◆ Gentoo Logic · Modeling Warehouse

Architecture comparison

Waymo vs Wayve

Waymo runs closer to the foundational model (Native (own / self-hosted weights)) than Wayve (Native (own / self-hosted weights)). Shared foundation: proprietary / self-built models (highlighted).

Waymo

Autonomous

Proximity
Native (own / self-hosted weights)
Models
proprietary / self-built models
Cloud · datastore
GCPBigtableSpannerGoogle Cloud StorageProprietary in-vehicle data stores
Compliance
ISO 27001 (inferred via Alphabet/Google Cloud hosting)[4]SOC 2 (inferred via Alphabet/Google Cloud ecosystem)[4]Automotive safety and validation frameworks for AVs, including systematic safety case construction[1][8]
Architecture

Waymo’s production driver uses a modular six-layer stack (sensor processing, perception, world modeling, prediction, planning, control) integrated with HD maps, simulation, and a foundation-model-based end-to-end driving architecture that fuses camera, lidar, and radar into a unified world representation.[1][2][8][9][10] Training, evaluation, and large-scale data handling run on Google/Alphabet’s internal and Google Cloud infrastructure, with closed-source models and tooling.[1][4][5][8][10]

Wayve

Autonomous

Proximity
Native (own / self-hosted weights)
Models
proprietary / self-built models
Cloud · datastore
AzureAzure Storage (Blob/Object)[7]Azure Databricks (Delta Lake/Parquet over object storage)[7]Azure Machine Learning experiment/model metadata stores[9][12]Internal time-series/telemetry stores for fleet data (unspecified, likely on Azure)[9][12]
Compliance
Azure built-in compliance (ISO 27001, SOC 1/2/3, GDPR support) inferred from exclusive Azure usage[7][12]Wayve safety framework for AV (Safety 2.0 paradigm, not a formal security cert but core to production governance)[8]
Architecture

Wayve runs a proprietary end-to-end embodied driving stack on Microsoft Azure, using large vision-world models (AI Driver, GAIA-3) trained on petabyte-scale fleet and simulated data via Azure Kubernetes Service, Azure Machine Learning, Databricks, and GPU clusters, then deploys a compressed model on in-vehicle compute for production driving.[7][9][12] The architecture replaces modular perception/mapping/planning with a single neural policy and related foundation models (driving, simulation, scenario classification, language explanation), all controlled via internal MLOps pipelines on Azure.[1][4][9][12]

← Full orbit map · Score your own stack →