Data Analytics News: Power BI Surges With July 2026 Features

Microsoft’s Power BI monthly release, a new AI governance alarm from Arctera, fresh deployment controls from DataRobot, and an eye-catching internal analytics platform from Cloudflare all land in today’s data analytics news. It’s a week heavy on tooling updates and sobering enterprise AI reality checks — exactly the kind of mix that separates teams keeping pace with the field from those falling quietly behind.

Power BI Surges With DAX User-Defined Functions and Agent Skills in July 2026 Update

Microsoft dropped its Power BI July 2026 Feature Summary this week, and the headliners are production-ready DAX user-defined functions (UDFs) and expanded Agent Skills for Power BI and Fabric Apps. DAX UDFs let semantic model authors encapsulate reusable logic in named functions — a long-requested capability that makes complex models dramatically easier to maintain. Agent Skills, meanwhile, let developers prompt an AI agent to build and refine semantic models and reports end to end: describing the goal in plain language, letting the agent build the semantic model from Fabric data, write the DAX, and generate report pages. Other additions include Model Options now accessible in the Power BI Service web interface (previously Desktop-only), DAX measure descriptions via triple-slash (///) inline syntax, and expanded modern visual defaults across reporting. For the data analyst community, this is one of the most significant data analytics news releases Microsoft has shipped since Build 2026.

Arctera: 78% of Organizations Expect AI Governance Risk to Rise, But Only 19% Can Prove It

The Arctera State of AI Governance 2026 report, released July 21, delivers a sharp warning for analytics and data teams: while 78% of organizations using AI expect communications risk to rise over the next 12 to 24 months, fewer than one in five (19%) have the logging, retention, detection, and scoring controls needed to prove what their AI actually did. More than half (55%) have core AI policies and training in place — but policy without an evidence layer is unenforceable. Based on surveys of 500 compliance decision-makers across finance, healthcare, and energy in the Americas and EMEA, the report underscores a governance gap that will become increasingly costly as AI regulation tightens globally through late 2026 and into 2027.

DataRobot Delivers Full Deployment Flexibility and AI Governance Controls

DataRobot launched updates this week giving enterprise data teams granular control over where AI models run and how they are governed. The new capabilities allow organizations to choose between cloud, on-premises, and regulated-environment deployments, with governance frameworks that travel with the model regardless of infrastructure. DataRobot is explicitly targeting financial services, healthcare, and government customers facing compliance constraints that rule out a single-cloud strategy. For data teams building pipelines to feed those models, getting query logic right before deployment is foundational — a principle explored in our SQL for Data Analysis guide to sorting results and finding unique values, which covers how clean, structured outputs prevent downstream model issues.

Cloudflare Details Town Lake Lakehouse and Skipper AI Analytics Agent

In a detailed engineering post that has been making waves in the data infrastructure community, Cloudflare revealed how it built Town Lake, an internal unified data platform, and Skipper, an AI analytics agent that runs on top of it. Billing workloads account for 53% of all platform queries, with Town Lake built on a lakehouse architecture using Trino, Apache Iceberg, Cloudflare R2 for storage, and DataHub for metadata governance. Skipper is the natural-language layer — allowing any Cloudflare employee to ask questions in plain English and receive correct, auditable SQL-backed answers in seconds. The architecture is notable for being built entirely on Cloudflare’s own infrastructure stack rather than third-party cloud data warehouses, making it a useful blueprint for engineering teams considering self-hosted analytics stacks. For practitioners building toward similar stacks, our Machine Learning with Python guide covers the modeling and inference concepts that underpin AI-powered analytics agents like Skipper.

DataCamp and Google Cloud Launch Joint Learning Collaboration

DataCamp announced a new collaboration with Google Cloud aimed at helping data practitioners build practical cloud and AI skills through role-based learning paths. The program extends DataCamp’s curriculum to cover application modernization, cloud security, and AI and data engineering on Google Cloud, reflecting employer demand for professionals who can work across cloud-native data stacks rather than just local or single-vendor environments. The partnership adds Google Cloud to a DataCamp roster that already includes Microsoft, AWS, and Databricks — making DataCamp’s training one of the more comprehensive multi-cloud skill ecosystems in the market.

Simplilearn and Duke University Partner on 20-Week Agentic AI Program

Simplilearn and Duke University have launched a 20-week agentic AI program designed for data analysts, engineers, and scientists who want structured, hands-on training in generative and agentic AI. The program features live online instruction, hands-on projects, and a dedicated Agentic AI workshop. The Simplilearn-Duke collaboration is notable for bringing academic rigor to what has largely been an informal, self-directed upskilling market — signaling that agentic AI competency is moving from differentiator to expected baseline for senior data roles.

From Power BI’s agentic analytics push to Cloudflare’s open-source lakehouse blueprint, this week’s data analytics news reflects a field where AI is no longer optional infrastructure — it’s the new baseline. Teams that treat tooling updates and governance surveys as background noise risk finding themselves several releases behind when it matters most.