CAEVES offers AI-ready cloud archiving for enterprises, cutting storage costs up to 70% while keeping data searchable as shown on IT Press Tour.
more. The economics of that model are now under pressure from two directions simultaneously. On one side, data volumes are growing faster than budgets. On the other, the rise of AI-assisted workplace tools has exposed a structural flaw: most of what organizations store is invisible to the AI systems they are investing in.
CAEVES Technology, Inc., a privately held startup headquartered in Morristown, New Jersey, presented its Intelligent Deep Storage platform at IT Press Tour #67 in Sofia, Bulgaria on March 31, 2026. The company’s pitch is direct: enterprises are paying premium storage rates for data that is rarely accessed, and that same data is simultaneously absent from the Microsoft 365 search and AI tools that knowledge workers depend on daily.
The Problem in Numbers
The scale of the challenge is not trivial. According to IDC projections cited by CAEVES, global data creation is on course to reach 291 zettabytes by 2027, with between 80 and 90 percent of that volume consisting of unstructured files — documents, engineering drawings, contracts, emails, reports. Of that unstructured mass, IDC estimates that 68 percent is created once and rarely, if ever, accessed again. Gartner, meanwhile, has estimated that roughly 30 percent of enterprise storage spending goes toward maintaining cold or redundant data that generates no active business insight.
The financial consequence is structural rather than incidental. Enterprise data volumes are reported to double approximately every two years, while storage budgets grow at a fraction of that pace. The result is a widening gap between what organizations hold and what they can afford to hold on infrastructure built for active workloads.
The AI dimension compounds the problem. Microsoft 365 Copilot and Azure AI Search index data that is reachable via the Microsoft Graph. When files are moved to cold archive tiers or legacy systems — including tape, aging NAS platforms, or third-party object stores — they fall out of the Graph entirely. From the perspective of Copilot, those files cease to exist. For organizations that have been accumulating decades of operational knowledge in file shares, that represents a direct limitation on the return they can expect from AI tool investments.
The CAEVES Approach
CAEVES was founded in 2024 and 2025 by a leadership team that previously built Talon Storage Solutions, a file caching and edge storage company acquired by NetApp in March 2020. After several years working within NetApp’s cloud data services division, the founders established CAEVES with the stated intention of applying similar principles — intelligent file system management, cloud tiering, and AI-era data accessibility — to the archiving problem specifically.
The product, now generally available in the Microsoft Azure Marketplace, is built natively on Azure infrastructure. It runs as a virtual machine instance within the customer’s own Azure tenant, accepting data via SMB and NFS file protocols and managing lifecycle tiering across Azure’s hot, cool, cold, and archive storage tiers. The company reports that tiering cold data off premium storage and onto Azure object storage can reduce archive costs by up to 70 percent on a per-gigabyte basis.
Crucially, CAEVES stores file content in native Azure Blob format rather than a proprietary file system layer. The company presents this as a differentiator from competitors such as Nasuni, which uses its own UniFS format, or Panzura, which operates on a proprietary CloudFS backend. With CAEVES, customers retain direct access to their data through standard Azure APIs regardless of whether they continue using the platform — a point the company frames as freedom from vendor lock-in.
Architecture and AI Integration
The platform’s architecture consists of a CAEVES VM instance that acts as the file gateway, a metadata and snapshot volume for managing file state, and Azure Object Storage as the primary data repository. Metadata and file indexes are maintained in Azure Table Storage, enabling search queries without requiring full data rehydration from archive tiers.
The AI connectivity is delivered through the CAEVES Copilot Connector, a Microsoft Graph connector that indexes file metadata and content into the Microsoft 365 Search index. This allows files stored in CAEVES — including those in cool or archive tiers — to appear in Microsoft 365 Search results and to be surfaced by Microsoft 365 Copilot in natural language queries. The company claims search response times 50 times faster than traditional tape-based retrieval workflows, with no IT ticket required to access a file.
Data protection is handled through a layered snapshot mechanism. CAEVES takes metadata snapshots at intervals between one and 24 hours, storing them in Azure Blob alongside the corresponding data payloads. The system also supports Azure Blob versioning, a 30-day soft delete policy, and WORM immutability for compliance archiving purposes. The company positions this combination as a replacement for tape-based backup and disaster recovery infrastructure, with restore capabilities that operate in-cloud or can be directed to a hybrid on-premises environment.
Security follows the Microsoft identity model. Authentication is handled through Azure Active Directory and Entra ID, with NTFS access control lists preserved through the tiering process. Item-level permissions in the Microsoft 365 index are maintained, meaning a file that a user could not access on the original file server will not surface in their Copilot results after archiving.
Pricing and Market Positioning
CAEVES operates a capacity-based subscription model sold through the Azure Marketplace, making purchases eligible against Microsoft Azure Consumption Commitments. Pricing is tiered by volume under management: the first five terabytes are free, followed by $0.03 per gigabyte per month for the 6–50 TB range, stepping down to $0.025 for 51–200 TB, $0.02 for 201–1,000 TB, and $0.01 per gigabyte per month for volumes above one petabyte. Contracts are annual but billed monthly based on peak capacity used in each cycle.
At 2,000 TB under management — the example scenario presented at the press tour — the combined CAEVES licensing and Azure infrastructure cost amounts to approximately $51,900 per month, or roughly $0.026 per gigabyte per month total. The company asserts that for organizations managing more than 200 TB of cold data, the tiering savings on premium storage typically offset the CAEVES license cost within the first billing month.
The go-to-market strategy relies primarily on the Azure Marketplace, Microsoft co-sell arrangements with Azure sales teams, and a channel of Microsoft Gold and Solutions Partner system integrators. A multi-tenant MSP mode is indicated on the roadmap but not yet available.
Roadmap and Forward Direction
The product reached general availability in February 2026, following a private preview in August 2025 and a public preview launched at Microsoft Ignite in November 2025. A minor release is planned for September 2026 to coincide with a Microsoft event, with a major version release scheduled for Q1 2027.
Upcoming capabilities include an Enterprise Management Plane — a dashboard providing ROI visibility, queue management, scheduler configuration, and compliance reporting — and an MCP (Model Context Protocol) server. The MCP server is intended to extend the platform’s context layer beyond Microsoft tools to AI frameworks including Azure AI Foundry, OpenAI, Claude, Gemini, and LangChain, enabling permission-aware retrieval across those environments.
The company’s stated longer-term ambition is to evolve from a storage optimizer into a data intelligence platform, adding content classification, data sensitivity scoring, and lineage tracking. It also describes autonomous storage operations as a future direction, where policy recommendations and execution are generated by the platform rather than configured manually.
Context and Competitive Landscape
CAEVES is entering a market that includes established names. Nasuni and Panzura both offer cloud-native file systems with object storage tiering, though at price points the CAEVES competitive analysis places significantly higher — approximately $0.085 per gigabyte per month for Nasuni and $0.07 for Panzura at the higher usage tiers. Neither offers native Microsoft Graph or Copilot integration, according to the CAEVES comparison, which describes Nasuni’s search capability as metadata-only and Panzura’s as an optional add-on.
Larger vendors including NetApp with FabricPool and Dell EMC with its ECS and PowerScale platforms offer archiving and tiering at comparable or lower per-gigabyte rates, but with proprietary namespace layers and limited or no Copilot integration. The hyperscaler platforms themselves — Azure Files, Azure Blob, and Azure NetApp Files — offer lower raw storage costs but require custom development or additional tooling to achieve indexed search and AI connectivity over archived data.
What distinguishes CAEVES in this landscape is not price alone, but the combination of a low archive storage rate, native Microsoft 365 and Copilot integration maintained across all storage tiers, open Azure Blob data format, and deployment entirely within the customer’s own Azure environment. Whether that combination proves sufficient to displace incumbents in enterprise storage decisions will depend on how organizations weigh AI readiness against established vendor relationships and the total integration effort involved in a migration.
The founding team’s prior company — Talon to NetApp in 2020 — provides a reference point for the company’s strategic trajectory, and CAEVES has been explicit that it is building with a similar outcome in mind.
Dr. Jakob Jung is Editor-in-Chief of Security Storage and Channel Germany. He has been working in IT journalism for more than 20 years. His career includes Computer Reseller News, Heise Resale, Informationweek, Techtarget (storage and data center) and ChannelBiz. He also freelances for numerous IT publications, including Computerwoche, Channelpartner, IT-Business, Storage-Insider and ZDnet. His main topics are channel, storage, security, data center, ERP and CRM.
Contact via Mail: jakob.jung@security-storage-und-channel-germany.de