This distributed data storage startup wants to take on the Big Cloud

The explosion of artificial intelligence firms has pushed the demand for computing power to recent heights, and firms resembling CoreWeave, Together AI and Lambda Labs have capitalized on this demand, attracting huge amounts of attention and capital due to their ability to offer distributed computing power.

However, most firms still store data with the three largest cloud providers, AWS, Google Cloud and Microsoft Azure, whose storage systems are built to keep data close to their very own compute resources reasonably than spread across multiple clouds or regions.

“Modern AI workloads and AI infrastructure are choosing distributed computing over the large cloud,” Ovais Tariq, co-founder and CEO of Tigris Data, told TechCrunch. “We want to provide the same option for storage because without it, compute power is nothing.”

Tigris, founded by the team that developed Uber’s storage platform, is building a network of localized data storage centers that it says can meet the distributed computing needs of contemporary AI workloads. The startup’s AI-powered storage platform “moves with your calculations, [allows] data [to] it mechanically replicates where GPUs reside, supports billions of small files, and provides low-latency access for training, inference, and agent workloads,” said Tariq.

To do all this, Tigris recently raised a $25 million A round led by Spark Capital and with participation from existing investors that include Andreessen Horowitz, TechCrunch has learned exclusively. The startup is acting against the operators present on the market, which Tariq calls the “Big Cloud”.

Tariq believes that these incumbent operators not only offer costlier data storage services, but also less efficient ones. AWS, Google Cloud, and Microsoft Azure have historically charged an exit fee (called a “cloud tax” in the industry) if a customer wants to migrate to one other cloud provider or download and move their data if they need to, say, use a cheaper GPU or train models in different parts of the world at once. Think of it like having to pay extra to go to the gym if you wish to stop going there.

According to Batuhan Taskayi, head of engineering at Fal.ai, one of Tigris’ customers, these costs used to make up the majority of Fal’s cloud spend.

In addition to exit fees, Tariq says there is still a latency issue with larger cloud providers. “Exit fees were just one symptom of a deeper problem: centralized storage that can’t keep up with the decentralized, high-speed AI ecosystem,” he said.

Most of Tigris’ 4,000-plus customers are similar to Fal.ai: generative AI startups building image, video and voice models that typically have latency-sensitive large data sets.

“Imagine talking to an AI agent that handles local audio,” Tariq said. “You want the lowest latency. You want your compute to be local, close together, and you want your storage to be local as well.”

He added that giant clouds are not optimized for AI workloads. Streaming huge datasets for real-time training or inference across multiple regions can create latency bottlenecks and decelerate model performance. However, access to localized storage means faster data recovery, which implies developers can run AI workloads reliably and cheaper using decentralized clouds.

“Tigris allows us to scale our workloads across any cloud by providing access to the same data file system from all of these places without requiring exit fees,” said Fal’s Taskaya.

There are other the reason why firms want to keep data closer to their distributed cloud solutions. For example, in highly regulated fields resembling finance and healthcare, one of the essential obstacles to implementing AI tools is the need for firms to ensure data security.

Another motivation, Tariq says, is that firms increasingly want to own their data, as demonstrated by how Salesforce earlier this 12 months blocked its AI rivals from using Slack data. “Companies are becoming more and more aware of how important data is, how it drives LLM and artificial intelligence,” Tariq said. “They want to have more control. They don’t want anyone else to have control over it.”

With the recent funding, Tigris intends to proceed building out its data storage centers to meet growing demand – Tariq says the startup has seen eightfold growth every 12 months since its founding in November 2021. Tigris already has three data centers in Virginia, Chicago and San Jose and wants to proceed expanding in the US, in addition to in Europe and Asia, particularly in London, Frankfurt and Singapore.