Jim O’Dorisio, the Senior Vice President and General Manager of HPE Storage, sits at the vital intersection of enterprise infrastructure and the cutting edge of digital innovation. As organizations worldwide grapple with the dual pressures of integrating artificial intelligence and defending against sophisticated cyber threats, O’Dorisio provides a strategic roadmap for navigating these complex waters. His perspective is grounded in the reality of the data center, where the hype of AI meets the practical necessity of storage scalability, data lineage, and the rising costs of traditional virtualization. This conversation explores how the modern enterprise is moving beyond the experimental phase of model training to the high-stakes world of real-time inferencing and resilient recovery.
The discussion centers on the pivotal shift in AI strategies, moving away from pure model development toward extracting tangible value through inferencing and robust data governance. We examine the evolving landscape of cybersecurity, where the focus has transitioned from perimeter defense to ensuring operational continuity through air-gapped environments and rapid recovery points. Furthermore, the dialogue addresses the economic shifts triggered by Broadcom’s 2023 acquisition of VMware, which has prompted many organizations to aggressively seek more cost-effective hypervisor alternatives to fund their innovation pipelines.
The industry is currently witnessing a significant pivot as companies move from the experimental stages of AI toward production. How are you seeing the focus shift from the heavy lifting of model training toward the more immediate demands of inferencing?
We have reached a point where the initial excitement of training massive models is settling into a more pragmatic understanding of how to actually extract value from data. It is no longer just about the raw power required to build a model; it is becoming much more about inferencing, where the AI actually performs the tasks it was designed for in a live environment. To get this right, an enterprise must have a crystal-clear data strategy that defines the specific use cases and the exact problems they are trying to solve. Without this focus, businesses find themselves lost in the noise of a market that has an abundance of offerings but very little direction. We are helping customers transition into this phase by providing the intelligent storage layers necessary to support the high-speed demands of real-time data application.
You’ve noted that many organizations struggle with activating their AI because they lack a clear understanding of their own data. What are the specific risks and challenges associated with deploying AI infrastructure without a firm grasp of data lineage and governance?
The biggest challenge we see is not necessarily the hardware or the software, but the lack of understanding regarding where data comes from and where it is going. If an IT executive stands up a massive infrastructure—complete with a three-tier architecture, advanced networks, and compute nodes filled with GPUs—they are only halfway there. The real danger lies in the governance phase; if you don’t understand your data lineage, you might inadvertently send sensitive information into an AI model that shouldn’t have access to it. This lack of visibility can lead to significant security gaps and compliance failures that are difficult to walk back once the model is in use. We are seeing a major push for more intelligent storage solutions that can provide the necessary guardrails to ensure data integrity across the entire AI lifecycle.
Cyber-protection remains a top-of-mind issue for boardrooms, yet the focus seems to be shifting from prevention to resilience. How are you advising customers to prepare for the reality that a ransomware attack is often a matter of “when” rather than “if”?
The days of believing a perimeter could stop every single threat are behind us, and our ransomware workshops are consistently packed because leaders realize the stakes have changed. The goal now is to stay out of the news and keep the business running even when an incident occurs, which requires a fundamental shift toward rapid recovery. We are implementing solutions where agents can monitor the environment, identify an unauthorized action that shouldn’t have happened, and then immediately recover the entire environment back to a specific point in time. This involves creating air-gapped environments that provide a safe haven for data, ensuring that even if the primary systems are compromised, the operation remains functional. It is a sensory shift in the data center—moving from the anxiety of a breach to the confidence of knowing you can restore operations in a matter of minutes.
The landscape of hybrid cloud is being reshaped by the 2023 acquisition of VMware by Broadcom, leading many to re-evaluate their long-term strategies. Why are we seeing such a surge in migration evaluations right now, and how is it impacting IT budgets?
The timing is largely driven by the natural expiration of renewal contracts, combined with the fact that Broadcom has introduced licensing approaches that have made costs significantly more expensive for many customer segments. It is getting harder to find any enterprise that isn’t at least evaluating their options, whether that means looking at our specific hypervisor offerings or exploring other alternatives in the market. Executives are looking for ways to create a more operationally efficient infrastructure so they can take the money they save on traditional licensing and reinvest it into AI or cyber-resiliency. This is a board-level conversation because it directly impacts the ability of a company to fund its competitive advantages while maintaining its core operational foundation. By optimizing the storage and virtualization layer, companies are finding the capital they need to drive their next generation of digital growth.
What is your forecast for the future of data management as AI and cyber-protection become even more intertwined?
I believe we are heading toward a future where storage is no longer a passive repository but an active participant in an organization’s security and intelligence strategy. We will see the rise of highly scalable, intelligent storage layers that are capable of self-healing and autonomous governance, effectively acting as the first line of defense against both cyber threats and data mismanagement. As AI agents become more integrated into our workflows, the infrastructure will need to be robust enough to detect and correct errors in real-time, preventing “hallucinations” or data leaks before they can impact the business. Ultimately, the winners in this space will be those who treat their data as their most valuable asset, protected by air-gapped security and powered by a strategy that prioritizes actionable insights over raw storage capacity. Efficiency and resilience will be the two pillars that define the successful enterprise of the next decade.
