Tigris Data secured a $25 million Series A round led by Spark Capital, with support from existing investor Andreessen Horowitz. Founded by the team that built Uber’s storage platform, the company is developing an AI-native distributed storage network that aligns data location with compute resources. The platform enables automatic replication to where GPUs operate, supports billions of small files, and offers low-latency access for training, inference, and agentic workloads.
Funding and investors
The funding will help expand Tigris’ data centers beyond its current U.S. presence in Virginia, Chicago, and San Jose. CEO Ovais Tariq confirmed plans to launch facilities in London, Frankfurt, and Singapore to serve the growing global demand for distributed compute. Tigris’ user base has grown eightfold yearly since its 2021 founding, now serving over 4,000 customers. Clients like Fal.ai report that egress costs previously accounted for most of their cloud expenses, an issue Tigris aims to eliminate through its localized, multi-cloud storage architecture.
→ Explore more developments signaling industry disruption.
Technical advantage
Tigris’ AI-native approach addresses a core weakness of traditional centralized cloud storage, which is tied to each provider’s infrastructure. Major incumbents such as AWS, Google Cloud, and Microsoft Azure have historically imposed egress fees when customers move data or switch compute environments. Tariq argues that these systems cannot sustain the demands of decentralized AI operations that depend on rapid data movement and minimal latency. Tigris positions its distributed model as both cost-efficient and performance-optimized for generative AI workloads.
Market adoption
Generative AI companies, including those building models for image, video, and voice, make up the majority of Tigris’ customers. These startups depend on proximity-based compute and storage to reduce latency and improve throughput. As Fal.ai’s head of engineering, Batuhan Taskaya, noted, “Tigris lets us scale our workloads in any cloud by providing access to the same data filesystem from all these places without charging egress.” The company’s localized storage model supports these clients’ ambitions for real-time inference and training across distributed environments.
Industry context
As enterprises adopt decentralized AI infrastructure, ownership and control of data have become critical concerns. Tariq cited the example of Salesforce restricting access to Slack data as evidence of companies wanting to secure their datasets. Tigris’ model appeals to this need by allowing customers to keep their data geographically distributed yet securely managed. This strategy directly challenges “Big Cloud” incumbents and reflects a broader migration toward multi-cloud resilience and data autonomy.
Strategic significance
Tigris Data’s Series A funding marks a shift toward a decentralized cloud economy optimized for AI operations. By eliminating egress fees and latency constraints, Tigris reduces costs that traditionally hindered distributed model training and inference. The company’s expansion into Europe and Asia positions it as a competitive alternative for startups scaling globally. If adoption continues at its current pace, Tigris could become a key enabler of geographically distributed AI ecosystems, aligning data mobility with compute availability for next-generation infrastructure.
Reference
Bellan, R. (2025, October 9). This distributed data storage startup wants to take on Big Cloud. TechCrunch. https://techcrunch.com/2025/10/09/this-distributed-data-storage-startup-wants-to-take-on-big-cloud/
