With the growing momentum in AI investments, a significant number of companies find themselves grappling with the challenge of operationalizing their AI-driven initiatives and achieving meaningful returns on investment (ROI).
The hurdles encountered are diverse, yet a prevalent issue revolves around data management. The data essential for training, running, and fine-tuning AI models is often fragmented, segmented, and poorly optimized. According to a 2022 survey conducted by Great Expectations, an open-source data benchmarking platform, 77% of organizations expressed concerns regarding their data quality.
Startups offering solutions to these data issues are attracting substantial investment.
On Wednesday, Weka, a platform specializing in building data pipelines capable of managing various data sources, types, and sizes, disclosed its successful $140 million Series E funding round, split into two parts ($100 million and $40 million). This round was spearheaded by Valor Equity Partners, with contributions from Nvidia, Norwest Venture Partners, Micron Ventures, Qualcomm Ventures, Hitachi Ventures, among others. This oversubscribed funding round propelled Weka’s post-money valuation to $1.6 billion, doubling its previous valuation.
CEO Liran Zvibel, along with Weka co-founders Maor Ben-Dayan and Omri Palmon, initially collaborated on the data storage startup XIV, which IBM acquired for $350 million in 2007. The trio remained with IBM for several years before pursuing independent ventures.
Zvibel, discontent with the persistent data management challenges, felt compelled to address the issue.
“I was frustrated with customers having to rely on disparate, siloed data solutions that were wasteful, costly, and complex,” Zvibel remarked. “The rise of cloud computing, high-performance computing, machine learning, and the early stages of AI workloads made this problem even more evident.”
Thus, in 2013, Zvibel, alongside Ben-Dayan and Palmon, set out to create an innovative suite of data tools aimed at revolutionizing data storage, management, and movement.
“We envisioned a platform robust enough to handle the performance demands of next-gen compute hardware and large-scale, data-intensive tasks in dynamic and distributed environments,” he explained. “To cater to modern workloads, it needed the capability to process tens of terabytes of data and be deployable anywhere.”
Weka’s primary offering is a parallel file system, a distributed file system designed to allocate and manage data tasks (such as file copying) across multiple locations (servers and workstations) simultaneously. In addition to this, Weka provides services and functionalities tailored for AI and machine learning, visual effects, and high-performance computing tasks in various settings including on-premises data centers, public clouds, and hybrid clouds.
Zvibel asserts that a significant advantage of Weka’s architecture is its ability to accelerate AI model training by minimizing the time needed to transfer data across storage locations. “A typical generative AI data pipeline entails multiple data set transfers, which hinders training time,” he noted. “Weka ensures that model training hardware is continuously supplied with data, enabling faster model training.”
Weka faces competition from data platforms such as DataDirect, Pure Storage, NetApp, and Vast Data. Vast Data, in particular, poses a substantial challenge, having secured a $118 million Series E funding round in December 2023, which tripled its valuation to $9.1 billion.
Despite the competition, Weka maintains a strong position with a client roster exceeding 300 brands, including AI startup Stability AI, 11 of the Fortune 50, and various undisclosed domestic and international government agencies.
Zvibel mentioned that despite Weka’s relatively sizable workforce (approximately 400 employees globally, with plans for a 25% expansion in the coming year), the Silicon Valley-based firm is on track to achieving cash flow positivity by December 2024.
“The decision to raise funds was based on favorable market conditions and proactive interest from investors, allowing us to secure terms highly advantageous for Weka,” he added. “Our average monthly burn rate is projected to be below half a million dollars before reaching that milestone. We’ve surpassed $100 million in annual recurring revenue and are maintaining a path of hyper-growth.”
