Overview
Primary Data developed software for managing data across different storage systems. David Flynn and Rick White founded it in 2013 after building Fusion-io. Its DataSphere platform aimed to let applications keep using their data while policies placed that data on suitable storage.[1]
TechCrunch reported a January 2018 shutdown after about $100 million in equity and debt.[6] The technology had a later life in Hammerspace, a separate venture. The useful distinction is between a worthwhile infrastructure idea and a product that enough customers will adopt before its funding runs out.[7]
Image 1 / 1
Founding Story
The team's earlier work exposed the problem. Flash storage introduced a fast new tier, while cloud adoption added another location for data. Existing enterprise systems remained in place. Smith's 2017 interview describes the resulting mix as the reason for managing data independently of the device holding it.[1]
The November 2013 company announcement named Flynn and White and reported a $50 million financing led by Accel. It planned a product introduction the following year. A history of building a successful storage company helped attract resources before this new product had demonstrated broad adoption.[2]
Smith joined as chief executive in late 2014 and said the initial all-flash concept broadened into software spanning different storage types. He was not an original co-founder.[1]
Timeline
- 2013: Flynn and White founded Primary Data; the company announced its $50 million initial financing in November.[1][2]
- Late 2014: Smith joined as chief executive.[1]
- August 2016: DataSphere became generally available on an annual subscription basis.[3]
- March 2017: DataSphere 1.2 expanded supported systems, scale, and cloud destinations.[4]
- August 2017: DataSphere 2.0 arrived alongside $20 million of investment and a $20 million credit line.[5]
- January 2018: The original company shut down.[8]
- October 2018: Hammerspace launched a new product using technology and staff from Primary Data.[7]
- April 2025: Hammerspace announced $100 million of new growth capital.[9]
What They Built
DataSphere separated information about files from the storage holding their contents. That information, called metadata, included location and usage. Applications could use a logical view of data while the software managed physical placement. This addressed the work of moving between fast, expensive storage and other capacity without changing each application.[4]
The control path supplied the location and policy decisions; the data path carried the actual file contents. Keeping those paths separate aimed to avoid putting every read and write through a central bottleneck. The 2017 examination describes support for Linux, Windows, and Mac hosts, multiple storage vendors, and more than one S3-compatible cloud destination.[4]
The August 2017 release added more detailed policy expressions, cloud activity statistics, and interface changes. Those features show a product attempting to handle real operational complexity. Their presence does not independently establish reliability, lower total cost, or an easy deployment at every customer.[5]
Market Position
Target Customers
Large enterprises with substantial existing storage were the intended buyers. Smith described about a dozen customers in early 2017, without naming them because of confidentiality agreements. That is a dated company account, rather than a verified record of paying production deployments or later renewals.[1]
The technical publication reported an initial focus on test and development, with plans to reach more critical production workloads. Moving into that next category required a stronger demonstration of operational safety and predictable support.[4]
Market Size
The reviewed sources do not establish a reliable addressable market for the precise product. Large amounts of stored data do not necessarily create a purchase. A buyer must face enough migration, utilization, or administration cost to justify another system and its deployment work.
Competition
The most practical alternative was often the existing process: buy more capacity, use each storage vendor's tools, or migrate manually. A new layer needed to offer enough benefit to overcome the cost of adopting it. Technical flexibility could also widen the range of environments that the vendor had to support.
Business Model
DataSphere sold as subscription software through direct sales and partners. The 2016 launch report describes annual availability, while management's later account describes platform and attached-service licensing. No independently verified price, retention rate, or sustainable sales margin was established in this review.[3][1]
Funding descriptions need their component parts. The 2017 announcement called its package $40 million, but identified $20 million of investment and a $20 million line of credit. A credit facility is a borrowing arrangement, and the announcement does not establish that all of it was drawn. It cannot be described as another $40 million of equity.[5]
The original 2013 round and subsequent package financed development and sales. Neither establishes product revenue or customer savings. The public record reviewed here does not provide a complete audited account of expenditure or final stakeholder losses.
Outcome and evolution
Later reporting confirms the January 2018 shutdown. In his October 2018 Hammerspace interview, Flynn said, “The Primary Data technology worked fine.” He added, “They were interested in moving to hybrid cloud.” He attributed the earlier outcome to its customer and single-data-center focus. This remains his interpretation, without an independent controlled comparison here.[8][7]
The same interview describes Hammerspace as using the earlier technology with a smaller team and a revised cloud-oriented product. It also identifies an adoption difference: the earlier system required an application outage to remount storage, while the new approach aimed to add metadata without that interruption. Such details make deployment friction a concrete part of the story.[7]
Hammerspace's current site offers policy-driven placement across existing storage, sites, and clouds. Its 2025 release announces new investment. These sources establish an active later company promoting a related approach; they do not prove every product claim or show that Primary Data's original owners and creditors were made whole.[9][10]
The outcome therefore leaves room for both product and commercial explanations. A difficult integration layer may need more refinement; buyers may also prefer a narrower or differently timed use case. The evidence does not isolate one causal factor. This review did not obtain a primary legal dissolution record, complete intellectual-property transaction terms, final creditor recoveries, or audited customer economics.
Key Lessons
- Test the cost of adoption alongside technical capability.
- Begin with a specific workload that customers need to improve now.
- Preserve the distinction between investment, debt capacity, revenue, and savings.
- Follow technology reuse separately from the original company's outcome.
Sources
- CEO Lance Smith origin/product interview
- Original2013financing announcement
- 2016product launch account
- DataSphere1.2technical examination
- 2017financing and2.0release
- Contemporaryshutdownreport
- Hammerspacefounder sequel account
- Later shutdown corroboration
- Hammerspace2025financing
- Hammerspacecurrent orchestration