Presented by Generative Gains
Thermal Trace
Pre-seed · QAR 1.85M · Qatar first, GCC next

Know what a failing fridge is costing before the gate does.

Cold-chain intelligence: physical risk and commercial value, in one reserve.

Cited · FAO / UNEP 2024
02The problem

A third of our food never gets eaten.

12.3SDG 12.3 · halve food waste, cut supply-chain loss 2SDG 2 · zero hunger 13SDG 13 · climate action
13%

of food is lost between harvest and retail.

19%

more is wasted after it reaches retail and homes.

8–10%

of global greenhouse-gas emissions come from food loss and waste.

Cited · TASMU · Qatar NFSS 2030
03Why here

An import-dependent region, a brutal summer.

Climate and imports

45 °C

Imports by sea and land, customs holds and border queues, across a summer that passes 45 °C. A fridge that fails at noon can spoil a load in hours.

Policy tailwind

TASMU analytics and Qatar's Food Security Strategy target efficient supply chains and less food waste. Thermal Trace works on the loss in transit.

National Food Security Strategy 2030
Food waste
−50%
Food loss
−30%
Thermal debt Maritime import+Customs hold+Border queue+45 °C summer=Lost shelf life
Thermal Trace is the operator-level layer that complements national systems.
Commercial Reserve · designed, staged in the pilot
04Two clocks

Safe is not the same as sellable.

Thermal ReserveThe physical clock
Commercial ReserveThe commercial clock
FocusProduct health: “Can it still be consumed?”
FocusEconomic life: “Can it still earn its value?”
ClockProduction → Transport → Deterioration → Shelf-life boundary
ClockArrival → Customer exposure → Sell-through → Replenishment
GoalProtect physical life. Calculated
GoalProtect economic life. Predicted

Health creates the possibility of selling. Demand creates the possibility of realising value. Commercial Reserve lives where the two meet.

CR = Qr × Pr × Vr
Qr quantity realistically realisable Pr probability of sale Vr expected value per unit
Layer 3 designed · staged in the pilot
05The product

The anatomy of value: four reserves, one stack.

Thermal Trace GROUND TRUTH Thermal Reserve THE ENVELOPE Commercial Reserve THE VELOCITY Decision Reserve THE ACTION WINDOW
Layer 4 · Decision Reserve Optimized

The time left to act before the good options disappear.

Layer 3 · Commercial Reserve Designed

The economic runway: value that can still be realised.

CR = Qr × Pr × Vr
Layer 2 · Thermal Reserve Calculated

The physical budget left before quality is compromised.

Layer 1 · Thermal Trace Measured

A continuous, timestamped memory of temperature exposure and stress.

06Layer 1 · Thermal Trace · What we measure

Six signals. Three numbers that matter.

Temperature
DHT11 today · DS18B20 next

The demo node reads one point. Ceiling and floor probes are the next step, so a warm layer shows up instead of hiding in an average.

Humidity
DHT11 today · SHT3x next

Rising humidity shows whether an open door is wetting the cartons.

Shock & tilt · simulated
3-axis MEMS · designed

Bruising from a dropped pallet is damage a thermometer never sees.

Gases · designed
Metal-oxide · VOC, C2H4

Ethylene is the fruit itself reporting that it has started to ripen.

Sub-surface · designed
NIR · 700–2500 nm

Finds internal bruising before anything is visible from outside.

Surface · designed
RGB camera · edge vision

Colour shift, mould, and a count of what is actually on the shelf.

They feed Deterioration detection Remaining Useful Time Thermal exposure ET Calculated

Temperature and humidity flow from a real sensor node today (DHT11, about ±2 °C); shock is simulated and the rest are the designed next layer. All publish the same message, so adding one changes nothing downstream.

Example scenario
07Layer 2 · Thermal Trace → Thermal Reserve

The label says 7 days. The trace says 4.5.

Layer 1 · Thermal Trace · cargo temperatureMeasured
8° 6° 4° 2° SAFE MAX 4 °C refrigeration fails 3.8° 7.4° E T = exposure
Thermal exposure · °C·min
ET = Σ max(0, Ti − Tsafe) · Δt
Deterioration rate · Arrhenius
k(T) = kref · e(Ea/R)(1/Tideal − 1/T)
Layer 2 · Thermal Reserve · safe life left
D = Σ k(Ti)·Δti → reserve = L·(1 − D)
Label date
7.0 d
Thermal reserve
4.5 d

CalculatedDeterministic Python. No AI model touches these numbers.

AI on a leashNeither may invent a number

Laya

System 1 · local · offline Predicted

Answers typed questions in one pass. Never generates text; corroborates the decision, never overrides it. Base model, confidence shown as uncalibrated.

DeepSeek V4.1 Flash

System 2 · cloud AI-explained

Explains the calculated facts in plain words. May rank only feasible options; never invents a temperature, ETA or cost.

08The shift

Redefining the perishable asset.

The biological clockDefensive · today

Legacy model: inventory is a degrading liability, managed by a static printed expiry date.

Result: late markdowns and write-offs.

The commercial traceOffensive · Thermal Trace

Thermal Trace model: inventory is an asset with a live value, priced from its real thermal history and live demand.

Result: value recovered before it is lost.
Countdown · example · Commercial Reserve designed
09Two divisions

Two clocks, one platform.

Thermal Trace tells us what is happening to the product. Two reserves read it.

Physical clockCalculated

Thermal Reserve

How much physical margin remains.

45 min left

For 3PLs, fleets and warehouses. Sold as a platform fine-tuned to each customer: sensors, calibration and integration.

3PLsFleetsWarehouses
Commercial clockPredicted

Commercial Reserve

How much economic opportunity remains, and what we should do about it.

POS · ERP · WMS → route → price to clear

For hypermarkets and HORECA. Sold as SaaS: joins the physical state with POS, ERP and WMS to route stock, and price it to clear.

HypermarketsHORECA
Entry point · Thermal Reserve · tuned platformControlled extension · Commercial Reserve · SaaS
Illustrative value curve · designed
10Layer 3 · The full commercial cycle

In transit until the profit is booked.

The trace follows the value from farm to ledger, not just from dock to dock.

Kinetic transit

On the truck, burning fuel and thermal life. Routing turns at-risk stock back into sales.

Thermal + Decision Reserve

Static transit

On the shelf, burning shelf life while it waits for customers.

Commercial Reserve

Consumer transit

Into the customer's cart: the moment value is realised.

Value realised

Consolidation

POS data flows back, profit hits the ledger, and demand triggers replenishment.

Value Ledger
Live demo · simulated data
11The software

One screen for the whole cold chain.

Thermal Trace · Command center
Thermal Trace command center: KPI strip, live map of five trucks in Doha, active incidents, sensor monitoring and simulation control 1 2 3 4 5
  1. 1
    Headline figuresActive trucks, at risk, food saved, loss prevented.
  2. 2
    Live mapEvery truck, warehouse and store, coloured by risk.
  3. 3
    Active incidentsRanked by severity, with the cause in plain words.
  4. 4
    Sensor monitoringEach truck's cargo temperature against its safe range.
  5. 5
    Simulation controlTrigger a failure and watch the system respond.
Data→Condition→Prediction→Optimization→Decision→Food loss
Example scenario
12Layer 4 · Decision Reserve

The window to act, and the move to make.

min( Ctransport + Cfood‑loss + Cdelay )
Subject to
  • ETA ≤ remaining thermal reserve
  • Quantity ≤ available capacity
  • Storage temperature compatible
AI CorroboratesExplainsNever invents a number
Reefer Compressor failing Thermal reserve 45 min
DestinationETAFits 45 min
HypermarketPlanned route
90 min✗ spoils
Nearby buyerRe-routed
20 min✓ fits
OptimizedREROUTEDecision reserve ≈ 25 min to commit. The optimizer picks; the AI only explains why.
Illustrative · products named for reference
13Layer 1 · Same sensors, two foods

Banana fails cold. Ice cream fails warm.

Signal
🍌 13–14 °C · 90–95 % RHToo cold browns it, ethylene ripens it
🍦 −18 °C or colderAny thaw-refreeze ruins texture
Gate check onlyOne logger, read after delivery
Bought trackersIn transit vs handheld at QA
TemperatureDHT11 today · DS18B20 next
Below 13 °C = chilling injury: peel turns grey days later, after it is sold on.
A rise above −18 °C grows ice crystals; it refreezes gritty and nobody can tell from outside.
Excursion found at the gate or by the customer. Load already damaged.
Tive Solo 5G, Sensitech TempTale GEO X, ELPRO LIBERO
HumidityDHT11 today · SHT3x next
Too dry and skins shrivel and scar; too wet and cartons go soft.
A door left open lets moist air in: frost on cartons proves a warm-up.
Soggy or dry cartons noticed only when unloading.
Tive Solo 5G, ELPRO LIBERO Gx
Shock & tilt · simulated3-axis MEMS · designed
A dropped pallet bruises; bruises appear 1–2 days later.
Crushed tubs and popped lids; tilt lets softened product leak.
Damage claims argued with no record of where or when.
Tive Solo 5G (12 g shock), Controlant Saga (vibration)
Gases · designedMetal-oxide · VOC, C₂H₄
Ethylene rising: the load started ripening in transit, so sell it first.
Not used. Sealed product, and MOx sensors are not rated this cold.
Over-ripe load discovered at the store, too late to reroute.
Handheld only: Felix F-900 ethylene analyser
Sub-surface · designedNIR · 700–2500 nm
Starch turning to sugar inside, before the peel shows it.
Not used. Packaging blocks it; temperature history tells the story.
Ripeness judged by eye at the dock.
Handheld only: Felix F-750, VIAVI MicroNIR
Surface · designedRGB camera · edge vision
Peel colour stage, brown spots, crown mould, count on shelf.
Frosted or misshapen packs, open lids, empty facings.
Staff walk the aisle a few times a day.
No in-transit product; fixed shelf cameras only
They feed Deterioration detectionRemaining Useful TimeThermal exposure ETCalculated
14The shift, in full

From managing biology to managing liquidity.

Dimension Legacy supply chain Thermal Trace
The goal Delay spoilage (defensive) Maximise yield (offensive)
Asset view A degrading physical object An asset with a live, priced value
Inventory rule FEFO: first expired, first out, by printed date LCRFO (planned): lowest Commercial Reserve, first out, by live physics
Intervention Passive temperature alarms Optimized routing today; early markdowns planned
Waste An invisible operating tax A quantified, bounded value at risk
Hypotheses to validate · not market prices
15Business model

Tailored to each customer. Sold as SaaS.

Thermal Trace platform · shared IP Trace engineFood-state modelData ingestionDeterministic governanceDecision workflowsValue ledger
Physical intelligence

Thermal Reserve

Configured to the customer's equipment, sensors and products. The deployment is customised; the platform stays reusable.

Market intelligence

Commercial Reserve

Precise calculations over the retailer's own POS, inventory and ERP data. The repeatable SaaS product.

Pay 1×, recover ≥ 3×every QAR 1 charged saves QAR 3+ of food.
1 · One-time setup Implementation QAR 90K Fixed setup cost. Hardware they need is billed separately, at cost.
→
2 · Yearly subscription · pick one Standard QAR 120K / yr One product · live monitoring, alerts and reserves · monthly value report · email support
2 · Yearly subscription · pick one Enterprise QAR 216K / yr Both products · customised dashboard · unlimited integrations · SSO and uptime SLA · full support
16The ask

Twelve months to prove it in the field.

QAR 1,850,000
≈ US$508,000 · 12 months · pre-seed or strategic round
  • Core five-person engineering team55.1%
  • Sensors, deployment, cloud, integration23.5%
  • Validation, customers, legal10.8%
  • Contingency reserve10.5%

What it buys in 12 months

  1. An anchor customer live at 3–5 facilities
  2. Avoided loss measured at ≥ 3× the platform fee
  3. Commercial Reserve tested in controlled trials
  4. An annual contract and a second customer
  5. QAR 300–500K contracted ARR target
12.3Every kilogram saved is logged as evidence.
Pre-seed budget · QAR
17The ask · How the money flows

Every riyal, from round to line item.

Engineering teamQAR 1,020,000 Hardware & platformQAR 435,000 Validation & go-to-marketQAR 200,000 ContingencyQAR 195,000 Core five-person team1,020,00055.1% Sensors, edge equipment155,0008.4% Deployment, calibration95,0005.1% Cloud, data, security100,0005.4% Enterprise integration85,0004.6% Food-domain validation65,0003.5% Customer development70,0003.8% Legal, IP, readiness65,0003.5% Contingency reserve195,00010.5% PRE-SEED ROUNDQAR 1.85M≈ US$508K · 12 mo
Planning assumptions · not market rates
18The ask · The team it hires

Five engineers, QAR 76,000 a month.

Role Owns Monthly (QAR) Annual (QAR)
CTO · Systems & Product Architect Architecture, scope, data contracts, security, delivery 20,000 240,000
Thermal Intelligence · Data Engineer Thermal state, exposure, risk models, shelf-life logic 16,000 192,000
Backend & Data Platform Engineer Ingestion, databases, APIs, identity, integrations 15,000 180,000
IoT & Edge Engineer Sensors, gateways, connectivity, calibration 13,000 156,000
Application & Workflow Engineer Operator UI, alerts, approvals, ERP and WMS workflows 12,000 144,000
Total gross compensation 76,000 912,000
Employment-cost provision ≈ 109,000
Total team budget ≈ 1,020,000

The CEO leads investors, anchor customers and sales. Food-safety experts are hired fractionally, from the validation line.

19The ask · What the rest buys

The other QAR 830K, line by line.

Hardware & platform 435,000
Sensors and edge equipment Sensors and gateways for 3–5 facilities: cold rooms, reefer trucks, dispatch areas 155,000
Deployment, calibration, field testing Install, calibrate, and test outages and alerts on site 95,000
Cloud, data infrastructure, security 12 months of hosting, storage and security for asset history 100,000
Enterprise integration, data engineering A few high-value flows into product, shipment, ERP and WMS records 85,000
Validation & go-to-market 200,000
Thermal and food-domain validation Fractional food-safety experts review thresholds and shelf-life models 65,000
Customer development and success The anchor pilot: site survey, value baseline, conversion 70,000
Legal, IP, commercial readiness Data-access agreements, pilot and annual contracts, IP 65,000
Contingency reserve released only with documented founder and investor approval 195,000
20The ask · What the money proves

From pilot to recurring revenue.

Months 1–4

Secure the pilot, agree how value is measured, install the devices.

Months 5–9

Run the intervention workflow and prove the avoided loss against the baseline.

Months 10–12

Convert to an annual contract and win a second customer.

The pilot must prove avoided loss ≥ 3× the platform fee. Month-12 target: QAR 300–500K contracted ARR. A target, not a forecast.
21The ask · Risks

Known risks, planned responses.

Risk Mitigation
Customer data access Sign data access and a named sponsor first
False precision Validated assumptions, uncertainty shown, domain review
Integration complexity A few high-value data flows, clear contracts
Long sales cycle Named decision-makers, success criteria, conversion terms
Building both divisions too early Thermal Reserve first, Commercial Reserve staged
10.5%

Contingency reserve

QAR 195,000 held back. Spent only with documented approval from founders and investors.

Vision · not built
22Beyond food

One primitive, many industries.

Any asset that deteriorates, cures or loses viability under the stress of time needs this architecture.

Biopharma zero-trust compliance

Physics stability budgets

Calculates active-cooling intercepts on the tarmac

Energy the grid trace

Physics grid signals, inverter telemetry

Calculates grid flexibility reserve

Space cargo the orbital trace

Physics propellant bleed, thermal cycling

Calculates mission reserve for manoeuvres

Defence the kinetic trace

Physics munition shock and degradation

Calculates operational readiness reserve

Trace → State → Reserve → Decision Built first for the food cold chain
Generative Gains
—Thermal Trace · by Generative Gains

Thank you.

Generative Gains is raising QAR 1.85M to prove Thermal Trace in the field: that what a failing fridge costs can be measured, and recovered.

  1. Najeeb AbdiMath · Frontend
  2. IslambekHardware engineer
  3. Param Anand TrimbakeFrontend
  4. Ahad HussainAI engineer
  5. Robin ThomasBusiness · Product
Generative Gains · Pre-seed · QAR 1.85M · Qatar first, GCC next github.com/Isl-d/Reboot-The-Earth-Cold-Chain-Logistics
← → navigate · F fullscreen · N notes · O overview · P PDF