FYRINVESTOR DECK · DRAFT ⌘P → save as PDF (background graphics ON)
FYR
A lighthouse for the underground

Positioning you can prove — so autonomous machines and people can finally share the same mine. Safer than exclusion. Productive around the clock.

Investor briefing · 2026fyr.mining
The problem · Safety

A kilometre down, nobody can prove where anyone is.

GPS reads zero underground. Position is a guess — and the guess is killing people. Fatalities among the world's largest miners are rising, not falling.

42
deaths at ICMM member companies in 2024 — the third consecutive annual rise
ICMM Safety Performance 2024
~90%
of those fatalities occur underground or in underground-supporting work
ICMM 2020–2024
#1
underground killer: vehicle interaction — 33% of underground deaths, ahead of fall-of-ground
ICMM 2020–2024
FYR02 · −40 M
The problem · Productivity

The industry's answer is exclusion. Exclusion is downtime.

  • Autonomy zonesAutonomous machines run behind physical barriers and light curtains. One person crosses the line — the whole level shuts down.
  • Blast re-entryAfter every blast the level stands empty for hours while gases clear — machines wait with the people.
  • Shift changeThe mine empties, travels, and refills — twice a day, every day, with production idling in between.

“The underground automation area is fully isolated with safety barriers.”

Epiroc, on today's state of the art in underground autonomy
FYR03 · −80 M
25%
What exclusion costs

A major copper mine's loading fleet ran at a quarter of its capacity.

Not because machines break — because the operating model keeps stopping them. Every hour of exclusion is capacity the mine already paid for and doesn't use. Safety is being bought with three quarters of the mine.

Resources Policy (2026), stochastic OEE study — shovels ≈25% OEE (range 8–52%); losses dominated by operational coordination, not breakdowns.
FYR04 · −120 M
The insight

Safe or productive is a false choice.

It only exists because position underground is a statistical estimate — five to ten metres, ninety-five percent of the time. No regulator, and no sane mine manager, lets a 60-tonne machine move near people on “95% of the time”. Make position a proof, and the trade-off dissolves.

Best published underground positioning accuracy: 5–10 m at 95% confidence (Mobilaris Onboard, 2019) — an estimate, not a bound.
FYR05 · −160 M
How · The physics
A radio pulse can arrive late.
It can never arrive early.

Ultra-wideband ranging measures time-of-flight at the speed of light. Reflections and obstructions only ever make a distance read too long — never too short. The error has one sign.

Combine measurements from several anchors and the one-sided errors carve out a proven localization volume: a region mathematically guaranteed to contain the tag. Not an estimate. A proof.

ANCHOR A ANCHOR B ANCHOR C TAG PROVEN VOLUME guaranteed to hold the tag each range reads at most r — the tag is provably inside every arc
FYR06 · −200 M
How · The system

Lighthouses on the walls. A proof around every machine.

  • 01Anchors — the fyrarBlast-hardened UWB anchors every 30 m, two cables per tunnel. ~2 hours of installation per kilometre.
  • 02Chips on everythingEvery person, vehicle and machine carries a tag with a live proven localization volume.
  • 03Protective shieldsEach shield = proven volume + the distance that unit needs to fully stop. Shields move with the machines, recalculated continuously.
  • 04Guaranteed stopShields overlap → the machines stop. Engineered to SIL-3 / PL-e — safety a certifier can sign.
AUTONOMOUS LHD · 20 KM/H WORKER · ON FOOT PROVEN VOLUME SHIELD = VOLUME + STOP DISTANCE no overlap → both keep working
FYR07 · −240 M
FYR live operations monitor — safety monitor active, event feed of predicted interventions
The product

Every person. Every machine. Proven, live.

One operating picture for the whole mine: positions with guarantees, predicted conflicts, and every intervention logged as evidence. The shield does the stopping — the control room finally sees the mine it runs.

FYR08 · −280 M
The payoff · Productivity

Mixed traffic turns excluded hours into producing hours.

Machines re-enter first after a blast. They keep hauling through shift change. A technician walks a live level and nothing shuts down — shields shrink speed instead. In our calibrated digital twin of an LKAB-style operation:

Today · exclusion25%
FYR · mixed traffic90%

The gap hours alone are worth mining: at Westgold's Big Bell, autonomous loaders working only shift changes and re-entry windows delivered 10,000–12,000 t/month above forecast. FYR extends that from the gaps to the whole day.

25% = Resources Policy OEE (2026), external measurement. 90% = FYR twin, modelled ceiling. Big Bell: Epiroc customer story, 2025.
FYR09 · −320 M
The tailwind · Regulation

The law is already ahead of the industry.

L9
South Africa mandates automatic machine intervention on underground vehicles — in force since Dec 2022, no transition period
MHSA reg. 8.10.1.2(b), DMRE
0
collision-avoidance vendors publish a SIL rating for their detection chain — the whole market certifies behaviour, not integrity
Survey of Booyco, Schauenburg, Torsa, Hexagon, Wabtec, Newtrax, 2025
2025
ICMM + OEM CEOs' target year for collision-avoidance technology capable of eliminating vehicle-interaction deaths
ICMM ICSV initiative

Regulators are writing cheques the industry can't cash. FYR is engineered to the standard the mandates imply — provable, certifiable, SIL-3 — before enforcement catches up worldwide.

FYR10 · −360 M
Why now

Four forces, one moment.

DemandDeeper, or nothingThe IEA sees a 30% copper supply gap by 2035 and $500–600B of new mines needed by 2040 — deeper, lower-grade, underground.
LabourNobody's coming71% of mining executives say talent shortage blocks production targets; ~70% of under-30s won't consider the industry.
AutonomyAdoption inflectedAutonomous haul trucks +84% in one year (2,080 → 3,832, 2024–25). Every one of them still needs people kept out of the way.
TechnologyProvable is possibleUWB silicon is commodity, rail has certified UWB positioning to SIL-4, and a Rust safety toolchain is TÜV-qualified to SIL-3. Every ingredient now exists.
IEA Global Critical Minerals Outlook (2025); McKinsey talent survey (2023); GlobalData AHS census (2025); TÜV SÜD / Ferrocene (2025), Piper Networks SIL-4 UWB (2021, rail).
FYR11 · −400 M
Market

We sell the layer everything else depends on.

$6.7B2025
$15.7B2030
Real-time location systemsour layer · 18.6% CAGR · MarketsandMarkets 2025
$3.9B2025
$5.9B2030
Mining automationdepends on position · 8.4% CAGR · MarketsandMarkets 2025
$0.7B2025
$1.2B2030
Mine collision avoidancemandate-driven · 11.2% CAGR · Tersus 2022

The positioning layer compounds faster than the automation market it enables — and mandates convert “nice to have” into law: South Africa's collision-avoidance market alone is heading from $234M to $876M (2024→2034).

FYR12 · −440 M
Competition

Everyone else monitors, interrupts, or excludes. We prove.

MonitorSituational awareness Statistical estimates — 5–10 m, 95% of the time. Fine for dashboards; no certifier will let it stop a machine. MOBILARIS · MST · NEWTRAX
InterruptProximity detection Field-strength triggers: “too close” alarms and slam stops. Zero published SIL ratings across the category. BOOYCO · SCHAUENBURG · TORSA · HEXAGON
ExcludeOEM autonomy Barriers, light curtains, cleared levels. Utilisation capped by every hour people need access. SANDVIK AUTOMINE · EPIROC
ProveFYR · guaranteed position A mathematical bound on every position, engineered to SIL-3/PL-e — the only basis on which people and autonomous machines can legally share ground. Moat: one-sided-error physics + functional-safety certification — years of certifier evidence a fast follower can't skip.
FYR13 · −480 M
The live mine: five lit levels descending into the dark, every person and machine tracked
What mining becomes

The mine never empties. The ore never stops. Nobody gets hit.

Machines re-enter first after every blast. Production runs through shift change, breaks and nights. People go underground when their judgment is needed — and walk live levels under a guarantee, not a guess. The industry's oldest trade-off, retired.

FYR14 · −520 M
Team placeholders — to fill

Built by people who've shipped underground.

Founder nameCEO · Co-founder

One line: the operating or domain credibility that makes this person inevitable for this company.

Founder nameCTO · Co-founder

One line: UWB / functional-safety / real-time systems track record.

NameSafety certification lead

One line: IEC 61508 / ISO 13849 assessment experience — the hire investors will ask about.

NameMining operations

One line: years underground at a tier-one operator; speaks the customer's language.

NameAdvisor

One line: operator-side or OEM heavyweight lending distribution credibility.

NameAdvisor

One line: standards / regulatory figure (EMESRT, ISO 21815, Minerals Council).

FYR15 · −540 M
The ask

Light the first mine.

Now → 12 moProve it underground

Production pilot with a tier-one Nordic operator: one level, live mixed traffic, evidence pack for certification.

12 → 30 moCertify it

SIL-3 / PL-e assessment of the positioning chain — the first functionally-safe positioning layer in mining.

30 mo →Scale where it's law

South Africa's Level-9 mandate first, then every jurisdiction that follows. Hardware + per-tag SaaS.

€X.X M

18 months of runway: pilot hardware, the certification programme, and the safety-engineering hires that make FYR the reference. amount + terms to fill

FYR — a lighthouse for the undergroundmagnus@saltfish.ai · fyr.mining