O DiskOver recebe US $ 7,5 milhões de Snowflake, NetApp e outros para facilitar a descoberta de dados não estruturados para a IA

Big-data discovery startup Diskover Data Inc. is currently bursting into the spotlight, announcing a $7.5 million seed funding round, its first acquisition, and key partnerships with industry giants Snowflake Inc. and NetApp Inc.

The startup’s momentum comes as it develops a new platform that allows companies to “structure their unstructured data” and make it more usable and secure for artificial intelligence models.

The round was led by Park Capital Partners and Hive, and also saw participation from Discover’s new partners Snowflake and NetApp. It intends to use the funds to expand its inbound and engineering teams.

DiskOver is on a mission to help businesses untangle the massive volumes of what could potentially be extremely valuable unstructured data — things like documents, PDF files, videos, audio recordings, text messages, scans of handwritten notes, receipts, and so on. These items account for up to 80 percent of all data stored by the typical enterprise, yet most of it does nothing except sit on their storage arrays, collecting dust. That’s because their unstructured nature makes them nearly impossible for businesses to track, govern, secure, and use in any meaningful way.

However, unstructured data is also the most critical fuel for large language models, which have a lot of potential in terms of business automation. That’s why DiskOver wants to help them understand what their unstructured data is, where it is, and how it can be used.

The startup does this by continuously scanning and indexing millions or even billions of unstructured data files across an organization’s IT environment, spanning clouds and on-premises servers. This allows it to generate metadata for those files that explains what they are and where they are stored, making them searchable. With that metadata, it can create a real-time inventory with tools to facilitate discovery, classification, and governance.

According to Diskover, its platform can play a critical role in curating information for AI data pipelines by identifying the most relevant files for training different models. Its technology checks all data it encounters against the company’s existing authentication systems, allowing it to honor existing permissions to ensure compliance.

Co-founder and chief executive Will Hall said his company has transformed what used to be a “murky, opaque data swamp” into a “structured, searchable, actionable resource” that is just begging to be used by AI developers.

“Se você deseja construir a IA que funciona, começa com o que você tem e a curadoria da maneira mais eficiente, e é isso que fazemos”, disse Hall. “Com dados não estruturados, compreendendo mais de 80% de todos os dados corporativos e a necessidade insaciável da IA ​​de entradas de alta qualidade crescendo diariamente, o discover é o ponto de partida para a IA corporativa”.

O DiskOver também se envolve em aquisições, que é extremamente raro para uma startup que acaba de anunciar sua primeira rodada de financiamento. Juntamente com o anúncio de financiamento, ele disse que está comprando uma startup ainda menor chamada Cloudsoda Inc.especializado em “gerenciamento de dados inteligente de Ai-I-Proy” e fornece o que parece ser muitas das mesmas capacidades.

“É um acoplamento ideal”, disse Hall. “Nossos respectivos pontos fortes são mutuamente reforçados. Tínhamos escala, eles tinham simplicidade. Juntos, agora temos a plataforma não estruturada mais intuitiva e pronta para a empresa no mercado”.

O analista da Constellation Research Inc. Michael Ni disse a Siliconangle que o que o Diskover está fazendo não é tão diferente dos provedores de software de gerenciamento de dados existentes como a Komprise Inc., posicionando -se para preencher a lacuna entre os lagos de dados brutos e as camadas de inteligência para conteúdo não estruturado. Mas, diferentemente dos jogadores mais estabelecidos, ele disse que se destaca um pouco devido à natureza de código aberto de sua plataforma.

“Isso oferece ao DiskOver um posicionamento exclusivo, em algum lugar entre as ferramentas de linha de comando de baixo nível e as soluções corporativas comerciais de maior custo”, disse Ni.

Segundo a NI, as principais vantagens que o discover é uma barreira mais baixa à entrada em termos de custo-efetividade, mais transparência e flexibilidade, nenhum bloqueio de fornecedores e uma capacidade comprovada de escalar.

“O DiskOver é construído no Elasticsearch, o que lhe dá credibilidade para a análise do sistema de arquivos em larga escala”, explicou Ni. “Isso o torna mais adequado para equipes de tecnologia que se sentem à vontade para gerenciar infraestrutura, como o Elasticsearch, que desejam evitar as despesas gerais de plataformas caras”.

Embora as reivindicações de grandeza da maioria das startups possam ser tomadas com uma pitada de sal, o fato de que o Snowflake e o NetApp estão em parceria com o Discover e financiando sugere que ele realmente pode estar em algo. O Snowflake disse que está disponibilizando a plataforma do Discover no mercado de Snowflake e também conectará seus recursos de inteligência de dados no local com seu serviço de integração de dados OpenFlow para permitir a orquestração de dados híbridos superiores.

Snowflake Ventures principal Harsha Kapre said more companies are adopting AI-first data strategies, which require being able to access all of their data.

“Enterprises can’t unlock the full value of AI without knowing what unstructured data they have and how to use it,” he explained. “Our partnership with DiskOver, in combination with Snowflake OpenFlow, makes this possible, acting as a super-connector for unstructured data at exabyte scale.”

NetApp is excited about DiskOver and is integrating its services into its own data pipeline infrastructure, which incorporates data sources from the network edge to the cloud. According to Gagan Gulati, senior vice president of data services at NetApp, the partnership will ensure that businesses can better emerge and activate unstructured data, regardless of where it resides. “This collaboration helps accelerate cyber resilience, AI readiness and storage efficiency to deliver outcomes that drive business value,” said Gulati.

Diskover says it had strong momentum even before today’s announcements. It says it already has more than 130 enterprise customers across a variety of industries, including media and entertainment, life sciences, manufacturing, energy and semiconductor design. It also established a business relationship with Dell Technologies Inc. in October 2024.

Neuralytix analyst Ben Woo said he was sold on Discover’s platform because almost every company that can afford to do so these days is investing in AI because of the huge advantage it can offer in terms of enterprise automation. But one of the biggest challenges they face is getting the data they need to fuel their AI initiatives.

“AI requires relevant and accurate data, and DiskOver helps businesses identify the data that will generate the most value,” Woo explained. “It will connect [the data] with the most critical enterprise applications and enable business leaders to make informed decisions to achieve their business goals.”

Image: Siliconangle/Dreamina