In this interview from the Databricks Data + AI Summit 2026, Samuel Bonamigo, senior vice president and general manager of EMEA at Databricks, joins Dee Fitzgerald, vice president and head of data and analytics at Danone, to talk with theCUBE's John Furrier about how global enterprises are building trusted data foundations to unlock AI at scale. Bonamigo underscores EMEA's strategic weight, noting that more than 2,500 attendees from the region — roughly 700 companies — traveled to San Francisco as the event's second largest delegation. He details Databricks' regional investment commitments of $850 million in the UK, $400 million in Germany and $300 million in France. Fitzgerald grounds the conversation in Danone's global scale — 27 billion euros in revenue, operations across 120 countries and over 1.5 billion consumers — and explains how a five-year Databricks partnership has sharpened the company's focus on data trust, quality and governance as the prerequisites for AI readiness.

The conversation also explores Danone's "Talk to My Data" initiative, which uses Databricks' Genie and Ontology capabilities to let business users query enterprise data in natural language and surface actionable insights without relying on static dashboards. Fitzgerald argues that ontology complements rather than replaces traditional metadata, with humans remaining in the loop to validate context and confirm data standards as AI takes on more of the underlying work. Bonamigo adds that capturing the full value of these capabilities requires deep investment in skills and enablement, not just infrastructure — bringing business and IT teams together around shared data definitions and measurable outcomes. From bridging fragmented data estates to fostering a "better together" ecosystem with partners, both guests make the case that disciplined data strategy, not AI tooling alone, will determine which enterprises are positioned to compete in the years ahead.

In this interview from Databricks Data + AI Summit 2026, Bryan Clark, director of product management at Databricks, joins Grant Veazey, chief technology officer and enterprise performance manager at Ensemble Health Partners, to talk with theCUBE's John Furrier about how unified data architecture is enabling the shift from AI pilots to production-grade agentic applications in healthcare. Clark details Lakebase, Databricks' approach to consolidating transactional and analytical workloads into a single-copy architecture that eliminates brittle data pipelines. By translating row-based Postgres formats directly into columnar open formats, Lakebase gives enterprises the best of both analytical and operational query engines without forcing trade-offs. Veazey explains how Ensemble Health Partners applied this immediately — cutting the iteration cycle for its clinical appeal agent from months of custom pipeline work down to hours.

Key themes include how Ensemble Health Partners is navigating the shift to an agentic-first data strategy, where observability and traceability of agent transactions feed back into the Lakebase architecture to support governance in a highly regulated environment. Veazey details how ontologies — which Clark frames as "agent dreaming," a background process that builds a persistent memory map of data relationships rather than reasoning from first principles each time — are critical for healthcare's complex web of denial codes, clinical documentation and payer interactions. The conversation also covers Veazey's approach to AI prioritization: not cost savings or automation, but solving problems that were previously out of reach, such as driving healthcare denial rates from 3% toward zero. From enabling cross-cloud data mobility that lets organizations dynamically shift workloads to lower-cost GPU capacity to building cyber resilience into every layer of the platform, the discussion illustrates how the convergence of transactional and analytical data is turning long-standing enterprise friction into measurable, production-ready results.

In this interview from Databricks Data + AI Summit 2026, Stephen Orban, senior vice president of product ecosystem and partnerships at Databricks, joins Mike Palmer, chief executive officer of Sigma Computing, to talk with theCUBE's John Furrier about the consolidation of enterprise data stacks and the shift from isolated AI experiments to governed, production-ready intelligence at scale. Palmer argues that the enterprise software market is at an inflection point, with hundreds of standalone SaaS contracts collapsing into platform-led solutions — a shift enabled by Databricks' ability to unify structured, unstructured and on-premises data under a single governed layer. Orban details the summit's flagship ecosystem announcements: transactable marketplace capabilities that allow customers to apply pre-committed Databricks spend toward partners like Sigma, Genie agent sharing for distributing AI agents to customer endpoints and a new storage ecosystem that makes on-premises data from providers including VAST, MinIO and Qumulo accessible directly within Databricks.

The conversation also explores how deep, differentiated partnerships are replacing generic integrations — Palmer describing Sigma's approach as becoming the first adopter of every new Databricks capability, from Unity Catalog to AgentBricks. Real-world deployments illustrate the stakes: DraftKings relies on the combined Databricks and Sigma stack to shift betting lines in real time, while JPMorgan is using the platform to rebuild core banking workflows across portfolio modeling, compliance and wealth management. Palmer frames enterprise AI readiness as a three-tier investment — trusted data, governed workflows and agentic automation — noting that organizations now effectively operate a round-the-clock workforce as agents run continuously in the background. Orban echoes the shift, observing that well-governed infrastructure built on Unity AI allows even non-engineers to build production applications under policy-enforced guardrails. From the rise of the citizen builder to the emergence of net-new services that weren't possible a year ago, both executives underscore that the most competitive organizations will be those that treat AI not as a productivity multiplier, but as a foundation for entirely new business models.

In this interview from Databricks Data + AI Summit 2026, Andrew Krioukov, general manager of Lakewatch at Databricks, joins Patrick Wright, chief technology and operations officer at National Australia Bank, to talk with theCUBE's John Furrier about the transformation of enterprise security from manual operations to agentic, data-driven defense at machine speed. Wright frames the threat landscape plainly: attackers are now hyper-automated, deploying agents and large language models to find vulnerabilities faster than the developers who wrote the software. Krioukov explains why security has become a big data problem that legacy SIEMs cannot solve at the required velocity. The Panther acquisition, announced at the summit keynote, accelerates Lakewatch's vision of ingesting every data type — telemetry, code, Jira tickets, Slack messages — to give agents the context needed to separate a legitimate user session from a compromised one.

The conversation also explores National Australia Bank's strategy of unifying security data with business operations data — financials, fraud signals and network availability — inside the same Databricks lakehouse rather than isolated systems. Wright details why he is now hiring software developers and data scientists instead of traditional security operators, arguing the modern SOC must extract signal from noise in milliseconds, not minutes. He also shares how Databricks' Genie agents are helping democratize information across the enterprise, giving every employee access to faster, more actionable intelligence. From supercharging security operations centers with risk-based detections to Wright's broader vision of a streamlined, AI-accelerated bank, the discussion maps a concrete path from fragmented data and AI experiments toward production-ready defense at scale.

In this interview from Databricks Data + AI Summit 2026 in San Francisco, Jonathan Frankle, chief AI scientist of Databricks, joins theCUBE's John Furrier to discuss the systems revolution reshaping how enterprises build, govern and scale AI in production. Frankle frames the platform across four interconnected layers — systems, models, applications and runtime — arguing that injecting intelligence into an organization changes everything. He details Lakebase, engineered specifically for agents, highlighting its GitHub-style branching and separation of storage and compute as foundational primitives for agentic workloads at scale. Frankle also explains how his team has spent the past year moving from evaluation to optimization, using reinforcement learning to build custom models that are faster, cheaper and more accurate than closed-source alternatives for targeted enterprise tasks.

The conversation also explores the industry's shift from token maximization to measurable business value — a pivot Frankle frames as the defining storyline of 2026. He urges organizations to treat AI deployment as a "fractal" problem, starting by working alongside the humans doing the work today, bootstrapping evaluations through MLflow and proving a concept before committing to fine-tuning. Frankle reframes ontologies as a practical form of agent memory — breadcrumbs that let agents accumulate knowledge across calls rather than rediscovering the same data every time. From the launch of Unity AI Gateway for intelligent model routing to Unity Catalog as the governance backbone for AI agents, he provides a roadmap for how organizations can move beyond the buzz phase and deploy trusted, production-grade AI with cost control and security built in from the start.

In this interview from Databricks Data + AI Summit 2026, Bob Pisani, chief technology officer of Addepar, joins theCUBE's John Furrier to discuss how AI and the Databricks lakehouse are transforming portfolio intelligence and unlocking the largely untapped data potential of private markets. Pisani explains how the platform, which oversees more than $9 trillion in assets across family offices, registered investment advisors and private banks, made an architectural bet on Databricks roughly four years ago. By adopting the Lakehouse and Unity Catalog, Addepar built an alternatives data management solution that applies LLMs and agents to extract structured positions and transactions from documents previously trapped in PDFs and spreadsheets — giving clients a unified view across alternative and public investments for the first time.

The conversation also explores how Addepar's graph-native foundation, built over 15 years, is generating powerful network effects by analyzing anonymized, cross-sectional patterns across the full $9 trillion on the platform. Pisani highlights how this lens has already produced counterintuitive market insights, including data showing capital continuing to flow into the US during a period when conventional narratives predicted capital flight driven by tariffs and global uncertainty. On trust, he outlines how Addepar applies the same permissioning and authentication model from the data tier all the way through its AI and agentic layer, ensuring access controls are never compromised as agents take on more complex workflows. Looking ahead, Pisani identifies M&A consolidation — fueled by private equity flowing into the registered investment advisor space — as the most underestimated force reshaping wealth management. From liberating alternative investment data trapped in legacy documents to powering an AI-first inversion of its entire product surface, Addepar offers a practical roadmap for how domain-specific platforms can turn proprietary data networks into a defensible competitive advantage.

In this interview from Databricks Data + AI Summit 2026, Simon Davies, senior vice president and general manager of APJ at Databricks, joins Anish Shah, chief operating officer of Jio Platforms, to talk with theCUBE's John Furrier about how a data foundation built at massive scale is accelerating the shift from fragmented enterprise data to production-ready AI. Shah details how Jio Platforms — spanning telecom, retail, media and energy for 190 million users — migrated 25 petabytes of data into Databricks in just six months, consolidating more than 12,000 tables and running over 1,500 jobs simultaneously. He underscores how the lakehouse architecture enables real-time analytics across previously siloed business units, creating cross-pollination of insights that was simply not possible before. The result, Shah notes, is end-to-end visibility now powering a broader ambition: building intelligence as a core business.

The conversation also explores Databricks' 85% year-on-year growth across Asia Pacific and the company's commitment to training 700,000 people across the region in data and AI skills. Shah reveals how Jio deployed Databricks Genie — a natural language data interface — directly to business users across supply chain, finance and network operations, generating entirely new use cases that the engineering team had never anticipated. Davies adds that Databricks is expanding into the Jio Azure region in India to address data sovereignty requirements and support the public sector, pointing to organizations such as National Australia Bank as examples pushing the boundaries of the platform's emerging security capabilities. From deploying autonomous services across telecom, retail and financial verticals to assembling a full-stack intelligence layer spanning compute, spectrum and agentic frameworks, Shah outlines why a solid data foundation is the essential prerequisite for any enterprise AI transformation.

In this interview from the Databricks Data + AI Summit 2026, Anand Pradhan, vice president and head of the AI center of excellence and mortgage data at Intercontinental Exchange, joins theCUBE's John Furrier to discuss how the company is building a governed, production-ready AI stack across on-premises infrastructure and cloud. Pradhan details ICE's hybrid architecture — combining NVIDIA hardware, VAST Data and Databricks Unity Catalog — to process millions of pages of unstructured data across mortgage servicing, real estate, fixed income and multilingual pipelines. He explains why a hyper-optimized data layer is the prerequisite for deploying agentic AI at the application layer, including AI chatbots, voice agents and chat tools built directly into MSP, the company's flagship mortgage servicing platform.

The conversation also explores how data governance anchors ICE's entire AI strategy — from catalog management and access control to benchmarking standards and audit trails. Pradhan describes the shift from systems of intelligence to systems of agency, where every agent action and conversation must be recorded and analyzed at petabyte scale, with VAST Data and Qumulo handling storage at that depth. He highlights Databricks' open source philosophy — rooted in Apache Spark and extended through open table formats and Unity Catalog — as a core reason ICE standardized on the platform, enabling vendor independence across hybrid deployments. From separating genuine AI momentum from hype to the discipline required to elevate a prototype into an enterprise-grade production system, Pradhan provides a practitioner's roadmap for how financial services organizations can meet the demands of agentic AI without sacrificing accuracy, security or regulatory compliance.