Clustered scale-out file system provider Qumulo has replaced Bill Richter as CEO after eight years with Doug Gourlay, a former Arista vice president and general manager of software, as it aims to grow in the applied AI market with its hybrid cloud filer software that can scale anywhere.
Richter became CEO in November 2016. Investors pumped in $32.5 million in a C round, $30 million in a C prime round the following year, $93 million in a D round in 2018, and $125 million in an E round in 2020. He used these capital infusions to develop the business, software, go-to-market organization, and channels. Qumulo currently has over 1,000 customers. The company uses full-featured, high-performance file system software that runs on-premise and in cloud-native form on Azure and AWS clouds. In fact, in the Azure cloud it outperforms Weka, known for its fast parallel file system performance. Richter currently serves on Qumulo's board of directors and as an executive advisor.
“We are thrilled that Doug will be leading Qumulo's expansion into its next phase of growth, building on the company's recent success. His impressive resume, vision and passion for client success make him the ideal fit to lead Qumulo,” Sameer Menon, a director at BlackRock, an investor in Qumulo, said in a statement.
Matt McIlwain, managing director at Madrona Venture Group and a member of Qumulo's board of directors since 2014, added: “Doug helped grow Arista Networks from an early-stage startup into a company with a market cap of over $100 billion, making him the ideal person to lead Qumulo's next phase of expansion.”
“I am thrilled to join Qumulo. The combination of great team, innovative on-premise and cloud software, applied AI opportunities, and the Scale Anywhere vision attracted me to the company. I look forward to executing on our unique vision and strategy for our customers, partners and channels. To quote America's first astronaut just before launch: Let's light this candle.”
We asked executives about this leadership change.
Blocks and Files: Why the change from Bill Richter to Douglas Gouray?
CTO Kiran Bageshpur: Leadership transitions are natural and often inevitable after more than eight years as CEO. This transition, while bittersweet, is a testament to our company's commitment to growth and evolution. We've found someone who can build on the foundation Bill has built: a great team, a strong culture, a high-quality product base, highly satisfied customers, and a broad, differentiated product offering targeted at mainstream enterprise customers.
B&F: What did Doug bring to Kumullo that Bill couldn't?
Bageshpur: Bill's core competencies are a solid financial background and incredible operational savvy. Doug's background is visionary product strategy, strong product/market leadership, and scaling businesses to consistently achieve over $1B in revenue. When we interviewed Doug, it became clear that his extensive background in networks and systems, combined with his close attention to the engineering, operational, and business priorities of a mainstream enterprise customer base, is incredibly synergistic with Qumulo's next phase of growth.
Most notably, we are seeing increased interest from many of our clients in reintegrating compute, network and storage, and having leaders who see this as a broader system deployment rather than a point product development is a key driver for Qumulo.
B&F: Will Qumulo, under Doug's leadership, move quickly to make customer information held in storage available for Search Augmented Generation (RAG) LLM?
Goulet: This was one of the first questions I asked Kiran and the executive team when discussing the evolution of Qumulo. Our industry architecture agnostic “run anywhere” model, combined with our cloud native implementation that levels consumption economics across public clouds, and our modern global namespace implementation, enables “data anywhere, everywhere, at once” and the ability to apply that data to solve complex business problems, which is especially important for mainstream enterprises.
Frankly, the market for hyperscale AI training/learning and model development does exist, but the talent pool to actually develop quality models is small, and a few well-funded organizations have made significant infrastructure investments in the space, leaving it in the hands of nation states, hyperscale technology companies, and specialized traders.
Cloud providers are developing capabilities for mainstream enterprise consumption primarily using “off-the-shelf” models such as OpenAI. Enabling enterprises to securely interact with well-formed, mature public models without making their data part of the public model will revolutionize the consumption of AI, moving towards a more “applied AI” model. RAG is a key technology we are actively integrating to unlock the potential of the exabytes of data being managed and stored across volumes containing billions of unstructured and semi-structured data sets.
B&F: Do you see this as a pressing need, one that exists both on-premise and in the public cloud, perhaps with different implementations?
Goulet: Qumulo's strategy and the execution of its products today standardize the implementation and management of clients' data whether it resides in an on-premise datacenter, a remote edge location, or the public cloud. More importantly, it standardizes the economic catalyst – unit economics – so that businesses have the freedom to choose whether to use the cloud, their existing datacenter, or a hosting facility, without a technology foundation constraining or predetermining a particular outcome.
On-premise and cloud-based consumption, the ability to offer consistent unit economics, the elimination of legacy barriers, and Hobson's “cloud vs on-premise” choice are unique differentiators for Qumulo that resonate with infrastructure heads to CIOs.
B&FWill Qumulo develop its own LLM to help customers use their Qumulo data assets?
Goulet: Never say never, but this is not currently in our action plans. We put our customers first, building things that are “relevant” and of high quality for our customers, and paying close attention to market, secular and technology trends.
Rather than trying to develop our own competing LLM in a crowded space where we are generally building superior technology, we are far more likely to use the same RAG-assisted model that enables us to leverage the huge investments in LLM development from OpenAI, Microsoft, Google, and others.
B&FDoes Qumulo on Azure integrate with OpenAI Azure capabilities?
Brandon Whitelaw, VP of Cloud and Strategic Partnerships: We cannot discuss the nature of our strategic partnership with Microsoft and Azure, particularly regarding future developments that have not been announced. We have a strong and deep partnership with Microsoft, are the only storage partner in the Azure Native ISV program, and are actively collaborating on a variety of fronts.
B&FWill Qumulo move quickly to become cloud-native on AWS and take advantage of AWS AI facilities such as Bedrock?
White Law: As you may know, we have already launched Cloud Native Qumulo on AWS, and we are actively working with our customers and AWS on various integrations.
