
Partners
Technology wedid not build,and say so.
We operate the infrastructure layer. Above it we partner with specialists — named, cited, and kept apart from our own record.
The integrated stack
One sovereign stack,silicon to Arabic AI.
Each layer runs in-country on Qode Core infrastructure. The partner platforms plug in as optional managed services.
- 66Applications & innovationUniversity AI Hub ventures; enterprise and government AI applicationsLaunchingQode Core AI Hub
- 55Arabic AI layerArabic language layers, Arabic retrieval, bilingual agents, right-to-left pages and portalsIn developmentQode Core
- 44Trust & governanceRuntime guardrails, tool-call validation, explainability, audit-ready evidenceTynapse
- 33Intelligence platformData ingestion, behavioural modelling, ML, generative and agentic AI, and IoT on one platformQODE Cortex · on Lorgan Enigma & Ellie
- 22Orchestration & monetisationProvisioning, workload routing, metering, billing, telemetry and utilisationHZ AI
- 11Sovereign AI infrastructure5 MW today toward 15 MW; Tier III to Tier IV; GPU clusters; UAE data residencyQode Core
- Delivered by Qode Core
- Partner platform, run by Qode Core in-country
- Not live yet
How the layers relate
Each layer runs in-country on Qode Core infrastructure. The partner platforms plug in as optional managed services — nothing above layer one is required to buy layer one, and every layer above it stays inside the same jurisdiction.
Insight 360 records what ran and on whose authority. The trust layer decides whether it should have run at all. They are adjacent disciplines and the site keeps them apart on purpose.

Tynapse · Trust layer
A model can tell youwhether it can act.Not whether it should.
Tynapse builds a verification layer between an AI system's decision and its execution: it checks policy, context and risk, then allows or stops the action — and logs why.
Qode Core and Tynapse have signed a partnership, and we have adopted their trust layer as the governance tier above our platform stratum.
Why governance has to run where the model runs
Tynapse builds a verification layer that sits between an AI system's decision and its execution. Before an action is taken, it checks the basis for the decision, the organisation's policies, the work context and the risk level, and returns a decision to allow or to stop — logging the basis and the policies it applied, so the call can be audited afterwards.
Qode Core and Tynapse have signed a partnership, and we have adopted their trust layer as the governance tier above our platform stratum. Where a customer's workload requires decisions to be verified before they execute, this is the mechanism we deploy — inside the customer's perimeter, on Qode Core infrastructure, under the same jurisdiction as everything beneath it.
A verification layer that sits in another provider's cloud creates exactly the dependency that sovereign infrastructure exists to remove: every decision your system makes would be adjudicated outside your jurisdiction, by an operator you do not control, on a path you cannot audit. Tynapse ships a customer-managed deployment — a containerised service that runs in the customer's own environment, including network-separated and air-gapped ones. That is the only form of it we would put on this platform.
The governance tier and the compute beneath it end up under one legal regime, which is the whole argument of this company applied one layer higher.
The trust layer · as published by Tynapse
Three components,one record.
Tynapse describes the layer in three parts. The outputs are a policy decision, an evidence trail, an attestation and — where it is warranted — a human review.
01
Detect
GuardScreen. Screens incoming prompts, retrieval context and tool calls before they reach the model, stopping unsafe or out-of-policy requests in real time.
02
Trust Judge
Adjudication. Evaluates each model output against internal policy, business rules and domain requirements before it is delivered or acted on.
03
Trust Ops
Record. Captures a structured, replayable record of behaviour for monitoring, audit and accountability — what happened, why, and what was applied.
Outputs
- Policy decision
- Evidence trail
- Attestation
- Operator review
Trust attestation set
- Status
- Sealed
- Decision
- AGT-7F3A · tool_call.execute
- Policy
- FIN-RISK-07 · verified PASS
- Model
- guardrail-v3 · context checked
- Latency
- 200 ms
- Integrity
- sha256
Example record, as published by Tynapse
Risk coverage — beyond a content filter
Toxic language and leaked personal data are the part of this problem the market already has tools for. The categories below are the ones that appear once an AI system is allowed to take actions rather than produce text.
Standard screening
- Toxic language
- Personal data leakage
Action-level governance
- Tool and transaction misuse — orders, refunds, approvals
- Rule bypass and manipulation, such as promotion abuse
- High-stakes decision errors — status, claims, compensation
The distinction is the one the whole partnership turns on. A system that only generates text has a content problem. A system that can execute has a governance problem, and it is not solved at the same layer.
Tynapse's record — theirs, not ours
The following are facts about Tynapse, drawn from their published material and from the press report linked below. They are reproduced here because a buyer is entitled to know who is in the stack. None of them is a Qode Core credential, and none of them should be read as one.
NVIDIA Inception Grand Challenge 2026Digital Today
Named in the top ten at the Asia-Pacific finals, held in Singapore on 22 September 2026. The company reported the result on 30 September 2026. Entrants were judged on technical capability and business viability.
Financial services deploymentDigital Today
The technology was tested in production settings including finance, through a project with KB Kookmin Bank verifying an AI financial advisory service.
RecognitionDigital Today
That project received the Financial Services Commission chairman's Cooperation Award of the Year, reported in June.
RuntimeDigital Today
For real-time use at scale, Tynapse states it is building its runtime on NVIDIA Hopper-series GPUs, using TensorRT-LLM and FP8 inference optimisation.
Deployment modelsTynapse
Tynapse publishes two: a managed secure cloud, and a customer-managed containerised deployment for closed, network-separated environments including private VPC and air-gapped. Qode Core deploys the second.
Frameworks referencedTynapse
Tynapse's material references the EU AI Act, Korea's AI Basic Act, ISO 42001, the NIST AI Risk Management Framework and the OWASP LLM Top 10 as the regimes the layer is built against.
Runtime guardrail frameworkTynapse
Tynapse states it is co-developing an AI-agent runtime guardrail validation framework with KB Kookmin Bank, Korea's largest commercial bank — a proof of concept, with the production path under evaluation.
Information securityTynapse
Tynapse's site lists ISO/IEC 27001:2022 certification for its information security management system.
A partner's accreditations do not transfer. Tynapse's record — like HZ AI's and Lorgan's below — speaks to the component Qode Core deploys, not to Qode Core's own controls, which are assessed on their own terms.

HZ AI · Orchestration
You own the hardware.HZ brings the demand.
A marketplace and control plane for GPU capacity. Operators set the price per GPU-hour; HZ AI brings the buyers and settles every job to its runtime.
Its platform is the orchestration and monetisation layer of our stack — and the platform beneath the capacity partnership we offer to other facility operators.
The platform, in HZ AI's terms
GPU capacity orchestration and monetisation · Burlingame · California
HZ AI runs a marketplace and control plane for GPU capacity. Operators list the NVIDIA GPUs they choose to sell and set their own price per GPU-hour; HZ AI brings the buyers, runs each workload isolated on a dedicated GPU, and settles every job to its runtime.
As described by HZ AI
Partnership
Qode Core has partnered with HZ AI. Its platform is the orchestration and monetisation layer of our stack: it puts surplus and burst capacity in front of buyers, meters it and settles it — and it is the platform beneath the capacity partnership we offer to other facility operators. Dedicated partner halls stay fully ring-fenced from it.
The platform, in HZ AI's terms
Demand aggregationHZ AI
Buyer workloads brought to listed capacity, at the price per GPU-hour the operator sets.
Isolated workloadsHZ AI
Each workload runs isolated on a dedicated GPU.
One control planeHZ AI
Every connected facility joins one pool of GPU capacity, managed from one control plane.
Metering & settlementHZ AI
A hold is placed on the buyer's prepaid credit before a job starts; every job is settled to its runtime.
Hardware supply — GPU servers specified and built for operators' racks
- NVIDIA H100
- H200
- B200
- B300
- RTX PRO
- AMD Instinct MI300X
- MI325X
- MI355X
Sources

Lorgan · QODE Cortex
One platform forAI, GenAI and IoT.
QODE Cortex is Qode Core's intelligence platform, built on Lorgan's Enigma and Ellie and hosted in-country on Qode Core infrastructure.
Data ingestion, behavioural modelling, machine learning and generative and agentic AI on one governed foundation.
What QODE Cortex brings together
Enigma and Ellie · New South Wales · Australia
Lorgan is a product and professional services company based in Australia. Its flagship platforms are Enigma, which enables system orchestration, data unification and AI-powered control across IoT and enterprise environments, and Ellie, which delivers context-aware insights, autonomous agents and continuous learning tailored to specific domains.
As listed in the National AI Centre's AI Directory
QODE Cortex
QODE Cortex is Qode Core's intelligence platform, built on Lorgan's Enigma and Ellie and hosted in-country on Qode Core infrastructure. It brings data ingestion, behavioural modelling, machine learning and generative and agentic AI onto one governed foundation — so a tenant reaches production without assembling a separate data, ML and GenAI toolchain first.
What QODE Cortex brings together
Data ingestion & IoTQode Core
Batch and real-time ingestion, including IoT streams, with automated data quality.
Behavioural modellingQode Core
Automated feature engineering that models the behaviour of customers, products and devices.
ML & analyticsQode Core
Build, explore and visualise machine-learning models and insights in one tool.
Generative & agentic AIQode Core
GenAI and agentic capability, through Ellie, on the same governed data foundation.
Industries Lorgan works across
- Energy
- Government
- Telecommunications
- Healthcare
- Finance
- Sports
- Manufacturing
Sources
- National AI CentreAI Directory — Lorgan Pty Ltd ↗
- Qode CoreQODE Cortex — our product, our description
Partner with us
Two ways in.
Supply capacity, or build what belongs above the infrastructure layer.
What we look for
- It deploys inside the customer's perimeter, not only as someone else's cloud service
- It has a published record we can cite rather than describe
- It solves a problem we would otherwise have to build badly
The two routes, in full
Capacity partnership. You operate a facility with GPUs that sit idle. With HZ AI, we bring the buyer, the serving stack and the billing, and you earn on hardware that would otherwise depreciate dark.
Technology partnership. You build a capability that belongs above the infrastructure layer and runs inside a customer's perimeter. Tynapse, HZ AI and Lorgan meet these criteria; the next partner will be held to the same ones.

Partnerships
Tell us whereyou would sit.
Supply, technology, or something the two categories above do not cover. The first conversation is short and it is with someone technical.


