Federal agencies may benefit from slower cloud adoption, Cloudera CEO says

“The cloud providers have really powerful technology stacks, but they are unique," which is why Cloudera sees an opportunity for the company's hybrid cloud solutions. Cloudera
Charles Sansbury believes a "cloud hangover" has led commercial enterprises to reconsider data workload strategies, giving the government an opportunity to gather lessons learned and be "fast followers."
The cloud computing market is changing with more organizations, including federal agencies, shifting away from migrating everything to the cloud to a more selective and hybrid approach.
The market is suffering from a bit of a “cloud hangover,” according to Charles Sansbury, CEO of Cloudera.
“We moved a bunch of things to the cloud, and we didn’t get the savings we expected,” he told Washington Technology, relaying what customers have been telling his company. “The shift to the cloud also didn’t really help in terms of quality results.”
Cloud momentum is still real, but appetites to move everything to the cloud are no longer there. Sansbury predicted organizations may move two-thirds of their workloads to the cloud and the remaining third will stay in-house, but still be massive.
Sansbury was in Washington, D.C. as part of Cloudera’s worldwide Evolve 25 tour for showcasing its technology, sharing insights on where the market is going and discussing challenges its customers face.
Stay in the know — Washington Technology’s Insider Membership gives you unmatched access to breaking news, in-depth analysis, and insights that federal contractors can’t afford to miss. Join today for 50% off.
During the conversation with Washington Technology, Sansbury drew a distinction between multi-cloud and hybrid cloud, with an emphasis that the latter is where Cloudera sees the future.
“A lot of organizations claim they are multi-cloud, but what they really doing is using two different cloud providers for two different things,” he said.
The challenge is portability and the difficulty of moving data and functions from one cloud provider to another.
The recent Amazon Web Services outage paralyzed many organizations and is just the latest example of the risk cloud-dependent organizations face today.
“The cloud providers have really powerful technology stacks, but they are unique. They didn’t architect their native functionality to run equally across AWS, or Azure or Google Cloud,” Sansbury said. “Multi-cloud is not workload portability.”
Cloudera began developing tools to address portability several years ago. The company has also invested $1 billion in research-and-development spend over the last three-to-four years to build capabilities across public clouds, private clouds and on-premise data environments.
The goal is for customers to move data seamlessly from one cloud to another and between the cloud and in-house data centers. That is a true hybrid cloud, Sansbury said.
“Hybrid is a combination of workloads running on premises, on the private cloud or public cloud,” he said. “What we can do is allow you to run workloads across any of those platforms.”
Cloudera’s acquisition of Taikun in August accelerates the company’s containerization strategy. The integration is still underway, but Sansbury said new products are just months away.
The shift to hybrid cloud is being accelerated by the adoption of artificial intelligence. Organizations want to leverage AI but don’t want to move proprietary data outside their security perimeter.
“AI grew up on the cloud, so the idea has been to take these large language models and training them on all the publicly available data,” he said.
But moving data to the cloud is risky, whether you are a defense agency or financial services company.
“They don’t want to do that and moving all the data to feed the models isn’t the right mindset,” Sansbury said.
Cloudera promotes the idea of a data lake house that pulls together data sources.
“Basically, it allows you to bring the model to where it’s convenient for you to have the data and not to have to move the data to a cloud-based structure in order to feed and train the models,” he said.
That scenario allows for what is being called private AI.
“It’s basically AI done on your proprietary data that you don’t allow to escape your security perimeter,” he said. “People find that very compelling.”
The public sector might have an advantage here because federal agencies have been slow to adopt the cloud-first strategies.
“In this case, it’s OK to be a fast follower as opposed to a pioneer,” Sansbury said.
Cloud service providers provide a lot of value in terms of ease of implementation and ease of upgrades, which is why the momentum continues. But there is a growing realization that the cloud is good for some workloads, not all.
“Customers now understand the different dynamics,” he said. “Now, they are thinking about matching the appropriate workload to the appropriate computing platform.”
He added customers have an opportunity to learn from the early commercial adopters when it comes to cost, downtime and complexity, he said.
Two years ago, a customer told the company they were moving everything to the cloud and removed Cloudera from their strategic vendor list.
But now that client is back because it is building a data center.
“We’re not moving everything to the cloud,” Sansbury said they told him. Cloudera is a strategic vendor again “because it turns out what you do is critical to how we use and manage our data to run our business," according to Sansbury.
A 2024 survey of chief information officers by the Barclays investment bank found that over 80% were bringing some data workloads back in-house.
“We think we’ve made some good decisions in terms of the direction of the company, and we’ve gotten a little bit lucky,” he said.