Cloudera 分析师：AI 自动化取决于可移植性
AI automation hinges on portability, says Cloudera analyst - SiliconANGLE
Cloudera 在 EVOLVE26 上把自己定位为面向 AI 的“anywhere cloud”，强调企业应能在本地、主权云和不同云之间自由迁移 AI 工作负载。SiliconANGLE 企业编辑 Paul Gillin 称，开放权重模型占 token 使用量的比例已从两年前的约 20% 升至如今的 80% 以上。Cloudera 还在其数据服务中加入面向数据管理、GPU 容量优化等场景的智能体工作流，但 CIO 在信任 AI 自动化前仍需弄清智能体如何被控制、留下何种审计轨迹。
正文
Enterprises racing toward AI automation are finding that where their models and data run matters as much as what the models can do. Regulatory pressure, rising token costs and fear of lock-in are pushing companies to demand more control over their AI workloads.
That demand is shaping how data platform providers pitch their next act. Cloudera Inc. started out in the Hadoop market and later moved to a unified data layer that spans clouds and on-premises infrastructure. It is now positioning itself as an anywhere cloud for AI, according to Paul Gillin (pictured, right), enterprise editor of SiliconANGLE Media Inc.
“I think big companies decided a long time ago. They didn’t want to be beholden to a single cloud provider and that’s sort of baked into the landscape now,” Gillin said. “I think what’s different about Cloudera is that they’re strong proponents of on-premises infrastructure, of sovereign clouds, of companies being able to run their workloads wherever they want and move from one to another.”
Gillin spoke with theCUBE’s Rebecca Knight (left) at EVOLVE26, during an exclusive broadcast on theCUBE, SiliconANGLE Media’s livestreaming studio. They discussed AI automation and provided analysis of the event’s keynote speech. (* Disclosure below.)
Sovereign AI and purpose-built agents drive AI automation
Sovereign and private AI came up frequently in the EVOLVE26 keynote. Surveys point to regulation as the main reason enterprises want to keep AI inside their own walls. Open-weight models give them a way to do it, having grown from about 20% of token use two years ago to more than 80% today, according to Gillin. Cloudera is also adding agentic workflows across its data services. General-purpose agents are more of a consumer phenomenon, he explained.
“They’re more of an office worker tool, but the Cloudera agents are tuned to managing data,” Gillin said. “To managing workloads, optimizing GPU capacity, the types of things that really concern CIOs, because they’re paying the bills for these very high training costs.”
Simply having an agent is no longer an edge, so Cloudera is doubling down on its long-standing multiplatform play. It is extending the portability enterprises wanted from their clouds to their AI models, according to Gillin. Before trusting that pitch with AI automation, CIOs will want to know how agents are controlled, what audit trails they leave behind and how agent decisions can be documented for regulators, he noted.
“This is a very big issue with agents right now,” Gillin said. “It’s still a wild west out there with agents.”
Stay tuned for the complete video interview, part of SiliconANGLE’s and theCUBE’s coverage of EVOLVE26.
(* Disclosure: TheCUBE is a paid media partner for EVOLVE26. Sponsors of theCUBE’s event coverage do not have editorial control over content on theCUBE or SiliconANGLE.)
Photo: SiliconANGLE