In the latest episode of Breaking Analysis, theCUBE Research’s Dave Vellante examines AMD’s Helios platform and the engineering velocity shaping artificial intelligence infrastructure. Vellante questions whether AMD’s distributed development model can compete with NVIDIA’s tightly integrated approach as software, hardware, partners and validation capacity increasingly determine long-term platform leadership.Drawing on SemiAnalysis findings, Vellante explains why testing coverage, model support, production feedback and improvement rates matter more than keynote benchmarks. AMD delivered an 18-fold interactivity gain in 30 days, yet cluster instability disrupted a key vLLM target and exposed a significant internal GPU capacity gap compared with NVIDIA.

AMD’s AI ambitions take center stage in the latest Breaking Analysis, as theCUBE Research’s Dave Vellante unpacks the company’s shift from processor comeback story to full-stack platform contender. Vellante traces how EPYC, Instinct, ROCm, Helios and strategic acquisitions now support a broader push into enterprise AI infrastructure.The episode examines why AMD does not need to unseat NVIDIA to succeed. Vellante frames its strategy around building the core, buying the gaps and ceding the ecosystem while weighing ROCm maturity, supply constraints and execution. He also explains how openness and heterogeneous computing could make AMD an indispensable second platform.

In the latest Breaking Analysis, the AGI hype gets a reality check. TheCUBE Research’s Dave Vellante and George Gilbert argue that the bigger enterprise opportunity is not chasing general intelligence, but building company-specific intelligence rooted in proprietary data, workflows, rules and trusted business context.The episode connects Databricks’ Data and AI Summit announcements to a bigger enterprise AI shift. Vellante and Gilbert unpack Genie Ontology, Unity Catalog, Omnigen and AgentBricks as building blocks for trusted role-based agents, showing how governed systems of intelligence can turn data investments into measurable business outcomes.

AI agents are getting plenty of buzz, but the real battle is over who builds the software stack behind them. On Breaking Analysis, Dave Vellante and George Gilbert talk about why agentic clients need a governed backend system of intelligence to turn semantic views, skills, artifacts and user traces into continuous learning.Vellante and Gilbert connect Snowflake Summit, Microsoft Build and Databricks’ Data and AI Summit through one big question: who owns the enterprise AI control plane? They argue winners pair agentic front ends with intelligent backends, using governance, reasoning traces, evaluations and composable skills to recommend, execute and optimize work.

In the latest Breaking Analysis episode, Dave Vellante and George Gilbert dig into how personal agents are reshaping enterprise AI strategy. Vellante and Gilbert examine the rise of a new system of intelligence, where data platforms, AI software stacks and large language model applications begin to transform workflows.Gilbert frames the discussion around a five-layer model for enterprise intelligence, from analytic and operational data to observability, context graphs and governed business logic. Vellante and Gilbert also tackle the tension between bottom-up agent adoption and top-down governance as enterprises pursue productivity without recreating disconnected silos.

The latest Breaking Analysis episode features theCUBE Research’s Dave Vellante and George Gilbert examining how the AI software stack is taking shape. Vellante and Gilbert connect systems of intelligence, digital twins, AI factories and agent observability to a bigger enterprise shift: reducing friction across operations while enabling more automated business models.Gilbert frames business rules as strategic assets, with live digital twins linking applications, analytics and agent-driven workflows. The discussion explores how observability, governance and reinforcement learning loops help enterprises manage probabilistic systems with deterministic controls. It also points to new platform economics, token-based consumption and monetization models for AI.

Enterprise computing is being rebuilt around the rack, and in the latest Breaking Analysis episode, theCUBE Research’s Dave Vellante and David Floyer dig into why. Vellante and Floyer examine Nvidia’s emerging AI factory model, DPUs, KV caches, semantic databases and the storage and networking changes pushing infrastructure into a new phase.The discussion centers on Floyer’s view that Nvidia is assembling a full-stack enterprise platform, making the rack the new unit of compute. Vellante and Floyer connect frontier-model growth to rising token demand, agentic workflows, semantic recovery, migration strategy, governance, interoperability and the cost, security and operational choices enterprises now face.

In the latest episode of Breaking Analysis, theCUBE Research’s Dave Vellante and George Gilbert take a look at Google Cloud Next 2026, where the emphasis is less on sudden change and more on steady execution. The conversation zeroes in on Google’s agent platform, the Gemini ecosystem and TPU v8 capacity, framing how these pieces connect to a broader system of intelligence for enterprise AI execution.Gilbert outlines why Google’s vertically integrated stack — from silicon through data and applications — creates a credible path to agent-driven automation. Vellante and Gilbert focus on governance, evaluation loops and incremental migration, showing how enterprises can move from fragmented data systems toward coordinated, agent-enabled service architectures that deliver dependable, business-level outcomes.

Dave Vellante of SiliconANGLE Media, Inc. provides a data-driven analysis of Google's artificial intelligence, AI-led cloud strategy drawing on theCUBE Research and Enterprise Technology Research, ETR data. Vellante examines Google’s full-stack advantages, TPU and GPU economics, BigQuery and transactional integration and the transition from human-scale analytics to agentic always-on execution; they contextualize market momentum, capital expenditure trends and technical implications ahead of Google Cloud Next.Vellante highlights Google’s advantage from integrated stack control and its TPU/GPU strategy and explains how this approach optimizes cost per token and enables real-time execution at scale. Analysts emphasize strong Google Cloud growth and rising AI and machine learning adoption in customer spending while urging attention to frontier model roadmaps, infrastructure economics, agentic data cloud capabilities, data-platform differentiation and partner ecosystem expansion.