Arize AI
Arize AI provides an enterprise AI engineering platform for tracing, evaluating, and improving LLM applications and AI agents, serving hundreds of enterprises across travel, retail, financial services, telecom, developer platforms, and federal/defense through its Arize AX SaaS and open-source Phoenix library.
- Company typePrivate
- Founded2020
- HeadquartersSan Francisco, United States
- Headcount101–250
- GTM typeB2B
- OfferingSoftware
What Arize AI does
Arize AI is a venture-backed private software company founded in 2020 and headquartered in Berkeley, California, that builds an AI engineering platform for observability, evaluation, and improvement of LLM applications and AI agents. The company's commercial product is Arize AX, which ships in two modules — Arize AX Generative (LLM/agent tracing, prompt management, online/offline evals, guardrails, sessions, and experiment management) and Arize AX ML & CV Observability (drift detection, performance heatmaps, explainability, and embedding drift for production ML/CV models) — both powered by a proprietary adb data store that connects natively to BigQuery, Databricks Unity Catalog, and Snowflake. The platform is built on Arize's own OpenInference semantic-conventions standard layered atop OpenTelemetry, and is paired with Phoenix, an open-source LLM observability library (10.2k+ GitHub stars, 3M+ monthly downloads) that anchors a developer-led go-to-market motion. Arize monetizes via tiered SaaS subscriptions (AX Free, AX Pro at $50/month, AX Enterprise custom), usage-based overages on spans and ingestion, freemium/open-source funnels into Pro and Enterprise, federal contracts (AFWERX SBIR), and partner-delivered services through an SI/consulting network including Infogain. Its named customer base spans hundreds of enterprises across travel (Booking.com, TripAdvisor, Priceline, Hyatt, Uber), retail/consumer (PepsiCo, Wayfair, Duolingo, AB InBev), financial services (America First Credit Union, Bill), developer platforms (Microsoft, GitHub, Twilio, MongoDB, eBay, ByteDance, Atlassian), telecom (AT&T), and government/defense (AFWERX, Defense Innovation Unit), with hyperscaler co-sell relationships with Google Cloud, Microsoft Azure, AWS, and NVIDIA. As of the most recent disclosures, Arize has raised approximately $131M in equity funding across four rounds (most recent: a $70M Series C in February 2025 led by Adams Street Partners and including M12, Datadog, PagerDuty, and Battery Ventures), sits at a reported $600M–$750M valuation band, and employs 101–250 staff across Berkeley, New York, and remote hubs in EMEA, APJ, and LATAM.
Arize AI firmographics
Firmographics- Name
- Arize AI
- Legal name
- Arize AI, Inc.
- Website
- https://arize.com
- Company type
- Private
- Founded year
- 2020
- Operating status
- Operating
- Headcount range
- 101–250 employees
- Short description
- Arize AI provides an enterprise AI engineering platform for tracing, evaluating, and improving LLM applications and AI agents, serving hundreds of enterprises across travel, retail, financial services, telecom, developer platforms, and federal/defense through its Arize AX SaaS and open-source Phoenix library.
- Ownership category
- akta.pro rank
Arize AI industry classification
Industry- Product category
- AI Observability / AI Engineering Platform
- NAICS
- Software Publishers (513210), Custom Computer Programming Services (541511)
- SIC
- Services-Computer Integrated Systems Design (7373), Services-Computer Programming Services (7371)
- akta.pro primary industry
- AI Observability, Monitoring & Evaluation Platforms (Drift, Quality, Safety) (HDAEANAF)
- akta.pro secondary industries
- End-to-End Enterprise AI Platforms (MLOps & Model Lifecycle Management) (HDAEANAA), AI Application Enablement Platforms (Copilot/Agent Frameworks, SDKs) (HDAEANAJ)
Keywords
Where Arize AI is headquartered
LocationHeadquarters
- HQ city
- San Francisco
- HQ country
- United States
- HQ region
- North America
Offices5 records
Markets served
Arize AI business model
Business model- GTM type
- B2B
- Offering type
- Software
- Cost components
- Personnel, Technology or R&D, Infrastructure, Marketing or Sales, Operations
Revenue model
- SaaS subscriptions (Arize AX): Tiered SaaS subscriptions for Arize AX: AX Free ($0), AX Pro ($50/month), AX Enterprise (custom). Includes paid Pro and Enterprise SKUs plus self-hosted Enterprise with data residency and multi-region options.
- Usage-based overage on spans and ingestion: Usage-based overage on top of Pro tier limits: $.0008 per additional span and $3 per additional GB of ingestion; consumed customers move to Enterprise custom contracts.
- Freemium / open-source funnel: Free AX tier plus Phoenix open-source (10.2k+ GitHub stars, 3M+ monthly downloads) used as a land motion feeding Pro and Enterprise conversions; startup pricing program also offered.
- Enterprise / federal contracts and partner-delivered services: Custom enterprise contracts (with dedicated support, uptime SLA, SOC2/HIPAA reports, training sessions) plus federal contracts (AFWERX direct-to-Phase II SBIR) and Global Service Network partner implementations of Arize.
Pricing tiers
| Model | Billing | Price |
|---|---|---|
| Freemium | Pay-as-you-go | AX Free — $0 for individuals/startups; 25k spans/month, 1 GB/month, 15-day retention; includes Alyx, online evals, product observability, community support. |
| Hybrid | Monthly | AX Pro — $50/month for small teams and startups; 50k spans/month, 10 GB/month, 30-day retention; higher rate limits, longer retention, email support. |
| Other | Multi-year contract | AX Enterprise — Custom pricing for enterprises; SaaS or Self-Hosted; includes dedicated support, uptime SLA, SOC2 reports and HIPAA, training sessions, adb Data Fabric, multi-region deployments and data residency. |
| Usage-based | Pay-as-you-go | Usage-based add-ons: $.0008 per additional span and $3 per additional GB of ingestion beyond Pro tier limits. |
Go-to-market motion1 record
Distribution channels1 record
Marketing channels11 records
Arize AI product offering
Product offeringCore offering
Arize AI sells Arize AX, an enterprise AI engineering platform that provides tracing, evaluation, and improvement workflows for LLM applications and AI agents in production. The platform is built on the company's OpenInference open standard (on top of OpenTelemetry) and the proprietary adb data store, and is sold alongside the open-source Phoenix library. Customers subscribe to tiered SaaS plans (Free, Pro, Enterprise) with usage-based overages and optional self-hosted deployments.
Product overview
Arize AI offers a platform-plus-modules architecture centered on two main commercial products — Arize AX (Generative) and Arize AX (ML & CV Observability) — built on top of the open-source Phoenix library and the OpenInference semantic-convention standard. Arize AX (Generative) provides end-to-end LLM/agent observability, evaluation, and improvement; Arize AX (ML & CV Observability) provides production ML/CV model monitoring, drift detection, and explainability. Both AX editions are powered by adb (Arize Data Fabric), the company's high-scale GenAI trace datastore that natively connects to BigQuery, Databricks Unity Catalog, and Snowflake. The in-app AI engineering agent Alyx (within Arize AX) and the parallel PXI agent (within Phoenix) automate root cause analysis, synthetic-data generation, eval creation, and prompt optimization, while OpenInference provides the cross-stack OpenTelemetry-based instrumentation layer. Phoenix itself is the free, open-source entry point used by AI engineers and Fortune 500 teams alike.
Differentiator
Problem solved
Functional benefit
Brands
- Arize AX: Enterprise AI engineering platform for generative and ML/CV observability; the primary commercial product line.
- Arize Phoenix
- Alyx
- OpenInference
- adb (Arize Data Fabric)
- PXI
Products and services
- Arize AX (Generative) Enterprise AI engineering platform for building, observing, evaluating, and improving generative AI applications and AI agents in production. Provides tracing, prompt playground and management, datasets/experiments, online and offline evaluations, search/curation, guardrails, monitors, and human annotations for ML engineers, AI engineers, and platform/MLOps teams at large enterprises.
- Arize AX (ML & CV Observability) Unified platform for ML engineering teams to monitor, debug, and improve traditional ML model performance in production, including performance tracing, explainability, automated dashboards/monitors, model and feature drift detection, embedding drift monitoring, cluster search, and dataset building across training, validation, and production environments.
- Arize Phoenix Open-source AI observability and evaluation platform for tracing, evaluating, experimenting, and iterating on LLM applications. ELv2 licensed with 10.2k+ GitHub stars and 3M+ monthly downloads; supports local, Docker, Kubernetes, and Cloud deployment; built on OpenTelemetry/OpenInference standards; now includes the PXI AI engineering agent.
- Alyx Arize's AI engineering agent that plans and executes complex AI engineering workflows autonomously, including instant root cause analysis, synthetic data generation, auto-generated evaluations, and prompt experimentation. Implemented with the same autonomous orchestration approach as Cursor and Claude Code, applied to AI engineering.
- adb (Arize Data Fabric) Managed datastore for GenAI traces, storing agent trajectories and context in open formats and connecting natively to BigQuery, Databricks (Unity Catalog), and Snowflake. Powers Arize AX's 1 trillion spans/month ingest and 1 billion evals/month throughput.
Quantifiable outcome
- 1 Trillion spans per month and 1 Billion evals per month processed on Arize AX
- +5 more outcomes
Companies that use Arize AI
Customer profileNamed customers46 records
Segments6 records
Ideal customer profiles2 records
Arize AI technology and API
TechnologyTechnology focussed Yes
API detail
- Has API
- Yes
- API docs
- API detail
Core technology
AI maturity
App detail
Integration22 records
AI capability16 records
Feature9 records
Arize AI partnerships and signals
Strategic signalPartnerships
16 partnerships are on record, tiered flagship strategic partner, standards partner, partner program member, implementation partner, regional channel partner, federal/government partner, acquired company, core technology partner and strategic partner.
- Google Cloudflagship strategic partnerPartnership to promote standardized telemetry for enterprise AI agents, leveraging OpenTelemetry and OpenInference standards for consistent observability across frameworks, models, and platforms. Google Cloud's Principal Architect AI/ML publicly endorsed Arize on the partners page.
- Open Agent Specification (BeeAI) / Linux Foundation ecosystemstandards partnerIntegration announced between the Open Agent Specification and Arize Phoenix/OpenInference to enable runtime-agnostic tracing, evaluation, and debugging for AI agents across different runtimes.
- Chronospherepartner program memberChronosphere's Partner Program integrates Arize alongside Embrace, Polar Signals, Checkly, and Rootly as specialized observability and reliability vendors to differentiate from all-in-one Datadog/Dynatrace/Splunk offerings.
- Infogainimplementation partnerStrategic partnership integrating Arize's observability platform with Infogain's Ignis AI innovation platform to help organizations evaluate, monitor, and scale complex AI systems from prototype to production.
- Tokyo Electron Deviceregional channel partnerDistribution and resale partnership covering Arize AI's enterprise platform in Japan.
- AFWERX (U.S. Air Force)federal/government partnerAFWERX selected Arize AI's AI Engineering Platform for R&D under a direct-to-Phase II SBIR contract.
- Velvetacquired companyArize AI acquired Velvet, an AI observability platform, in March 2025 to expand support for AI observability and LLM evaluation; Velvet's team and technology enhance Arize's model analysis, diagnostics, and performance offerings.
- Microsoft (Azure / M12)flagship strategic partnerStrategic collaboration announced November 2024 to enable more effective enterprise deployment of generative AI on Azure, integrating Arize's evaluation and tracing capabilities with Azure AI services. Deepened through the Series C in February 2025 with Microsoft Ventures (M12) participating and expanded Azure AI integration.
- LlamaIndexcore technology partnerJoint platform for evaluating LLM applications, with a one-click integration between LlamaIndex and Phoenix. LlamaIndex CEO Jerry Liu publicly endorsed Phoenix.
- AWS (Amazon Web Services)flagship strategic partnerStrategic partnership covering Amazon Bedrock Agents observability, joint case studies (TheFork on AWS), and an in-person AWS AI Builder Lab event with LlamaIndex, Anthropic, and Arize in San Francisco. AWS's WW Agentic AI Product Lead endorsed Arize on the partners page.
- NVIDIAcore technology partnerTechnology integration with NVIDIA NeMo and NIM microservices (AT&T deployed NeMo-based customer care agents with Arize AI for evaluation). NVIDIA's Senior PM, Generative AI endorsed the partnership.
- Databrickscore technology partnerIntegration enabling customers to send one OpenTelemetry trace stream to both Arize AX and Databricks Unity Catalog; Arize Data Fabric now supports Databricks for syncing production agent traces, evaluations, and annotations into customer-owned storage.
- MongoDBstrategic partnerStrategic partnership on AI application observability, especially RAG, supporting engineering teams tracing data flow through AI systems.
- Couchbasestrategic partnerStrategic partnership bringing monitoring, evaluation, and optimization capabilities to AI-driven applications built on Couchbase.
- PagerDutystrategic partnerIntegration enabling richer alert metadata and faster model debugging; PagerDuty's Senior Director Data Science publicly endorsed Arize on the partners page.
- CrewAIstrategic partnerJoint focus on observability and evaluation for CrewAI-built multi-agent systems in production; CrewAI's CEO publicly endorsed Arize on the partners page.
Scale indicators11 records
Recent moves6 records
Expansion highlights6 records
Arize AI competitors and assessment
Company assessmentDirect peers
- LangSmith (LangChain): LangSmith is the first-party tracing and evaluation product from LangChain, the dominant LLM framework vendor. It competes head-to-head with Arize AX in the AI engineering platform category, especially for teams building on LangChain/LangGraph, and benefits from tight framework integration that Arize must integrate around.
- Langfuse: Open-source LLM observability and evaluation platform with a hosted cloud offering. It is an open-core direct competitor to Arize Phoenix plus AX Pro and is frequently compared alongside Phoenix in OSS LLM observability discussions.
- Comet (Opik): Comet's Opik is an open-source LLM evaluation and observability toolkit integrated with the broader Comet ML experiment tracking platform. Direct overlap with Arize AX on LLM-as-a-judge, online/offline evals, and tracing — explicitly benchmarked against Phoenix in independent tests.
- Braintrust: Braintrust is an AI evaluation and observability platform ranked #1 in valuation among AI governance startups at ~$800M. It directly competes with Arize AX in LLM evaluation, prompt management, and online evals, with comparable enterprise positioning.
- Helicone: Open-source LLM observability proxy focused on logging, evals, and cost tracking for production LLM traffic. Competes with Arize on lightweight observability for AI engineers and has overlap with Phoenix's developer-led entry tier.
- WhyLabs: AI observability platform focused on data quality, drift detection, and model monitoring for production ML/LLM systems. Direct overlap with Arize AX ML/CV Observability and Phoenix, particularly on embedding drift and quality monitoring.
Broad incumbents
- Datadog: Datadog is the largest infrastructure observability platform and is expanding aggressively into AI/LLM monitoring (LLM Observability, AI Guard, Watchdog for AI). It is both an Arize strategic investor and a direct competitor at the enterprise layer, where Datadog can bundle native AI observability into existing contracts.
- Dynatrace: Enterprise observability and APM incumbent with Davis AI and growing AI/ML monitoring capabilities. Competes with Arize at the enterprise procurement layer, particularly in regulated industries, as a broader one-stop observability vendor.
- Splunk (Cisco): Splunk's observability and log analytics stack (now part of Cisco) is increasingly positioning around AI workload monitoring. It competes with Arize for enterprise AI/ML monitoring budgets and benefits from Cisco's broader enterprise footprint.
- Honeycomb.io: Observability platform for production engineering teams with high-cardinality tracing. Adjacent to Arize in the broader observability category and a potential integration partner for OTel-based AI telemetry.
Market position
Strengths4 records
Weaknesses4 records
Competitive moat5 records
Key risks5 records
Key highlights7 records
Customer concentration
Arize AI social profiles
Digital presenceArize AI compliance and trust
Trust signalCompliance5 records
Arize AI financial estimates
Financial estimateRevenue estimate
Valuation estimate
Arize AI leadership team
Management profileNumber of profiles
Profiles14 records
Arize AI subsidiaries and ownership
Company hierarchySubsidiaries1 record
Arize AI funding detail
Funding detailFunding overview
Funding rounds4 records
Investors14 records
Funding detail is available on the Subscription and Enterprise plan.Contact sales →
Arize AI M&A and investment
M&A and investmentM&A2 records
Investments
M&A and investment is available on the Subscription and Enterprise plan.Contact sales →
Frequently asked questions about Arize AI
What does Arize AI do?
Arize AI sells Arize AX, an enterprise AI engineering platform that provides tracing, evaluation, and improvement workflows for LLM applications and AI agents in production. The platform is built on the company's OpenInference open standard (on top of OpenTelemetry) and the proprietary adb data store, and is sold alongside the open-source Phoenix library. Customers subscribe to tiered SaaS plans (Free, Pro, Enterprise) with usage-based overages and optional self-hosted deployments.
Is Arize AI a public or private company?
Arize AI is a private company. It is classified as venture growth investor backed and is currently operating.
When was Arize AI founded?
Arize AI was founded in 2020. It employs 101 to 250 people.
Where is Arize AI based?
Arize AI is headquartered in San Francisco, United States, in the North America region.
How does Arize AI make money?
Four revenue lines are on record. SaaS subscriptions (Arize AX) is the primary driver. The others are usage-based overage on spans and ingestion, freemium / open-source funnel and enterprise / federal contracts and partner-delivered services.
Who are Arize AI's main competitors?
Direct peers on record are LangSmith (LangChain), Langfuse, Comet (Opik), Braintrust, Helicone and WhyLabs. Broad incumbents are Datadog, Dynatrace, Splunk (Cisco) and Honeycomb.io.
Does Arize AI have an API?
Yes. Arize AI offers a full REST API for programmatic prompt management with systematic versioning, allowing users to create, list, retrieve, and delete prompts across projects. The platform also exposes SDKs (Python and JavaScript) and an OpenTelemetry-compatible ingest API for sending traces, spans, and evaluation telemetry to Arize AX. The Python SDK features a redesigned interface centered on a single ArizeClient with a standardized client pattern and configurable arguments. Developer documentation is at arize.com/docs/ax.
What industry is Arize AI in?
Arize AI's product category is AI Observability / AI Engineering Platform. Its primary akta.pro industry code is HDAEANAF, AI Observability, Monitoring & Evaluation Platforms (Drift, Quality, Safety), with a secondary code of HDAEANAA, End-to-End Enterprise AI Platforms (MLOps & Model Lifecycle Management). Its NAICS code is 513210 and its SIC code is 7373.