Cloudian

Data Lakes/Lakehouses

Cloudian builds exabyte-scale object storage software designed to run on customer premises.

Reviewed by 7wData
Share Email

On this page

Profile

On-premises object storage software that mimics Amazon S3, designed for enterprises and government agencies to store exabytes of data locally and connect to AI models.

Cloudian builds exabyte-scale object storage software designed to run on customer premises. Founded in 2011 by Michael Tso and Hiroshi Ohta as a spinoff from Gemini Mobile Technologies (established 2001), the company initially targeted data-intensive applications with S3-compatible storage. It relocated to San Mateo, California in 2013 to accelerate North American growth.

The company has raised approximately $360 million across multiple rounds, with recent funding including a $60 million Series round in January 2023 and a $23 million growth equity round in June 2024. Cloudian serves approximately 900–1,000 customers globally, including major enterprises such as Pfizer, Verizon, the U.S. Department of Defense, and financial institutions like PostFinance, which has deployed over 2 petabytes of data on Cloudian infrastructure.

In November 2025, Cloudian launched HyperScale AI Data Platform built on NVIDIA's reference architecture, enabling enterprises to index unstructured documents (PDFs, images, spreadsheets) into vector form for real-time AI inference at the edge. The platform claims 8x faster vector operations than CPU-based alternatives and eliminates per-token cloud costs. Parallel moves included achieving NVIDIA-Certified Storage designation (March 2026), launching support for second-generation AWS Outposts racks (April 2025), announcing a strategic reseller partnership with CTERA (November 2025), and integrating PyTorch connectors with RDMA to optimize AI workloads.

Market surveys indicate 89 percent of organizations are repatriating data from public cloud and 93 percent are evaluating cloud-to-premises shifts for AI workloads, positioning Cloudian to benefit from sovereign AI and data residency trends. The company employs several hundred staff but has endured layoffs since 2022, with at least one major reduction announced with minimal severance.

Track Cloudian and 240+ vendors.

48k+ subscribers read the daily AI & data note. One email, both newsletters. Unsubscribe anytime.

Who buys this

  • Large enterprises running big data and analytics platforms (financial services, healthcare, manufacturing)
  • Government and public sector agencies requiring data sovereignty and regulatory compliance
  • Service providers and managed service operators building cloud-like services on customer premises
  • Organizations deploying AI inference at the edge and on-premises, avoiding cloud per-token costs

Publicly disclosed clients

  • PostFinance (Swiss financial institution; 2+ PB deployed)
  • U.S. Department of Defense
  • Pfizer
  • Verizon
  • State of California
  • Johns Hopkins University
  • National Cancer Institute
  • RaboBank

Strengths and what to watch

Strengths

  • Proven customer retention and scale: 900+ customers with multi-petabyte deployments; PostFinance alone manages 2 PB of regulatory, analytical, and transactional data.
  • AI-native positioning: HyperScale AIDP integrates NVIDIA GPUs and vector databases for on-premises sovereign AI, capturing rising demand for cloud repatriation (89% of organizations surveyed in 2025).
  • Cost and sovereignty advantage: Delivers S3-compatible storage at $0.03/GB with no per-token inference fees, appealing to cost-conscious enterprises and regulated industries.

Watch for

  • Layoff history and talent retention: Company endured significant workforce reductions since 2022 with minimal severance; ability to execute complex AI product launches depends on engineering continuity.
  • Competitive pressure from public cloud providers: AWS, Google Cloud, and Microsoft Azure are adding on-premises options (Outposts, Sovereign Cloud, Stack) and undercutting Cloudian on breadth of services.
  • Execution risk on HyperScale AI Data Platform: November 2025 launch is recent; customer adoption, reference architectures, and competitive validation against startups (e.g., Mistral, Hugging Face) remain unproven at scale.

Recent moves

Key Information

Industry
Data Lakes/Lakehouses
Founded
2001
Headquarters
San Mateo, California

Frequently Asked Questions

What is Cloudian and what does it do?

Cloudian builds exabyte-scale object storage software designed for on-premises deployment. Founded in 2011, it provides S3-compatible storage, allowing enterprises and government agencies to store massive datasets locally and connect directly to AI inference without relying on public cloud infrastructure or per-token costs.

Is Cloudian S3-compatible?

Yes. Cloudian delivers S3-compatible object storage that runs on your premises, eliminating cloud dependencies. It mirrors AWS S3's interface, so existing applications work without modification. At $0.03 per gigabyte with no per-token AI inference fees, it undercuts public cloud pricing significantly.

What is Cloudian's HyperScale AI Data Platform?

Launched in November 2025, HyperScale AI Data Platform indexes unstructured documents—PDFs, images, spreadsheets—into vector form for real-time on-premises AI inference. Built on NVIDIA's reference architecture, it delivers 8x faster vector operations than CPU alternatives while eliminating cloud per-token inference costs.

Why are organizations moving data from cloud back to on-premises?

Market surveys show 89 percent of organizations are repatriating data from public cloud due to cost, sovereignty, and compliance concerns. Cloudian enables on-premises AI inference and storage without per-token cloud fees, capturing the rising demand for data residency and sovereign AI workloads.

What types of organizations use Cloudian?

Cloudian serves 900+ customers globally, including Pfizer, Verizon, PostFinance (2+ petabytes deployed), U.S. Department of Defense, Johns Hopkins, and the State of California. Primary segments include large enterprises in finance and healthcare, government agencies requiring data sovereignty, and organizations deploying on-premises AI inference.

Does Cloudian work with NVIDIA GPUs for AI workloads?

Yes. Cloudian achieved NVIDIA-Certified Storage designation in March 2026 and integrates NVIDIA GPUs through HyperScale AI Data Platform, built on NVIDIA's reference architecture. PyTorch connectors with RDMA optimization enable high-performance AI inference on unstructured data at the edge without cloud costs.

How Cloudian compares

Direct head-to-head against 3 competitors. Picked by 7wData.

This company

Cloudian

Positioning
On-premises object storage software that mimics Amazon S3, designed for enterprises and government agencies to store exabytes of data locally and connect to AI models.
Customer segments
Large enterprises running big data and analytics platforms (financial services, healthcare, manufacturing)
Strengths
Proven customer retention and scale: 900+ customers with multi-petabyte deployments; PostFinance alone manages 2 PB of regulatory, analytical, and transactional data.
Watch for
Layoff history and talent retention: Company endured significant workforce reductions since 2022 with minimal severance; ability to execute complex AI product launches depends on engineering continuity.
Recent moves
Cloudian HyperStore Achieves NVIDIA-Certified Storage Designation, Enabling Exabyte-Scale AI Infrastructure

MinIO

Positioning
S3-compatible on-premises object storage for AI data and agentic computing at multi-exabyte scale; active across more than half the Fortune 500.
Customer segments
Fortune 500 enterprises and AI infrastructure teams self-managing exabyte-scale data lakes to avoid public cloud dependency.
Strengths
149% ARR growth; multiple 8-figure exabyte-scale enterprise deals closed in 2024; claims more than half of Fortune 500 as customers.
Watch for
Community Edition features stripped June 2025, project archived early 2026; self-hosted users face forced migration to commercial AIStor at $96,000 per year minimum.
Recent moves
Introduced AIStor Free and Enterprise Lite subscription tiers December 2025; AIStor Tables reached general availability February 2026.

NetApp StorageGRID

Positioning
On-premises and hybrid object storage integrated inside NetApp's unified storage portfolio; targets enterprises modernizing unstructured data infrastructure.
Customer segments
Large enterprises and regulated industries already committed to NetApp; IT storage architects managing multi-petabyte unstructured data environments.
Strengths
StorageGRID 12.0 (September 2025) adds AI-optimized caching and cross-grid replication; named GigaOm Object Storage Leader and Fast Mover 2025.
Watch for
Mindshare declined from 5.3% to 5.1% year-on-year by January 2026; buyers cite mandatory NetApp ecosystem commitment as a lock-in risk.
Recent moves
StorageGRID 12.0 released October 2025 with AI caching layer, 60TB flash drive support, and cross-grid bucket branching for isolated datasets.

Scality

Positioning
Enterprise S3-compatible and NFS object storage for data sovereignty, cyber resilience, and policy-driven tiering across edge, core, and cloud.
Customer segments
Regulated industries, media organizations, and government agencies requiring on-premises sovereignty with hybrid cloud tiering.
Strengths
ARTESCA Cyber Guarantee ($100,000 per customer, March 2026) and Autonomous Data Infrastructure with agent-driven policy tiering launched May 2026.
Watch for
Dual-product portfolio (RING for petabyte scale, ARTESCA for smaller deployments) complicates procurement and support scoping for buyers with mixed environments.
Recent moves
Autonomous Data Infrastructure (ADI) launched May 2026, adding agent-driven policy tiering across four storage tiers; F5 partnership expanded February 2026.

Sources

  1. cloudian.com — Company mission, product portfolio (HyperStore, HyperScale AIDP, HyperIQ, HyperCare), customer count (900+), and market positioning
  2. cloudian.com — Global presence, headquarters location, and company overview
  3. www.crunchbase.com — Funding history, total raised ($360.1M), latest funding rounds (Series, growth equity), and investor identities (Digital Alpha, Eight Roads, Intel Capital)
  4. cloudian.com — Named customers (Pfizer, Verizon, DoD, PostFinance, RaboBank, Johns Hopkins, National Cancer Institute, State of California, and 20+ service providers)
  5. news.mit.edu — Founders (Michael Tso MIT '93, Hiroshi Ohta), vector database integration (July 2025), and strategic AI positioning
  6. siliconangle.com — HyperScale AI Data Platform general availability (November 2025), NVIDIA AI reference design basis, vector performance claims (8x faster), and CEO Michael Tso quote
  7. www.storagenewsletter.com — CTERA partnership announcement (November 2025), mutual reseller agreement, and integrated data fabric capabilities
  8. cloudian.com — AWS Outposts second-gen announcement (April 2025), 500TB to exabyte scalability, pricing ($0.03/GB), and HyperCare inclusion