The explosion of AI companies has pressed the demand for computing power for the new ultimate, and companies like KorwV, together with AI and Lambda labs, have created capital on that demand, attracting a lot of attention and capital to the skill of providing distributed counting capabilities.
However, most companies still store data with Big Three Cloud suppliers, AWS, Google Cloud and Microsoft Azur, whose storage systems were made to keep them close to their own counting companies without spreading multiple clouds or regions.
“Modern AI Workloads and AI infrastructure are choosing computing distributed instead of big clouds,” Tigris Data Co-founder and CEO Ovis told Tariq TechCrunch. “We want to supply the same option for storage because the calculation without storage is nothing” ”
Tigris, founded by the team created by the Uber’s storage platform, is creating a network of local data storage centers that claim that modern AI work pressure can meet the calculations that can meet the needs. Startup’s AI-Native Storage Platform “is removed with your calculation, [allows] Data [to] Create automatically where GPUs are, supporting billions of small files and providing short-compliance access for training, estimates and agents, “said Tariq.
To do all this, Tigris recently raised a round of $ 25 million series that was operated by Spark Capital and saw the participation of existing investors, including Andresen Harovits, TechCrunch exclusively. The startup is going against the arrivals, who call Tariq “big cloud”.
Tariq thinks that these arrivals not only provide more expensive data storage services, but also provide a less skilled. Histor Histor has charged address fees in AWS, Google Cloud and Microsoft (dubbing “cloud tax” in the industry) If a customer wants to transfer to another cloud supplier, or if they want to download and remove their data, use a cheap GPU or train model in different parts of the world together. If you want to stop going there, think like paying your gym extra money.
According to Tigris, one of the customers of Tigris, Falle.AE’s engineering head, Batuhan Taskier, these expenditures were once considered most of the cost of fruit clouds.
TechCrunch event
San Francisco
|
October 27-29, 2025
Excluding address fees, Tariq says that there is still a delay problem with the greater cloud suppliers. “The address fees were only one symptom of deep problems: centralized storage that could not keep up with the decentralized, high-speed AI ecosystem,” he said.
Most tigris 4,000+ customers are like Fal.AE: Generator AI Startups Building Images, Video and Voice models, which include larger, delayed sensitive datasets.
“Imagine talking to an AI agent who is doing local audio,” said Tariq. “If you want the minimum delay
He added that the big cloud is not favorable for AI work stress. Large datasets can create latency barriers to run training or real-time estimates across multiple zones, slowly slow the model performance. However, being able to access local storage means that the data is recovered faster, which means that developers can spend more effectively using AI Workloads reliably and decentralized clouds.
“Tigris allows us to scale our work stress in any cloud by providing access to the same data file system from all these places without charging,” the fruit taskaya said.
There are other factors that companies want to get data around their delivery cloud options. For example, in highly controlled cases such as Finanance and Healthcare, a large road block is to ensure data protection for taking AI equipment.
Tariq says another inspiration is that companies want to own their data growingly, how to indicate how to the Salesforce early this year Its AI has blocked rivals From the use of slack data. “The companies are becoming more and more aware of how the companies are data, how important it is to fuel the LLM, how it is giving AI fuel.” “They want to be in control they don’t want someone else to be under control.”
Tigris intends to build its data storage centers to support the growing demand with fresh funds – Tariq has said that the startup has increased by 8x every year since its establishment in November 2021. Tigris already has three data centers in Virginia, Chicago and San Jose, and want to extend to the United States in London, Frankfut, especially in Europe and Asia.