TL;DR — Key Takeaways
- Cloudera expands its AI toolkit: New blueprints combine NVIDIA AI workflows with Cloudera’s data management, security and governance capabilities to simplify production deployments.
- Anywhere Cloud targets hybrid environments: Cloudera plans to deliver the platform by the end of November, providing centralized management of data and AI workloads across IT environments.
- Costs are forcing AI strategy changes: A Cloudera survey finds 89% of respondents have limited or redesigned AI use cases because of costs associated with rising token consumption.
Cloudera today added a set of blueprints for deploying artificial intelligence (AI) workflows on its data lake platform that were developed in collaboration with NVIDIA.
Announced at a Cloudera EVOLVE26 event, the company also reiterated a commitment to deliver a previously announced Cloudera Anywhere Cloud platform by the end of November. Additionally, Cloudera is committing to supporting that platform for the next six years, with updates now being provided every two weeks to ensure any vulnerabilities that are discovered are quickly remediated.
The NVIDIA AI Blueprints were built using the NVIDIA Agent Toolkit and are an extension of the Accelerators for ML Projects (AMPs) that Cloudera provides. The NVIDIA Agent Toolkit leverages NVIDIA NIM microservices and NVIDIA Nemotron models to make it simpler to invoke inference engines in the cloud or in an on-premises IT environment.
Cloudera CTO Sergio Gago told conference attendees the overall goal is to make it simpler for IT teams to separate the management of compute, storage and infrastructure at a higher level of abstraction. At the core of the Cloudera Anywhere Cloud platform is an orchestration framework that provides IT teams with a control plane for managing and governing data wherever it resides at that higher level of abstraction. That framework is based on a platform that Cloudera gained with the acquisition of Octopai in 2024.
The arrival of the Cloudera Anywhere Cloud comes at a time when many organizations are rethinking their approach to AI. A Cloudera survey, for example, finds 89% of surveyed organizations have limited or redesigned use cases involving AI because of cost issues stemming from rising token consumption.
The survey also finds that every respondent reports their organization has experienced higher or more volatile AI costs, with 99% moving to reduce them by, for example, moving a workload to another IT environment (45%) or reducing processing frequency, model size, context size, or output volume (40%). Just under a quarter (24%) have paused or canceled an AI use case altogether.
Additionally, nearly all respondents (98%) said their organization has pulled back on AI deployments due to failure, hallucinations, or unexpected behavior, with 68% of respondents needing to pull back workloads sometimes or frequently.
It’s not clear at what rate organizations are deploying AI workloads in production environments, but many of them are now in the process of focusing their efforts on a narrower range of initiatives that have proven value. The overall goal is to reduce the total cost of deploying AI workloads while at the same time re-engineering workflows in a way that makes the organization more competitive. Applying what are largely probabilistic technologies to workflows that are often deterministic is, of course, proving more challenging than many initially anticipated.
One way or another, it’s now more a question of how long it will take organizations to successfully deploy AI workloads at a cost they can afford, rather than whether they will. The issue is whether that goal will be achieved by most organizations in 2027 or, quite possibly, the year after.
