Methodology

How we derive the figures on Frontier Milestones Invest, and where the line sits between fact and opinion.

Verified technical stage (0–5)

For each field we publish a five-stage ladder to the goal (below). An actor’s verified stage is the highest ladder stage reached by an achieved milestone, plus a half-stage when work at the next stage is in progress. It is computed deterministically from the verified dataset — a derived fact, not a forecast. Funding, IPO and M&A events never count toward a technical stage.

Market-implied stage (0–5) — OPINION OPINION

Within a field cohort we fit the relationship between valuation and verified stage — regressing log(valuation) on verified stage to obtain a “value-for-maturity” line. A company’s valuation is then read back onto that line as an implied stage. This is an opinion about what the market appears to price, calibrated per field — not a fact about the company.

The gap — OPINION OPINION

Gap = market-implied stage − verified technical stage: the company’s distance from its cohort’s value-for-maturity line, in stage units. Beyond ±0.5 it is labelled market-ahead or progress-ahead; within, aligned. It is shown only where a real valuation and a cohort fit exist — listed companies first, private companies once round data is structured — and never as a recommendation.

The stage ladders below are a first pass under domain review. The gap activates as live market data connects.

Field stage ladders

Nuclear Fusion Goal Net-positive, grid-connected fusion power

1 Plasma & confinement High-temperature or long-duration plasma demonstrated
2 Scientific breakeven Ignition / Q≥1 in the lab
3 Repeatable gain & engineering Target gain, peer-validated physics or major plant hardware
4 Pilot plant licensed/built Demo/pilot plant permitted, under construction or design-reviewed
5 Commercial power Net-positive plant delivering power to the grid

Quantum Computing Goal Useful, fault-tolerant quantum computing

1 Scale & supremacy 1,000+ physical qubits or quantum supremacy
2 Below-threshold QEC Below-threshold error correction (logical < physical)
3 Many logical qubits Tens of logical qubits / strong error suppression
4 Fault-tolerant scaling Commercial fault-tolerant systems sold / modular scaling
5 Useful FTQC Cryptographically-relevant or commercially useful FTQC

Small Modular Reactors Goal Commercial small modular reactor power

1 Design First-of-kind design developed
2 Licensed design Design certified/licensed by a regulator
3 Construction permit Construction permit issued / groundbreaking
4 Under construction First concrete/basemat or demo reactor criticality
5 Commercial online Commercial SMR delivering power

Humanoid Robots Goal Mass-produced general-purpose humanoids

1 Advanced platform Advanced electric humanoid hardware
2 Paid pilots Paid warehouse/factory pilots
3 Commercial deployment Commercial deals / sustained autonomous shifts
4 Volume manufacturing Dedicated factory at volume (≈10k/yr)
5 Mass-produced general-purpose Mass-produced general-purpose humanoids

Autonomous Driving Goal Profitable, nationwide robotaxi service

1 Public driverless First public driverless operation
2 Paid in a city Paid driverless approved in a city
3 Scale in cities High volume in operating cities (100M+ miles, 100k+ rides/wk)
4 Multi-market expansion Multi-market / freeway / broad expansion
5 Profitable nationwide Profitable, nationwide service

Reusable Rockets Goal Fully & rapidly reusable super-heavy launch

1 Booster landing First orbital booster landing
2 Booster reflight Reflight of an orbital booster
3 High reuse cadence Many flights per booster / second-mover reuse
4 Super-heavy reuse demo Fully reusable super-heavy demonstration
5 Routine super-heavy reuse Fully & rapidly reusable super-heavy in service

Brain-Computer Interfaces Goal FDA-approved commercial brain-computer interface

1 Human implant (research) Human cortical implant for control
2 Next-gen human implants Next-gen implants in humans (multiple teams)
3 Regulatory clearance Regulatory clearance / first commercial approval
4 Pivotal trials / scaled use Pivotal trials or scaled clinical use
5 FDA-approved commercial BCI First FDA-approved commercial BCI

Air Taxis (eVTOL) Goal Scaled, profitable passenger eVTOL service

1 Prototype / type certificate Prototype flight or first type certificate
2 Piloted city/airport flights Piloted inter-airport / city flights
3 Certification progress Advanced certification stage (e.g. FAA Stage 4 of 5)
4 Certified + initial ops Type certification complete + initial commercial ops
5 Scaled passenger service Scaled, profitable passenger service

Commercial Space Stations Goal Commercial stations replace the ISS in LEO

1 Design funded Commercial station design funded
2 Module qualification Station module qualification testing
3 First station to orbit First commercial station reaches orbit
4 Operational crewed station Operational crewed commercial station
5 Replaces ISS in LEO Commercial stations take over LEO

Frontier AI Goal AGI — broadly human-level capability

1 Mainstream LLM GPT-4-class model (~1e25 FLOP)
2 Frontier scaling + reasoning 1e26-FLOP scale + reasoning models
3 Beats abstraction benchmarks Beats hard abstraction benchmarks (ARC-AGI)
4 Robust general agency Robust general agentic capability across new benchmarks
5 AGI Broadly human-level capability

Leading-Edge Chips Goal 1.4 nm-class chips in volume production

1 EUV volume EUV lithography in volume production
2 GAA node First gate-all-around node
3 2 nm-class + backside power 2 nm-class volume + backside power delivery
4 Sub-2 nm ramp Sub-2 nm node ramping
5 1.4 nm-class volume 1.4 nm-class node in volume