Startup Profile: Attimis Unifies Data for AI
Enterprise AI feeds on corporate data—clean, clear, well-managed data. Whether the task is training or inference, the need is the same. Yet many enterprises face a major challenge in supplying good data to AI pipelines: namely, unstructured, silo’d data is often stranded in geographically dispersed locations for a range of applications.
Enter Attimis, a five-year-old startup based in California that provides a global namespace across heterogeneous storage platforms, unifying unstructured data across multiple silos for global AI and analytics workloads.
How Attimis Is Different
How is this different from other data orchestration platforms on the market? According to CEO and cofounder Damian Kowalewski (ex-HPE, Meta, RStor), the differentiator is in the architecture of Attimis’s product, OneBucket.
“We are transparent and scalable,” he said. “We don’t build a database table to provide a single endpoint for data. We are stateless.” This approach eliminates bottlenecks, improves performance, and enables dramatic scalability, he claims: “Some of our customers have hundreds of petabytes of data.”
Early Partnerships and Integrations
Attimis was founded in 2021 but emerged from stealth early this year. And it’s been charging hard ever since. Its OneBucket product is available on a per-Pbyte per year subscription or as an annually charged developer test bundle for $10,000. It’s scored Seagate and TCS as global partners. Through a deal with Dell, OneBucket supports Dell ObjectScale, PowerScale, and OneFS platforms.
There are also integrations with Databricks and Snowflake, whereby customers can point analytics associated with those data management platforms at data in OneBucket.
Why not simply run with data in those platforms? Kowalewski says it’s a matter of waiting. “Data ingestion [into a data lake] can take too much time,” he said. “Customers can simply point platform analytics to OneBucket and let the data stay where it was.”
Attimis has also struck a partnership with Aquila Clouds, a FinOps platform that delivers observability into multicloud and hybrid cloud environments to reveal where costs are allocated and how workloads are (or aren’t) optimized to fit budgets. The information gleaned by Aquila Clouds is used to leverage OneBucket’s data orchestration and data deduplication capabilities.
"Visibility without action is just an expensive bill," stated CEO Kowalewski in a blog post by cofounder, Sales and GTM, Shayne Stubbs. "By partnering with Aquila Clouds, we are giving customers the 'Brain' to see the waste and the 'Muscle' to fix it."
A Moving Target
Attimis has entered the data management market at a crucial juncture, when enterprises are engaged in AI implementations, particularly using inferencing, or the technique of tailoring AI models to work with data specific to an organization.
According to Futuriom’s Enterprise AI Index (subscription required), enterprises in the vertical markets of retail, financial services and insurance, healthcare, and manufacturing are adopting AI applications at a brisk pace. Companies in these areas need help getting their data into the necessary workflows to create their AI apps.
Attimis is intent on addressing this enterprise market, but it’s also focused on delivering data for analytics and AI to neoclouds and alternative cloud providers, which also are in line for partnerships with the startup.
Challenges in the Market
Despite its innovative approach, Attimis faces its own set of challenges. Competition looms from the likes of CTERA, Hammerspace, Komprise, and others that also seek to converge corporate data from multiple sources for use in AI and are working hard to address market needs with their own innovations.
So far, Attimis supports S3-compatible data, though it’s working on links to other hyperscaler data APIs, and Kowalewski says an announcement is imminent. Given Attimis’s go-to-market pace so far, it seems we can expect that and other news shortly.
Startup Profile: Attimis
Headquarters location: Sunnyvale, Calif.
Founded: 2021
Founders: Damian Kowalewski, CEO; Domenico Guardini (ex-RStor, Meta), CTO; George Yakovlev (ex-RStor, Meta, Google), Chief Scientist; Shayne Stubbs (Sand Hill Consulting, ex-Aviatrix), Sales and GTM
Target market: Enterprises and neoclouds intent on unifying unstructured data across multiple silos for global AI and analytics workloads.
Prominent investors: Ascolta Ventures, Raptor Group, Tanzin Capital
