TheCUBE spent 2025 talking directly with top tech executives as they worked through how computing is being rebuilt inside their organizations, driven by real-world constraints rather than abstract roadmaps.

Across those conversations, a consistent shift emerged away from theory and toward execution. Enterprise leaders are now focused on how infrastructure is evolving, how software must adapt to new demands and how silicon is being pushed to support intelligence at scale.

“First of all, of all the CEO keynotes I’ve watched over the past 30 years, you use the word computer science and physics more than anyone else,” John Furrier (pictured, left), executive analyst at theCUBE Research, told Jensen Huang (right), president and chief executive officer of Nvidia, during this year’s GSA Awards. “You mentioned this industry, community and what it means to the society and the world — the growth, not only the growth but the impact.”

Huang and other top tech executives spoke to theCUBE analysts during a variety of events and interviews throughout 2025 as a part of theCUBE, SiliconANGLE Media’s livestreaming studio’s ongoing media coverage. TheCUBE’s 2025 event coverage highlighted the shift from AI theory to real-world execution across infrastructure, software, silicon and enterprise operations.

Here are 31 exclusive interviews from theCUBE that show how 2025 reshaped enterprise technology.

1. Top tech executive on reinventing computing for the AI era

Computing is being reinvented from the ground up as static, pre-written software gives way to systems that generate intelligence in real time, shaped by context, memory and the questions being asked. Nvidia’s Jensen Huang described this shift toward globally distributed “intelligence factories” — from centralized AI data centers to edge and embedded systems — during a conversation with John Furrier at the GSA Awards.

Read more of theCUBE’s exclusive interview with Huang.

2. How AWS is rethinking AI infrastructure

AI infrastructure is being redesigned around campus-scale systems that deliver shared benefits across customers. AWS CEO Matt Garman explained how scale translates into enterprise value in a discussion with theCUBE’s at AWS re:Invent.

Check out theCUBE’s complete interview.

3. Why AI factory demand isn’t slowing

AI factory demand continues to rise as models grow more complex and token volumes surge across workloads. Dell Technologies CEO Michael Dell talked with Furrier and Dave Vellante about why the curve keeps climbing, as a part of theCUBE + NYSE Wired: AI Factories series.

Read the full story on theCUBE.

4. Securing the SOC at machine speed

Cyber threats now operate faster than human response times, reshaping how security operations must function. CrowdStrike CEO George Kurtz discussed why agentic SOCs restore speed and resilience with Dave Vellante and Rebecca Knight at CrowdStrike Fal.Con.

Don’t miss the full interview on theCUBE.

5. When data, not compute, sets the pace

As AI shifts toward inference and autonomous agents, data quality and governance increasingly define performance. Data discipline matters more than raw compute, according to Vast Data CEO Renen Hallak during an interview with theCUBE’s at theCUBE + NYSE Wired: Physical AI & Robotics Leaders event.

Watch the full interview from theCUBE.

6. Infrastructure choices behind enterprise AI

Enterprise AI infrastructure decisions are shifting toward efficiency, simplicity and predictable outcomes. This is causing data center priorities to evolve, HPE CEO Antonio Neri said in an interview with at Nvidia GTC.

Read more of theCUBE’s exclusive conversation with Neri.

7. Enterprise AI adoption lags infrastructure reality

AI infrastructure investment is outpacing real-world enterprise adoption, creating uneven progress across industries, according to Diane Bryant, independent director for Broadcom Inc., Celestial AI Inc., Haemonetics Corp. and mmTron. She explained why SaaS leads deployment while custom implementations lag, during a conversation at theCUBE + NYSE Wired: Physical AI & Robotics Leaders event.

Check out theCUBE’s complete interview.

8. Enterprise cloud hits execution mode

Cloud adoption discussions have moved from reassurance to real execution as deployments scale. This is the main driver behind VMware Cloud Foundation’s growing traction and delivery momentum, said Broadcom’s Krish Prasad in a conversation with theCUBE at VMware Explore.

Read the full story on theCUBE.

9. Turning agents into real enterprise tools

Enterprise AI agents must be reliable, governed and measurable to create lasting value, according to Databricks CEO Ali Ghodsi. Practical automation is replacing hype, he explained in a conversation during theCUBE + NYSE Wired: AI Factories – Data Centers of the Future series.

Watch theCUBE’s full sit-down.

10. Crypto stress tests reveal market maturity

Market shocks are revealing which parts of the crypto ecosystem are resilient under pressure. Fortunately, for investors, Bitcoin and Ethereum have held firm during this period, said Fundstrat CIO Tom Lee in a discussion during theCUBE + NYSE Wired: Crypto Trailblazers event.

Hear the full story on theCUBE.

11. Snowflake improves its AI data foundation

Snowflake is elevating unstructured data with semantic context and GPU efficiency to improve AI accuracy. This means cost control and precision matter more than ever, explained Snowflake co-founder Benoit Dageville when he spoke with theCUBE at RAISE Summit.

Watch the full interview from theCUBE.

12. Marketing meets real-time data and AI

Modern marketing now depends on real-time data paired with AI-driven relevance and speed. Enterprises can gain an advantage through this personalization, noted Snowflake CMO Denise Persson in a conversation at Snowflake Summit.

Don’t miss the full interview on theCUBE.

13. Redefining enterprise software for the agent era

Enterprise software is evolving from static tools into digital labor powered by autonomous agents. Salesforce CEO Marc Benioff spoke with theCUBE’s George Gilbert and Dave Vellante on The Road to Service-as-Software podcast about how this is shifting productivity and execution.

Read more of theCUBE’s exclusive event analysis.

14. Why collaboration now drives semiconductor leadership

Semiconductor leadership increasingly depends on collaboration across manufacturing, software and ecosystems. Partnerships are helping to shape this momentum, Broadcom President Charlie Kawwas explained in a conversation with theCUBE during the GSA Awards.

Check out theCUBE’s complete interview.

15. Open-source shaping the next AI wave

Open source is becoming central to scaling AI efficiently as agentic workloads grow, said Red Hat CEO Matt Hicks at RAISE Summit. He discussed how openness keeps inference affordable and flexible.

Watch theCUBE’s full sit-down.

16. IBM’s agentic AI inflection point

Agentic AI marks a shift from assistive systems to technologies that take action across workflows. TheCUBE’s Dave Vellante covered these new orchestration demands in more detail in a conversation with Ritika Gunnar, general manager for data and AI at IBM, at IBM Think.

Read the full story on theCUBE.

17. From experimentation to production-ready AI

Moving AI into production requires balancing personalization, governance and cost across environments. Enterprises are working to make that transition, Bruno Aziza, group VP for data, business intelligence and AI at IBM, discussed with theCUBE during IBM Think.

Don’t miss the full interview on theCUBE.

18. Energy-first AI factories reshape where compute gets built

AI progress is increasingly constrained by power, pushing infrastructure closer to abundant energy sources. That shift enables massive, high-density AI factories built around renewables rather than legacy hubs, according to Chase Lochmiller, co-founder and CEO of Crusoe Energy Systems, at theCUBE + NYSE Wired: Physical AI & Robotics Leaders event.

Hear the full story on theCUBE.

19. Why private cloud modernization can’t wait

Legacy infrastructure is struggling under AI-driven demands for density and efficiency. In a conversation with theCUBE’s Rob Strechay during Dell’s “5 Steps to a Smarter Private Cloud” event, Juan Martinez, senior director of OEM/ODM enterprise and HPC, at AMD, explained why modernization timelines are accelerating.

Read more of theCUBE’s exclusive event analysis.

20. Seamless AI upgrades meet enterprise trust

Oracle is embedding AI directly into databases to avoid disruptive migrations, according to Oracle EVP Juan Loaiza. This is important for trust, governance and continuity to remain intact, he pointed out in an interview with Dave Vellante during an exclusive CUBE Conversation.

Check out theCUBE’s complete interview.

21. Google Cloud’s full-stack AI momentum

Google Cloud’s AI strategy reflects a shift from experimentation to enterprise-scale execution, said Google Cloud CMO Alison Wagonfeld. She discussed how full-stack investment drives value in a talk with John Furrier at RAISE Summit.

Watch the full interview from theCUBE.

22. When data becomes differentiated AI

Differentiated AI starts with disciplined data evaluation and strong model grounding, according to Unconventional AI CEO Naveen Rao in a conversation with theCUBE at RAISE Summit, he discussed how infrastructure and intelligence intersect.

Hear the full story on theCUBE.

23. Optical interconnects take center stage

AI system performance increasingly depends on efficient processor communication rather than raw compute. This makes optical interconnects crucial for performance, Celestial AI CEO Dave Lazovsky explained during theCUBE + NYSE Wired: Mixture of Experts event.

Watch theCUBE’s full sit-down.

24. Turning AI factories into real business value

AI factories are becoming enterprise platforms that translate GPU investment into measurable outcomes. In a talk with theCUBE’s John Furrier at SC25, DataDirect Networks CEO Alex Bouzari discussed orchestration strategies that are making this a reality.

Check out theCUBE’s complete interview.

25. Rethinking cybersecurity for an AI-driven threat landscape

AI has shifted cybersecurity toward detection-first models that assume breach. This is why hunting matters more than perimeter defense, Palo Alto Networks founder Nir Zuk emphasized in an interview with theCUBE at RSAC 2025.

Don’t miss the exclusive insights on theCUBE.

26. Networking becomes the AI factory backbone

AI factories now resemble distributed supercomputers where networking defines performance, according to Gilad Shainer, SVP of marketing at Nvidia. He discussed why interconnects are strategic during theCUBE + NYSE Wired: AI Factories series.

Read the full story on theCUBE.

27. Building AI factories at cloud scale

Efficient scaling has become the defining challenge for AI factory design. In a talk with theCUBE’s John Furrier and Dave Vellante during theCUBE + NYSE Wired: AI Factories series, Roman Chernin, co-founder and chief business officer of Nebius, discussed the realities of deployment.

Don’t miss the full interview on theCUBE.

28. Energy demand reshaping AI infrastructure

AI growth is colliding with unprecedented energy demand across data centers. This power constraint is reshaping infrastructure planning, Hut 8 Chairman Bill Tai explained in a discussion with theCUBE at theCUBE + NYSE Wired: Physical AI & Robotics Leaders event.

Watch the full interview with theCUBE.

29. Open platforms at AI scale

AI-native applications demand open platforms optimized for constant throughput and adaptation. In an interview with theCUBE’s John Furrier and Dave Vellante during theCUBE + NYSE Wired: AI Factories series, Together Computer CEO Vipul Prakash discussed why flexibility outperforms fixed stacks.

Read more of theCUBE’s exclusive event analysis.

30. Optics reshape AI scale-up

Optical I/O is changing how AI factories scale by reducing latency, power consumption and physical constraints. Ayar Labs co-founder and CTO Vladimir Stojanovic discussed the fabric-first shift during theCUBE + NYSE Wired: AI Factories series.

Check out theCUBE’s complete interview.

31. Storage density reshapes AI infrastructure

High-capacity flash is redefining how AI systems are fed efficiently at scale, according to Greg Matson, SVP of products and marketing at Solidigm. He discussed why balancing performance, power and footprint matters, during Dell Tech World 2025.

Hear the full story on theCUBE.

For all of theCUBE’s 2025 event coverage and exclusive industry conversations, visit our YouTube channel.

Photo: SiliconANGLE