In this interview from IBM Think 2026, Ric Lewis, senior vice president of infrastructure at IBM, joins theCUBE's John Furrier and Dave Vellante to discuss how IBM is evolving its infrastructure portfolio to serve as the foundation for enterprise agentic AI. Lewis opens by highlighting Z's sustained momentum — growing 20 to 30% program to program for nearly a decade as new workloads and on-chip AI capabilities gain traction. He details IBM's plan to integrate ARM natively into the next generation of Z systems at the hardware level, enabling enterprises to run ARM ecosystem apps at full performance alongside traditional workloads without emulation. Lewis frames hybrid as inevitable: because AI quality depends entirely on data quality, and that data is distributed across clouds, on-premises systems and edge environments, IBM is engineering infrastructure that brings AI capability directly to the data.

The conversation also explores the three hurdles slowing enterprise AI adoption: hybrid data access, data quality and scalability. Lewis introduces a sharp distinction — enterprise data is either a gold mine or a landfill, and applying AI to a landfill only accelerates the digging. IBM's storage platform addresses this with Content Aware, a feature designed to surface and structure AI-ready data within large repositories. On sovereignty, Lewis notes that IBM processes 70% of the world's financial transactions through Z systems already deployed in banks and financial institutions, making sovereign AI IBM's home court advantage. The discussion turns to how client intimacy has driven IBM's deepest innovations — inline fraud detection embedded directly in Z processors, developed years before the generative AI wave, now returning hundreds of millions of dollars to financial clients by stopping fraudulent transactions at the point of occurrence. From fit-for-purpose quantum computing as a next-generation co-processor to IBM's internal deployment of AI across a 13,000-person global support organization, Lewis outlines why infrastructure specialization — not consolidation — defines the path forward.

In this interview from IBM Think 2026, Jean-Michel Garcia, chief technology officer of BNP Paribas, joins theCUBE's John Furrier to discuss how Europe's largest bank is moving from AI experimentation to industrialized deployment at scale. Garcia, who oversees infrastructure, application security and production operations for a team of roughly 13,000, reflects on BNP Paribas's structured approach to AI governance — establishing a group-level framework that now encompasses nearly 1,000 active use cases generating approximately 750 million in annual value. He explains how the bank avoided the pitfalls of its prior digital transformation by deliberately balancing local experimentation with the longer-term discipline required to extend AI across the full enterprise perimeter.

The conversation also explores BNP Paribas's evolving AI factory and token strategy, including the trade-offs between running open-source models on-premises and sourcing tokens externally as frontier model performance shifts rapidly. Garcia details how the bank is separating inference from training workloads and pursuing hardware-agnostic token economics to reach a neutral cost position. He discusses the pragmatic approach to developer tooling — evaluating IBM's Bob, Cursor and GitHub Copilot in parallel rather than converging prematurely on a single platform, preserving flexibility as the ecosystem matures. From the imperative to protect client data in regulated environments to the challenge of making strategic infrastructure decisions without any historical precedent, Garcia outlines what it means to be a frontier practitioner navigating the hybrid AI era at the scale of a global bank.

In this interview from IBM Think 2026, Sripriya Srinivasan, general manager of core and ALM software products at IBM Corp., joins theCUBE's Dave Vellante to discuss the launch of IBM Sovereign Core and how enterprises can move from AI pilots to governed, production-scale deployment. Srinivasan traces the evolution of digital sovereignty across four dimensions — data, operational, technology and AI — arguing the fundamental driver is not regulatory compliance but operational resiliency. She details Sovereign Core, a software platform made generally available at the event, built around a control plane, base services and an extensible catalog spanning IBM, third-party and open source technologies. A key design principle places the full control plane within the client's own boundary — a direct contrast to what she calls "Sovereign 1.0" solutions, where keys and governance still reside with outside vendors.

The conversation also explores the technical foundations enabling Sovereign Core's flexibility, including a "build once, deploy anywhere" architecture inherited from IBM's internal hybrid cloud platform and reinforced by Red Hat components including OpenShift, Quay and ACM. Srinivasan introduces the "golden patterns" framework — a set of pre-tested, trusted configurations that establish a secure baseline while still allowing enterprises to bring their own models and third-party services within a governed boundary. She describes a go-to-market approach centered on managed service providers, hardware partners and IBM's existing regulated-industry customer base, noting that sovereignty concerns are just as pressing in US financial services and healthcare as they are in Europe or Asia-Pacific. The consumption-based pricing model, built on resource units rather than bespoke contracts, gives enterprises a predictable on-ramp. From compressing AI deployment timelines from 12 to 18 months down to hours, to building a platform the client fully controls, Srinivasan outlines how IBM is positioning Sovereign Core as the enterprise control plane for the next era of AI.

In this interview from IBM Think 2026, Mohamad Ali, senior vice president of IBM Consulting, joins theCUBE's John Furrier and Dave Vellante to discuss how enterprises are moving from AI experimentation into orchestrated, production-ready systems — and how IBM is positioning itself as the control layer for hybrid human-and-digital workforces. Ali walks through three flagship customer success stories from the Think keynote: Providence Health now recruits nurses 12 days faster using watsonx agents integrated with Oracle; Fortitude has compressed insurance claims processing from weeks to days; and Pearson Advantage applies IBM's own employee badging system to credentialize digital workers. Underpinning all three is IBM Consulting Advantage, a management layer running 4,000 digital workers across 450 active projects — on any AI stack, whether watsonx, Anthropic or OpenAI — while preserving enterprise governance and observability.

The conversation also explores how IBM is translating agentic AI into measurable business outcomes. Ali details how the firm decomposed its own operations into 490 workflows, re-engineered 70 of them and redirected $4.5 billion from a $25 billion cost base, with consulting profits expanding 20% from 2024 to 2025. The Aramco partnership illustrates the upper bound of what's possible: working alongside a company that has built its own LLM to solve industrial-scale problems like pipeline corrosion prediction, with potential to generate entirely new businesses. Ali also describes a third phase of AI value, where enterprises are moving beyond workflow efficiency to unlock new revenue models through dynamic pricing and real-time market intelligence. From championing reverse mentorship — requiring his top 100 executives to build and manage digital workers in Python — to reframing shadow AI as a feature rather than a risk, he outlines how IBM is establishing the governance infrastructure for the era of autonomous enterprise work.

In this interview from IBM Think 2026, Javier Olaizola, global managing partner for hybrid cloud and data at IBM Consulting, joins Jason Kelley, global head and managing partner for core business applications at IBM Corp., to talk with theCUBE's John Furrier about how enterprises are operationalizing agentic AI at scale and why IBM is positioning itself as the orchestration layer for enterprise transformation. Olaizola notes a decisive shift in boardroom conversations — clients have moved from backward-looking compliance checks to forward-looking exposure assessments — as geopolitical tensions, new regulations and rapid AI evolution converge. Kelley adds that AI is pressing enterprises to rethink their business models from first principles, separating the winners who realign their operating models around AI from those who apply it only at the fringe.

The conversation also explores what Kelley frames as the "orchestration era" — a recognition that modern enterprises carry accumulated debt across process, technology, systems and data, and that scaling AI requires a conductor to align all of it toward measurable outcomes. Both guests underscore that hybrid cloud has moved from a strategic debate to the de facto standard, with technology now demanding the same boardroom attention as a P&L. Olaizola introduces IBM's concept of "AI primitives" — governance, trust and security as the foundational layer enabling confident deployment across sovereign environments, illustrated by a consumer-industry client scaling a successful AI agent from a single country to a hundred. From co-developing AI-first assets with ecosystem partners to pursuing industry-specific workloads, Kelley and Olaizola outline why IBM Consulting's role as orchestrator — not just implementer — defines its competitive advantage in the agentic AI era.