The News: The NVIDIA acquisition of SwiftStack as part of its strategy to enable the scaling and rapid performance of data analytics, artificial intelligence (AI) and high-speed performance computing (HPC) for customers. This acquisition is the result of an ongoing relationship—NVIDIA has been a SwiftStack customer for the past year, is a preferred storage partner of NVIDIA, and has significant expertise building supercomputers, products. and solutions supporting AI, HPC, and accelerated workloads. SwiftStack customers include Verizon, PayPal, and Rogers Snapfish. The transaction, for which no financial terms were disclosed, is expected to close in coming weeks. Read more at Nasdaq.
NVIDIA Acquisition of SwiftStack Facilitates Cloud-to-Edge Data Management for AI and HPC
Analyst Take: For NVIDIA, this acquisition will remove a blind spot in its overall portfolio and will include storage orchestration. It clears the way for the company to implement robust AI platforms completely with a variety of storage platforms.
The deal provides NVIDIA with a software-driven data storage and management platform for acceleration of deep learning, analytics, and high-performance computing applications across multiclouds. As mentioned above, SwiftStack is already a preferred storage partner of NVIDIA, moreover, SwiftStack’s data management technology is also already integral to NVIDIA’s GPU-powered AI infrastructure designs.
The SwiftStack data platform—currently in version 7—enables users to store, manage, control, and use petabytes of unstructured data from cloud to edge. It is available either as a SaaS application or on-premise software installation. It supports public cloud (AWS, Azure, GCP), on-premises and edge deployments.
SwiftStack was one of the largest contributors to OpenStack’s Swift object storage platform, which is still integral to its cloud storage architecture. At the core of SwiftStack’s solution portfolio is a cloud file system that enables applications to access and consume storage using either SMB or NFS interfaces. A controller manages all storage resources and data placement policies. A single storage namespace enables applications and workflows to leverage SwiftStack services across public and private cloud infrastructure.
Key verticals that use SwiftStack include autonomous vehicles, media and entertainment, global service providers, life sciences, and web-based businesses.
SwiftStack states that it will continue to maintain, enhance, and support an existing set of open source tools, including ProxyFS, 1space and Controller–after the acquisition closes.
SwiftStack is a Complementary Offering to NVIDIA’s Core AI Hardware/Software Portfolio
Once the NVIDIA acquisition of SwiftStack is completed, NVIDIA will be able to go beyond selling GPUs and AI software to pursue more adjacent deals for enterprise multicloud data management. Going forward, NVIDIA and other vendors of AI platforms, tools, and libraries will need to diversify into robust cloud data management in order to address the one-stop shopping needs of enterprise customers large and small.
NVIDIA’s Focus Here is on The Stack
NVIDIA’s market-leading GPU-based offerings appear poised for growth and adoption in 2020 and beyond, but more competition in its core chip market is inevitable. That’s not necessarily a bad thing—certainly not for consumers. But it’s clear that NVIDIA is thinking beyond chips, which is what you’d expect from a market leader. It’s also clear the company’s strategic focus is on the stack. It’s about the data center, it’s about enterprise computing, and about the entire stack utilized by the enterprise—and it’s about the future of AI for enterprises.
Fill-in Acquisition May Rankle Some of NVIDIA’s Many Cloud Data Management Partners
As noted above, it is possible to regard NVIDIA’s fill-in acquisition of SwiftStack as a tactic that is both an offensive move and a defensive move. An offensive move in that the company is expanding the addressable market for its GPU-based AI hardware/software solution portfolio. And also a defensive move — bracing its revenue mix against the commoditization of its core GPU-based hardware and AI pipeline workbench offerings.
This is also a calculated risk on NVIDIA’s part, in that acquiring SwiftStack may alienate the many cloud data management solution providers in the vast channel ecosystem through which it goes to market. At the very least, the large public cloud providers (AWS, Microsoft Azure, Google Cloud Platform, etc.) who have built substantial AI-as-a-service businesses on the firm’s GPUs may have misgivings about the potential for NVIDIA’s SwiftStack offerings to compete with their own multicloud data management offerings.
NVIDIA is highly respectful of its partner ecosystem, so it’s unlikely that the company will use these newly acquired offerings to compete directly with partners’ solutions for cloud-to-edge data management. Likewise, it’s unlikely that the company will seek to acquire or build software to address adjacent data management requirements—such as discovery, profiling, cleansing, and transformation—which would dilute its core focus on the AI data pipeline and cross-over into the larger enterprise data management space where NVIDIA would scarcely be able to compete with the likes of Informatica.
Instead, NVIDIA will probably leverage SwiftStack’s tech to accelerate the performance of its existing NGC software for training of deep learning and machine learning models on GPU clusters running on hybrid clouds, multiclouds, and more complex cloud-to-edge distributed environments.
NVIDIA’s Acquisition of SwiftStack Gives it a Credible Platform for the Cloud-to-Edge AI Data Pipeline
NVIDIA’s acquisition of SwiftStack provides it with a much-needed high-performance platform for the core data management challenges involves in modeling and training AI and HPC applications for cloud-to-edge deployment.
It is apparent that NVIDIA recognizes its current market-leading status in AI chipsets won’t necessary last forever. By focusing on the enterprise stack—the data center, enterprise computing, and all things related to the future of AI in the enterprise, the company plots a path forward that allows it to sustain its impressive growth record.
That said, taking on SwiftStack’s software-defined multicloud storage offerings is also a calculated risk, in that acquiring SwiftStack may alienate the many cloud data management solution providers in the vast channel ecosystem through which it goes to market. But that possibility is remote, due to the strong likelihood that NVIDIA will use SwiftStack’s tech to accelerate the performance of its existing solutions for training of deep learning and machine learning models on cloud-to-edge GPU clusters.
No one seriously expects that NVIDIA will use this acquisition as a springboard to seek more acquisition—beyond its core AI DevOps pipeline focus—in the enterprise data management arena.
Futurum Research provides industry research and analysis. These columns are for educational purposes only and should not be considered in any way investment advice.
Other insights from Futurum Research:
Image Credit: SwiftStack
The original version of this article was first published on Futurum Research.
James has held analyst and consulting positions at SiliconANGLE/Wikibon, Forrester Research, Current Analysis and the Burton Group. He is an industry veteran, having held marketing and product management positions at IBM, Exostar, and LCC. He is a widely published business technology author, has published several books on enterprise technology, and contributes regularly to InformationWeek, InfoWorld, Datanami, Dataversity, and other publications.