Skip to content
HolonomiX
Menu

HolonomiX

The Data Representation-Topology Company

Products, platforms and bespoke software & computational engineering

HolonomiX develops products, platforms and bespoke software through data representation and computational engineering. We design data structures, numerical operators and execution paths around accuracy, memory, latency and deployment requirements. Our work spans scientific computing, data platforms, cryptographic lifecycle systems and mission decision support.

Products and platforms

HX-SDP is the flagship platform for teams consolidating AI inference data services. HX-Provenance is the deployed cloud proof point for teams preserving portable, independently verifiable records. Specialized systems address cryptographic transition, AI-factory thermal analysis, multidisciplinary engineering, agent-evaluation incident evidence and mission decision support. Each product has its own deployment route and evaluation requirements.

HX-SDP

Representation-first AI inference data platform

Consolidate structural classification, retrieval, serving, search, governance and proof-bearing operations into a representation-first data plane for suitable AI workloads.

HX-Provenance

Digital evidence beyond the originating system

Preserve signed receipts and portable evidence bundles for AI outputs, compliance records and other critical digital artifacts, with independent verification after handoff.

HX-PQC

The full post-quantum cryptographic lifecycle

Discover cryptographic dependencies, execute scoped migration, protect live workloads and continuously verify posture with signed lifecycle evidence.

HX-AIFactoryTwin

Thermal decisions before capital commitments

Evaluate thermal-fluid scenarios for high-density AI infrastructure with explicit model boundaries, validation states and signed evidence for each attested run.

HX-MDAO-TP

Validity-gated multidisciplinary analysis and optimization

Run coupled multidisciplinary analyses and optimization workflows within stated validity domains, with explicit convergence states and inspectable signed evidence.

HX-AEIR

Signed reconstruction evidence for agent-evaluation incidents

Validate, sign, store, export and independently verify the evidence used to reconstruct agent-evaluation incidents in a customer-controlled Google Cloud project. Available on Google Cloud Marketplace.

HX-Fusion

Evidence-backed mission decisions

Reconcile supported evidence into an answer, a request for missing information or abstention. The published local runtime candidate retains decision history and reproducible endpoint packets.

End-to-end HX-SDP latency path

The serving path has ten steps. Steps 03 to 06 run inside HX-SDP.

  1. 01Network, auth and input
  2. 02Query rewrite
  3. HX-SDP serving data plane
    1. 03Structural retrieve
    2. 04Permission filter
    3. 05Rerank and score
    4. 06Top-k assembly
  4. 07Queue
  5. 08Prefill
  6. 09First token
  7. 10Decode stream

HX-SDP advantage

  • Consolidates retrieval stages
  • Eliminates data-plane copies
  • Reduces serving-state overhead
  • Runs as one signed runtime

What changes

  • Retrieval, filtering, reranking and context assembly live inside HX-SDP.
  • Model serving and decoding remain in the inference runtime.
  • One structural representation across the hot path.

Data has structure.HolonomiX structural compute, rendered in one representation.

HX-SDP combines the serving path in a representation-first data plane.

HX-Provenance · Private Appliance on Google Cloud

The appliance VM, records and signing keys remain in the customer's Google Cloud project.

  • Customer applications

    Data and ML pipelines, CI/CD and compliance tooling.

    POST /v1/provenance/exactAdmin API key. Issues an ML-DSA-65 signed receipt.

  • Third-party auditors

    Regulators and your customers' customers.

    POST /v1/provenance/verifyPublic and read-only, no key. Or verify offline with the release-matched verifier.

  • Google Cloud Marketplace

    Delivers the VM image at deploy time. Not in the runtime path.

Google Cloud · End-customer tenant

The customer's GCP project, Google Cloud Marketplace Pattern 1. The whole product workload runs here, with no partner-tenant, on-premises or other-cloud components at runtime.

VPC network, customer-managed · ingress HTTPS :443 only

Compute Engine · HX-Provenance appliance VM

n4-standard-4 · 4 vCPU · 16 GiB · Ubuntu 24.04 LTS · CPU-only

  1. Front endnginx TLS termination :443TLS 1.2/1.3, AEAD only. Proxies requests to the service on loopback.
  2. Servicehx-provenance on FastAPI 127.0.0.1:8001ML-DSA-65 signing and verification (FIPS 204), with a per-instance key generated at first boot.
  3. Data planePersistent disks, Hyperdisk BalancedBoot 10 GiB · corpus data 100 GiB
  • Customer-provided, optional
  • Cloud Storage

    Your own bucket and project. Evidence-bundle exports.

  • Cloud Logging and Monitoring

    Your own project. Ops Agent ships service and nginx logs and host metrics.

Runtime data pathDeploy time or optional

Billable
Compute Engine, 1 × n4-standard-4. Hyperdisk Balanced, 10 GiB boot and 100 GiB data.
Customer-provided, optional
Cloud Storage bucket. Cloud Logging and Monitoring.
Not used by the appliance
Cloud SQL, BigQuery, GPU and TPU.
Data custody
HolonomiX holds no customer data. No partner tenant at runtime.

HX-Provenance keeps the appliance, keys and records within the customer environment.

HX-Provenance on Google Cloud

HolonomiX and Google Cloud help organizations preserve independently verifiable evidence for AI outputs, compliance records, cyber incident evidence, legal documents, and other business-critical records. Authorized reviewers can validate those records without relying on the systems where they were created.

HX-Provenance deploys inside the customer's existing Google Cloud environment. Records, signing keys and evidence storage remain under customer control, fitting into existing governance processes and applications.

Post-quantum digital signatures support verification after the originating system has been upgraded, migrated or retired. Retained evidence, the matching verifier and the expected issuer key provide the path for later review.

The joint solution brief reports 214.8 receipts per second across 1,000 requests with zero errors on its documented 2-vCPU deployment.

HX-Provenance and HX-Provenance for Vertex AI have separate Google Cloud listings. HX-Provenance is also deployed through AWS Marketplace. Trial eligibility and terms are specific to each listing.

First page of the HolonomiX and Google Cloud solution brief.
The published solution brief describes customer control, independent verification and the documented deployment.

Product availability

Each engagement defines the workload, deployment boundary, acceptance criteria, evidence package and next commercial decision before execution. HX-SDP is available through a Private Appliance or scoped pilot.

Availability reviewed 2026-09-05; HX-AEIR updated 2026-09-22; HX-Fusion updated 2026-09-22.
ProductAvailabilityAccess
HX-SDPPrivate Appliance · Scoped pilotRequest HX-SDP access
HX-ProvenanceGoogle Cloud · AWS · Private Appliance · Scoped pilotDeployment options
HX-PQCPrivate Appliance · Scoped pilotContact HX-PQC sales
HX-AIFactoryTwinPrivate Appliance · Scoped pilotDiscuss a thermal analysis
HX-MDAO-TPScoped pilotRequest an MDAO pilot
HX-AEIRGoogle Cloud Marketplace · Private Appliance · Scoped pilotView on Google Cloud Marketplace
HX-FusionLocal runtime release candidate · Scoped integrationDiscuss HX-Fusion integration

Claims resolve to artifacts

Commercial availability, technical performance, product status and verification are different forms of evidence. Each remains attached to its own scope and supporting record.

Measured results carry the workload, hardware, precision, baseline, methodology and limitations that produced them. Receipts, manifests, bundles and validation records provide an inspectable path from a claim to its supporting material.

Signing, custody and verification requirements are stated per product. Published records and gated proof packs remain identified in the evidence registry.

Verify an HX-Provenance example

The public demonstration contains a fictional record, its signed receipt, an evidence bundle and matching standalone verifiers. It exercises valid inputs, altered records and an incorrect issuer key. Verification runs offline after dependency installation.

This synthetic example demonstrates the mechanism. A valid signature establishes integrity under the expected key; it does not establish that the record's contents are correct.

Representation is the control surface

Data, memory and compute form one economic system. HolonomiX begins with the form of the data before execution starts, because representation determines how much state a system stores and expands for computation, how often it must move, and where useful operations can act.

The same data can create different demands on memory and compute. A representation that exposes useful structure can reduce stored state and support operations without expanding the full object. The benefit depends on workload structure, operation, fidelity requirements, representation rank and deployment conditions.

PropertyDense representationStructurally aware representation
StateExplicit values in dense arrays.A representation that uses the workload’s sparsity, low rank or other structure.
Memory movementDetermined by array layout, access patterns and the operation.Can decrease when locality or compact state avoids moving unnecessary data.
ComputationDense kernels act on the materialized arrays.Supported operations act on the chosen representation without expanding all of it.
ScalingStorage follows the dense array shape and precision.Storage and work follow the retained structure and its overhead.
A workload with useful structure can need less state, data movement or computation. These benefits depend on the representation and operations; dense kernels can be the better choice when structural overhead outweighs the savings.

Where representation and evidence change the decision

AI infrastructure

Decide how much serving state to materialize, what the inference data plane costs to operate, and how retrieval-dependent outputs are evidenced.

HX-SDP

Engineering systems

Evaluate designs and thermal scenarios before committing capital, with validity and convergence states that remain inspectable.

HX-AIFactoryTwin · HX-MDAO-TP

Cryptographic transition

Decide what to migrate, in what order, and how to demonstrate posture to reviewers across the post-quantum transition.

HX-PQC

High-consequence records

Preserve records that must remain independently verifiable after the originating application, vendor or infrastructure is gone.

HX-Provenance · HX-AEIR

Mission decisions

Reconcile changing inputs, retain uncertainty and contradictions, and identify what further evidence is needed before a supported answer is possible.

HX-Fusion

Product demonstrations and field notes

Watch product demonstrations, evidence walkthroughs and technical briefings. Follow deployment milestones, partner developments, representation research and commercialization through HolonomiX Field Notes.

Software and computational engineering

A defined technical engagement can deliver a numerical component, a product integration, a private appliance or a custom application.

Scientific software

Numerical methods, GPU execution, operator design, thermal-fluid simulation and multidisciplinary analysis.

Structural data platforms

Representation-aware indexing, retrieval, persistence and integration for AI and analytical systems.

Cryptographic lifecycle systems

Discovery, migration planning, key custody, portable evidence and verification tools.

Mission decision support

Deterministic data fusion, evidence-backed answers, APIs and custom applications.