Artificial Intelligence has rapidly become one of today’s most important technologies. AI can unlock new efficiencies, help us find amazing new insights, and work at unprecedented scale and complexity.
But AI is still far from where it needs to be. While it has become easier for enterprises to build an impressive AI pilot, it’s still too hard to roll these pilots into production and capture the value of AI at scale.
General purpose LLMs are often slow and costly and data in AI projects is often poorly governed. It’s no surprise Gartner found that on average, more than half of AI projects never make it into production—and those that do take 8 months to go from AI prototype to production.
This is in addition to the data problem we continue to navigate for our customers. They spend long periods making sure their proprietary data, the heart of value creation, is secure, accurate, and well governed. It’s a critical step, because flawed data increases the risk the model will "hallucinate" plausible but entirely false information, leading to enterprise chaos, financial loss, or reputational damage.
Most of the industry doesn't have a lot of good solutions for this dilemma. And the problem looks even tougher as enterprises seek to deploy specialized AI applications and build agentic AI. Agents are the AI applications we envision learning, deciding, and managing complex systems on their own—but they’ll fail if they’re slow and risky.
At Rubrik, though, we’ve always found better ways to handle data-centered challenges. Today we are excited to announce that Predibase, a cutting-edge platform for AI model training, fine-tuning, and deployment, will be joining Rubrik, a company trusted for delivering secure, governed data. Together, we can equip AI teams with the tools and building blocks needed to quickly bring enterprise-grade AI applications to production—faster, safer, and more effectively than ever before, for enterprises everywhere.
A better way to build AI
Predibase was founded to help enterprises get AI initiatives into production, developing standard open source tools for customizing and optimizing AI models to accommodate proprietary data and specific business needs.
They then released LoRA eXchange, an open source framework developers use to serve thousands of fine-tuned AI models on a single GPU. In March of this year, the company introduced its cutting-edge end-to-end platform for reinforcement fine-tuning, enabling organizations to improve model quality on tasks through real-time feedback.
Besides model customization, Predibase focuses on efficient, real-world deployment. Their tools allow customized models to run and scale—economically and quickly. This turns what was cutting-edge research into a tailored, efficient, and reliable business asset. By lowering costs, improving performance, and simplifying deployment, Predibase accelerates enterprise AI adoption, unlocking tangible business value.
When these capabilities are paired with secure data platforms like Rubrik's, they can be transformative, helping ensure trusted data powers responsible and impactful AI. Just as Rubrik has simplified access to secured, governed data for AI, Predibase is solving the performance and cost issues around deploying large language models for agentic and other AI applications.
Rubrik, the crucial next step
Throughout Rubrik’s history we’ve evolved to meet our customers' needs. We began by transforming the antiquated backup and recovery industry, shifting focus from legacy disaster recovery to cyber resilience—equipping organizations to withstand and recover from cyberattacks. As data became constantly in motion and operations went fully online, we further evolved our platform to be the ultimate defense against ransomware, ensuring businesses could always stay operational.
Today, our secure data platform breaks down silos, combines robust recovery with the clean, secure data essential for powering advanced AI. AI applications must honor existing access controls and prevent unauthorized information sharing. An AI system that inadvertently exposes confidential customer data or makes biased decisions based on unvetted, insecure information isn’t innovation—it’s devastation. And we’re here to stop that.
Rubrik’s data lake provides the necessary access controls, permissions, and policies to securely power AI applications. Predibase delivers operational value on top of improved, customized models, enabling enterprises to build, pilot and deploy AI safely at scale. When teams are equipped with Predibase and our secure data lake, they are empowered to accelerate AI adoption.
What Predibase has done with model tooling, low-friction customization, model serving, and process efficiency is a classic example of engineering a technology to work better, in an ever-widening number of cases. Predibase will continue to innovate and define new frontiers in enterprise agentic AI.
Turning security costs into value-creating investments
For a long time, data security has been a defense against cyber attacks—an essential insurance policy against the inevitable threats. At Rubrik, we pride ourselves on being that unshakable foundation. When an attack happens, our customers can get back online in minutes, without skipping a beat, proving the value of resilience. That capability remains paramount.
But we also recognize a new and untapped potential exists in the very data we secure, thanks to the way we built Rubrik. The combination of our data lake with a platform like Predibase will mean that secured data isn't just a corporate necessity; it is a key enabler for building AI, accelerating its deployment and lowering TCO. Data can become a powerful competitive asset. From here, it’s easy to see a world where AI is deployed both responsibly and effectively, creating the necessary trust and value required by any landmark innovation.
Our platform strategy allows us to innovate fast and deliver effective solutions to critical challenges at the epicenter of data, AI, and security. We continue to add capabilities and investments to enable secure GenAI across environments. And with Predibase, we will make strides to accelerate our customers’ journey and to bring secure data to power AI. This will be an exciting evolution that benefits us long-term, and we invite you to join us on this ride.