Qubits in Silicon (ACS Nano, 2024)
• Spiroligomer-Based Macrocycles for Atomically Precise Membranes (Angew. Chem., 2023)
• DNA Origami and Its Applications in Synthetic Biology (Advanced Science, 2026)
• Fabrication of Functional 3D Nanoarchitectures via ALD on DNA Origami Crystals (2025)
• Self-Replicating DNA-Based Nanoassemblies (JACS, 2024)
• Atomically Precise Cu Nanoclusters: Synthesis and Catalytic Applications (Nano-Micro Letters, 2024)
• Electronic Control of Silicon Surface Atomic Structures with Two-Probe STM (ACS Nano, 2025)
• Ultra-high vacuum STM-assisted atomic-scale manufacture of quantum devices (Chinese Science Bulletin, 2025)
• Inverted-mode STM covalent mechanosynthesis (arXiv:2512.24431, Dec 2025)
• Freitas, R.A. Jr., “Molecular Workstation Roadmap I” (May 2025)

Institutions and Organizations

• Zyvex Labs
• Silicon Quantum Computing
• MSEP.one, Molecular Systems Engineering Platform
• Foresight Institute
• Institute for Molecular Manufacturing
• CBN Nano Technologies (CBInsights)
• Tiptek
• Robert Freitas Publications
• NIST Atom Manipulation Program
• NIST Atom-scale Devices Program

Government Programs and Funding

• DARPA Atoms to Product (A2P)
• DOE AMMTO Small Business Awards (2024)
• NNI FY2025 Budget Supplement
• DOE APM Infographic
• DOE OSTI: Platform Technology for High-throughput APM
• Canada $40M Investment in CBN Nano Technologies

Risk Analysis

• 80,000 Hours: Risks from Atomically Precise Manufacturing
• EA Forum: Risks from APM Problem Profile

Industry Analysis

• Prime Movers Lab: APM Part 1
• Prime Movers Lab: APM Part 2
• Prime Movers Lab: APM Part 3
• New Institute for Molecular Machines (NextBigFuture, Dec 2024)
• SQC DARPA QBI Stage B Announcement
• SQC Silicon Processor Fidelity Results

Nobel Prize Context

• 2024 Nobel Prize in Chemistry (Protein Design / AlphaFold)
• 2025 Nobel Prize in Chemistry (Metal-Organic Frameworks)
• 2016 Nobel Prize in Chemistry (Molecular Machines)

Patents

• CBN Nano Technologies Patents (Justia)
• Freitas Mechanosynthesis Patent US7687146B1
• Positional Diamondoid Mechanosynthesis Patent US8276211B1

Wikipedia and Reference

• Mechanosynthesis
• Molecular Nanotechnology
• Drexler-Smalley Debate
• K. Eric Drexler
• Robert Freitas
• Foresight Institute
• Diamond Mechanosynthesis (Zyvex)
• Diamond Mechanosynthesis Literature Review (GitHub)