Industrial AI research & products · Doha, Qatar

Oil & gas intelligence, from [ first principles ]

Systems that watch millions of live signals, forecast drift, hold operating envelopes, and explain every conclusion.

Talk to an engineer Meet Axiom →
LIVE · PLANT TELEMETRY SCROLL DOHA, QATAR
[ 01 ] · The flagship instrument
Axiom

The ground truth for every asset you operate: built by agents, held to physics, trusted by engineers.

Axiom reconciles PI, SAP, and every maintenance record into one living knowledge graph, then keeps agents watching it around the clock, so the asset record never drifts from the plant again.

The result is an asset record that cannot lie.
SYSTEMS OF RECORD AI AGENTS VERIFIED ASSET GRAPH PI SAP CMMS DOCS RECONCILE VALIDATE TRACE MONITOR
Agent activity · live trace 4 agents running
14,208 tags contextualized
612 assets resolved
98.7% cross-system agreement
See Axiom in your plant →
GTG-805 · BEARING TEMP TREND ▲ WATCH 23-TI-401D · 879.5°C · OUT OF SOC 1H 16M F-501 · EFFICIENCY 87.6% · GAP −1.2 PTS EMBER · $2.43M/YR RECOVERABLE MERIDIAN · $69.5M IN TRANSIT · 2 AT RISK AXIOM · 254 RECORD CONFLICTS ARBITRATED ENVELOPE · 41 LIMITS UNDER REVIEW SHIFT REPORT · AUTO-DRAFTED 05:45
GTG-805 · BEARING TEMP TREND ▲ WATCH 23-TI-401D · 879.5°C · OUT OF SOC 1H 16M F-501 · EFFICIENCY 87.6% · GAP −1.2 PTS EMBER · $2.43M/YR RECOVERABLE MERIDIAN · $69.5M IN TRANSIT · 2 AT RISK AXIOM · 254 RECORD CONFLICTS ARBITRATED ENVELOPE · 41 LIMITS UNDER REVIEW SHIFT REPORT · AUTO-DRAFTED 05:45

Catalyx builds AI products for oil & gas operators across the Middle East and the world: systems that watch millions of live signals, forecast drift, hold operating envelopes, and explain every conclusion. Trained on your data. Constrained by physics. Valuable from day one.

Talk to an engineer The instruments →
Day 1
Time to first value, by design
100+ yrs
Combined plant operations experience
14,208
Tags contextualized in one agent run
Any platform
Fabric · Databricks · Cognite · on-prem
Built with 100+ years of oil & gas experience
Shell bp Chevron ExxonMobil QatarEnergy Saudi Aramco ADNOC TotalEnergies
Where our engineers ran operations and led digital programs before building Catalyx.
[ 02 ] · What we do

Three disciplines.One operating edge.

Most industrial AI treats the plant as a dataset. We treat it as a physical system first, then let the models earn their place.

№ 01

Deep science

Thermodynamics, reaction kinetics, and transport phenomena anchor every model. Physics-informed learning constrains predictions with mass and energy balances, so they hold when conditions drift beyond the training data.

Eq. 1 · Physics-informed loss
ℒ(θ) = ℒdata + λ · ℒphysics
№ 02

Core engineering

Our people come from control rooms, turnarounds, and reliability programs, not slide decks. We know what a bad alarm shelf costs, what a spurious trip does to a quarter, and what a good insight is worth.

Fluencies
ISA-95 ISA-18.2 HAZOP API RP 754
№ 03

AI technology

Agentic pipelines, causal inference, and Bayesian survival modeling, applied where they compound the physics and the field experience, never as a substitute for them. Every enriched value traces to a source document.

Eq. 2 · Causal setpoint effect
Δy = 𝔼[ y | do(u) ] − 𝔼[ y ]
[ 03 ] · The instruments

Three instruments.Value from day one.

Each one is a research method made operational: pre-built for oil & gas, deployed on your data platform, inside your security perimeter.

CX Ember: furnace fleet efficiency, combustion diagnosis and fouling economics
FURNACE AI INTELLIGENCE
$2.43M/YR RECOVERABLE

CX Ember

Efficiency-gap attribution, combustion diagnosis, and fouling economics: recoverable value quantified per furnace.

METHOD · Combustion thermodynamics · Fouling kinetics
CX Envelope: operating windows, SOC and interlock limits with live excursion tracking
AI-DRIVEN OPERATING LIMITS
SIL-CREDITED · HAZOP-TRACED

CX Envelope

Every critical tag governed against its safe operating envelope: HAZOP lineage, live excursions, and AI-recommended limit revisions.

METHOD · Constraint modeling · HAZOP lineage
CX Meridian: global tracking of critical equipment and materials in motion
SUPPLY CHAIN INTELLIGENCE
$69.5M IN TRANSIT

CX Meridian

Critical equipment in motion across vessels and lanes: ETA drift and at-risk shipments flagged before they bite the turnaround.

METHOD · Spatiotemporal ETA inference
[ 04 ] · The method · CX Foundry

AI-agent data engineering,and the science it unlocks.

The hardest 80% of industrial AI is the data engineering nobody sees. Catalyx agents do that work: discovering schemas, resolving tags to assets, validating quality, keeping pipelines current, so advanced data science starts on day one, not month nine.

01 Discover Agents crawl historians, DCS exports, SAP tables, and document stores, building a live inventory of every signal the plant produces.
02 Resolve Tags map to assets, assets to functional locations, documents to both: an asset-centric model reconciled across systems of record.
03 Validate Quality rules, physical-range checks, and cross-system agreement scores: disagreements are surfaced for arbitration, never papered over.
04 Publish The governed model lands on your platform (Fabric, Databricks, or Cognite) with lineage intact and pipelines that self-heal on drift.
Agent run · live trace● running
▸ discover 14,208 PI tags · 3 historians
▸ resolve tags → 612 assets (SAP FLOC)
▸ validate 98.7% quality · 41 flags raised
▸ arbitrate PI vs SAP disagreement ×254
▸ publish asset model → Databricks
✓ contextualization complete · 2h 14m
Weeks → days
Contextualization cycle
Self-healing
Pipelines repair on drift
[ 05 ] · Data & platforms

Every signal your plant produces.Every platform you already run.

Sources we speak natively
SAP SCADA DCS OSIsoft PI LIMS CMMS Engineering docs

Time-series, transactional, and unstructured data, contextualized into a single asset-centric model by the Foundry agents.

Platforms we deploy on
Microsoft Fabric OneLake-native
Databricks Lakehouse & Unity Catalog
Cognite Data Fusion Industrial DataOps
Your environment Cloud, hybrid, or on-prem
[ 06 ] · The company

Founded in Doha.Fluent in the plant.

Catalyx Technology was established in Doha, Qatar by engineers who spent their careers inside oil & gas plants: commissioning units, running control rooms, and leading reliability programs. That background is our unfair advantage: we build for the constraints of real operations, and we serve operators across the Middle East and worldwide.

HQ · Doha, Qatar Middle East & Global Upstream · LNG · Refining · Petrochem
a. Run by plant people. Every product decision is made by someone who has held a shift log.
b. Evidence over demos. First engagements run on your data, against your baselines, in your perimeter.
c. Science that ships. Research-grade methods, delivered as products, not papers, not pilots that stall.
25+
Years energy experience
40+
Projects delivered
3
Continents served
4
AI products shipping
TRACK RECORD 25.2854° N, 51.5310° E
[ 07 ] · Begin

Bring a problem. Leave with [ evidence ]

Bring a sample of plant data to the first meeting. We'll bring the science, the engineers, and the agents, and show value against your own baseline.

Request a briefing hello@catalyx.tech