Skip to content
HolonomiX

HolonomiX products

7 products developed around structural computing and independently verifiable evidence. Compare the problem each product addresses, the result to evaluate and the deployment route for your team.

Product roles

Start with HX-SDP, the flagship platform, when evaluating consolidation of AI inference data services. Choose HX-Provenance when critical records need portable evidence that remains verifiable after handoff; it is the deployed cloud product. The specialized systems address cryptographic transition, thermal decisions, multidisciplinary optimization agent-evaluation incident reconstruction and mission decision support. Each product page identifies the intended team, evaluation inputs, evidence and limits.

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.

Runtime and deployment boundaries

HX-SDP serving data plane

Structural retrieval, permission filtering, reranking and top-k assembly share the HX-SDP data plane. Model prefill and decoding remain in the inference runtime. The consolidation benefit depends on the workload and services in scope.

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-Provenance customer-controlled deployment

The Google Cloud appliance runs in the customer's project. Applications submit records to the appliance, while records, signing keys and optional storage remain under customer control. The published diagram makes the runtime and custody boundary explicit.

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.

Deployment availability

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

HX-SDP is available through Private Appliance and scoped pilot only. HX-Provenance has Google Cloud and AWS marketplace routes as well as private deployment. HX-AEIR is available on Google Cloud Marketplace, with Private Appliance and scoped pilot routes also available. HX-PQC and HX-AIFactoryTwin are available through Private Appliance or scoped pilot. HX-MDAO-TP is available through a scoped pilot. HX-Fusion is a published local runtime release candidate for scoped integration.

Product-specific evidence

The evidence discipline is shared, but the artifacts serve different purposes. A receipt identifies a signed result, a CBOM records cryptographic dependencies, and a validation bundle records a model run and its checks. Verification requirements depend on the product, release and signing mode.

ProductEvidence to review
HX-SDPExact-result and operation receipts, with workload-specific runtime and retrieval results
HX-ProvenanceSigned receipts and portable evidence bundles
HX-PQCCBOM, signed deterministic quotes, migration events, runtime evidence and posture snapshots
HX-AIFactoryTwinSigned execution evidence binds runtime profile, release, inputs and recorded outcome; Developer Edition benchmark cards identify requested and effective grids.
HX-MDAO-TPValidity status, convergence, acceptance results, provenance and formal certificates for specified invariants
HX-AEIRSigned receipts, deterministic evidence exports and offline verification records
HX-FusionRuntime validation and a bounded public-data exercise, with canonical result and packet recovery checks

Private deployment and scoped pilots

Private Appliance engagement defines the customer environment, integration requirements and key custody. A scoped pilot establishes the workload, scale, validity conditions, acceptance criteria and evidence required before broader deployment. Compare the measured result with your current process and account for infrastructure, integration, licensing and ongoing operation before the next commercial decision.

Discuss product deployment