Did you miss a session from the Future of Work Summit? Head over to our Future of Work Summit on-demand library to stream.
It’s rare that a startup specializing in data storage gets a unicorn valuation of $1 billion, but MinIO of Palo Alto, Calif., pulled it off when it announced a $103 million Series B funding transaction today.
MinIO earned its newfound corporate status based on its multicloud, AWS S3-compatible object storage suite of apps that the companies says reliesversatility and AI separate it from competitors.
While it was originally designed for S3 use cases, MinIO CEO and cofounder AB Perisamy said it can be used in any cloud environment. In this day and age of patched-together and often distantly separated storage systems, a fast data platform that plugs everything together isn’t common.
“We have more than 11 million active deployments on the public cloud, private cloud, and edge,” Perisamy said. “This includes public cloud deployments on Google Kubernetes Engine, Amazon’s Elastic Kubernetes Service, Azure Kubernetes Service, and private cloud deployments on Red Hat OpenShift, VMware Tanzu, HPE Ezmeral, SUSE Rancher, as well as millions of colo and edge deployments.”
MinIO is a high-performance, Kubernetes-native object store for a broad range of data storage use cases. It can deliver tens of gigabytes of data per node in throughput, Perisamy said. The performance characteristics have made MinIO the object store of choice for machine-learning frameworks, analytics applications, databases, web applications, and other performance-oriented workloads.
Questions for the CEO
VentureBeat asked Perisamy a few questions about the steady development of the seven-year-old company.
VentureBeat: Few storage companies ascend into the unicorn category. How has MinIO been able to move to this level?
Perisamy: The primary reason that MinIO was able to achieve this valuation is that we don’t look like a traditional storage company; we look more like AWS S3 (but not a service, and available anywhere). Consider the following:
- MinIO only does object storage. Almost everyone else is trying to build a block, file, object hybrid solution with predictable results. Our focus on object storage is a massive advantage.
- MinIO is software-defined and we run on any hardware. Think about it: As an open-source company, the hardware configurations we have been exposed to are dizzying. Most of the industry consists of hardware appliances.
- MinIO is cloud-native. We were born into the cloud and have adhered to the principles of containerization and orchestration since our inception. This allows MinIO to run anywhere. Appliances cannot be containerized and orchestrated and so they only run in a limited footprint (on-prem essentially). Running everywhere also provides MinIO with a broader class of workloads to target, which make us primary storage for our clients and community.
- MinIO is the world’s fastest object store. This, too, contributes to our adoption as primary storage because we support databases, advanced analytics, and AI/ML workloads, where other object stores (outside of AWS S3, which is also performant) cannot.
- MinIO is open source. This enables frictionless adoption of our product while the AGPL v3 license enables commercialization, once scale is achieved by the customer. There are more organizations and individuals downloading MinIO in a given month than the traditional storage players add in a year.
- FMinIO is lightweight. The binary that powers the data center at PayPal or the cloud object store in AWS is less than 100MB. That means MinIO can run at the bleeding edge of the network, in a 5G POP or an ARM chip. Edge computing is a massive opportunity, and MinIO enables enterprises to have a consistent data store from edge to cloud. No other object store can power such a wide range of physical locations.
VentureBeat: How is MinIO making use of AI in its solution? How is this setting the company apart from competitors?
Perisamy: There are two components here: how MinIO is integrated with the AI/ML frameworks and how MinIO is using AI in the product itself.
In terms of how MinIO is integrated with the AI/ML frameworks, it helps to note that the AI/ML market is distributed. Data exists everywhere and is generated everywhere. The object store with which these frameworks interact has to be available everywhere. Due to this requirement, AI/ML frameworks are increasingly standardized on MinIO. (See Kubeflow.) While it supports multiple storage offerings, it is standardized on MinIO. There are dozens of other examples, depending on your definition of AI/ML (Spark/Presto for example).
One of the key features that attracts the AI/ML ecosystem to MinIO is the performance angle. On a throughput basis, MinIO can push 2.6Tbps over just 32 nodes of NVMe. This is orders of magnitude faster than other object stores and far more scalable than SAN/NAS solutions.
In terms of how MinIO is using AI in the software itself, we have deployed expert systems logic into our MinIO Health Diagnostics componentry. Part of the SUBNET customer portal, this capability performs a full diagnostics sweep including advanced analytics to make root cause assessments.
Going forward, MinIO will use AI for diagnostics and operational acuity but sees greater utility for the customer by integrating with various AI frameworks.
VentureBeat: How would you summarize all of this?
Perisamy: The result is that MinIO has won with developers, data scientists, and data engineers. They are the engine of value creation in the enterprise, and they consider two primary options in object storage – MinIO and AWS S3. The market recognizes this fact, thus the valuation.
VentureBeat: How is MinIO continuing to innovate?
Perisamy: MinIO continues to extend its feature leadership with the addition of click-to-deploy capabilities on AWS, Google Cloud, and Microsoft Azure. This complements MinIO’s capabilities in information lifecycle management, ransomware protection, and active-active, strictly consistent multi-site replication.
MinIO has seen a high rate of adoption from organizations around the globe. Growth metrics include:
- In the last year, ARR grew by over 201%.
- In the last year, MinIO’s customer count grew by over 208%
- Total Docker pulls of 756,000,000, now averaging over 1 million per day
- More than 31,000 GitHub stars and more than 16,000 members of the MinIO Slack community.
The investment was led by Intel Capital with participation from new investor SoftBank Vision Fund 2, and existing investors Dell Capital, General Catalyst, and Nexus Ventures. The new financing brings MinIO’s total funding to $126 million.