2025 Data + AI Summit: Plotly Recap and Reflections

The Moscone halls were full again. Databricks reported 22,000 in-person attendees, and over 65,000 more tuned in via livestream. With 700+ technical sessions, keynotes, and product announcements, this year’s Data + AI Summit (DAIS) reinforced that the modern data stack is evolving quickly.

For Plotly, DAIS was a clear signal that the market is shifting toward agentic analytics and more integrated development experiences. We saw leaders talk about lakehouse governance and AI agents, and we ourselves showcased how Plotly’s AI-native approach extends these trends.

While Databricks introduced more data infrastructure Lakebase, Lakeflow, and Agent Bricks, Plotly focused on how organizations can solve the last-mile gap between the data science lab and actionable operations with interactive, production-grade data apps. Dash Enterprise, and the newly announced Plotly Studio, give teams the tools to build agentic analytics systems around these layers without waiting on IT or struggling with BI constraints.

Plotly on the ground

We showed up in full force this year as a Select ISV partner and an industry leader, with over one billion downloads of Plotly open source, focused on shaping the future of Python-based data applications. My one key takeaway was the enormous appetite for agentic, shareable, end-to-end analytics apps amongst the data folks we spoke to.

Plotly’s booth stayed active all week. We gave live demos, showed off Dash, and chatted through awe-inspiring feats from our customers. We showed how Plotly Studio generates full-fledged data apps from datasets in minutes, and how Plotly Cloud enables easy deployment and authentication.

Our San Francisco community meetup brought together passionate Plotly users from across industries. Once again, Plotly Studio was the star. Several attendees said they were ready to use Studio immediately.

Through our Agentic Data Challenge, we offered a sneak preview of Plotly Studio and Plotly Cloud. Challenge participants explored how easily Studio’s expertly-trained AI generates apps with no code (unless you want to). You can sign up for early access here.

Hundreds stopped by to meet with us and explore how Dash Enterprise integrates with Databricks environments, from smart meter data to live trading dashboards.

Speaker sessions that featured Plotly

Plotly featured on stage and behind the scenes. These sessions by expert speakers highlighted the power of agentic and real-time analytics, enabled by Plotly:

• Optimizing Smart Meter IIoT Data in Databricks for At Scale Interactive Electrical Load Analytics – Dave Gibbon, Plotly
• Real-Time Market Insights — Powering Optiver’s Live Trading Dashboards with Databricks Apps and Dash – Huy Nguyen, Optiver
• Building Real-Time Trading Dashboards with DLT and Databricks Apps – Matt Slack and Matthew Moorcroft, Databricks
• Using Databricks to Power News Sentiment, a Capital IQ Pro Application – Debbie Connolly, S&P Global

Seeing Plotly power decision-makers gives us the steam to keep building. Impressive achievements across these talks and more included streaming telemetry, financial market analysis, and operational forecasts, and real-time inpatient anomaly detection. Dash apps are how teams put models to work.

Keynotes at a glance

Each year, keynotes give a glimpse of where data and AI is heading, and what affects the careers of data professionals most. This year, five major themes stood out for us:

Agents as brick layers

With Agent Bricks, MLflow 3, and serverless GPU compute, the focus shifted to agents that orchestrate entire workflows. The emerging ability to continuously evaluate and improve the accuracy of agents is exciting. In this first release, Agent Bricks is a code-first agent framework similar to LangGraph.

Today, Dash Enterprise enables data teams to incorporate chat interfaces into their data app for their end users powered by any LLM. This is provided by Chatbot Builder which is included in Dash Enterprise. The combination of Agent Bricks and Dash Enterprise Chat Builder is powerful. Data apps built in Dash Enterprise have no limitation on the number of end users and no end user licensing requirement. This means you can support as many end users as you need through Dash Enterprise for functionality built on Agent Bricks.

With Dash Enterprise, end users of your data app (with its embedded chat UI) do not need Databricks accounts as they do in Databricks Apps and AI/BI Dashboards. So, your data app can grow freely!   

Governance gets integrated

Unity Catalog appeared in nearly every keynote and is now integrated across the Databricks stack. It serves as the metadata catalog for lineage, access controls, and policy enforcement. This integration is helpful for data teams with a large existing investment in data lakes built in Databricks. The Plotly team is looking into integration with Unity Catalog for Dash Enterprise and is interested in your feedback.

Speaking of data lakes and governance, they’ve gotten complicated with lots of moving parts. If you’ve been wishing for a simpler data lake architecture that reduces the complexity of managing storage file formats, table formats and metadata catalogs, it has arrived.

DuckLake was recently introduced by DuckDB Labs. By unifying an open table format and open catalog management into a simple relational database, the complexity of querying and managing many metafiles in JSON and Avro in blob storage is eliminated and replaced with familiar SQL tables while the underlying data files can remain in a familiar and open file format, Parquet. This simplification reduces the number of operations, obviates the need for proactive pruning and results in a speed improvement. DuckLake provides the flexibility to choose among several popular relational databases as the integrated table-catalog layer, such as PostgreSQL, MySQL, SQLite, DuckDB and MotherDuck. As Plotly Dash Enterprise embeds DuckDB today, we’re excited about this new open and simplified approach to building data lakes.   

Come together, OLTP and OLAP

I like to call this current moment in the evolution of the disaggregated data stack the “separate-but-sythesized” era. Single-box Postgres instances are all you need for most OLTP workloads, especially ephemeral agentic workloads. For batch analytics workloads (OLAP), there are larger working sets and computations that require columnar storage, separation of storage and compute, plus a collection of other techniques. Built on Neon’s innovative work on storage-compute separation and branching, Lakebase integrates the storage layer of Neon Postgres with Databricks SQL. This simplifies the data synchronization among transactional and analytical systems providing the next incremental iteration of separate-but-synthesized DBMS systems. With Dash Enterprise’s enterprise connector to Databricks SQL, we’re excited to see what new agent data is available for your Dash Enterprise apps.  

Pipeline building gets more accessible

Lakeflow’s visual pipeline builder and GenAI assistant aim to bring ETL to a wider audience. The acquisition of Arcion in 2023 along with Spark are the foundational technology underpinnings. Lakeflow has matured into a traditional data integration layer for batch ETL, streaming, data orchestration, and job scheduling to assist data engineering teams working within Databricks environments. 

Apps again

After launching in 2023, Lakehouse Apps was renamed to Databricks Apps last Fall and became generally available this Summer. Built in part on Plotly open source, Databricks Apps is a way to build a simple app from data in Databricks SQL; just mind the resource limits for memory, CPU, and file size. Don’t forget that users of your app will require Databricks accounts. 

While this may fit some app scenarios, Plotly Dash Enterprise enables flexible, production-grade apps with proven results from Fortune 100 customers. Dash Enterprise places no limits or license requirements on end-user access. This helps your data app get broad adoption internally and/or externally without the economic anxiety when your app goes viral. 

You’ll also get more mileage on computing resources with Dash Enterprise since you can deploy your app anywhere (AWS, Azure, GCP, or your own data center) and self-deploy your way using virtual machines, Kubernetes, or through Plotly’s Dash Enterprise Managed Service and SaaS (coming soon with Plotly Cloud).

The benefits include flexible management, deployment, cost, and more. The future of data apps is AI-accelerated. If you follow Plotly closely, you’ll recognize that we’re revolutionizing the fundamental approach to creating open, custom data apps. We’ve combined our expressive and flexible Python data app and visualization libraries, the world knowledge of LLMs and our 10+ years of Python data app expertise to enable agentic data app development.

Now, you can go from a dataset to a functioning data app in minutes. We’ve gone beyond the generic chat interface to a structured, specification file in natural language for modifying data apps and exposing the full set of code to keep apps open and extensible. This new capability is enabled by Plotly Studio and you can discover more in our announcement blog by Chris.    

These trends reflect what we’re hearing from our own users. Teams want to move faster, explore data more freely, and build apps without barriers. Plotly’s agentic analytics efforts support this need from data teams

The broader signal from the keynotes: the demand for governed, AI-augmented, and interactive analytics is growing. That’s exactly the space Plotly is building for.

Market and analyst signals

• Databricks reported 11× growth in production models and a 377% spike in vector database usage. These systems need to solve the last-map gap with effective user interfaces for humans working with AI and ML to achieve the best business outcomes. Plotly Dash Enterprise delivers custom data apps with the right collaboration, customizability and control on any data platform, in any cloud or on-premises environment and in your favorite deployment form factors of virtual machines and Kubernetes. Plotly Cloud (in early access) extends this with a SaaS solution. 
• 81% of teams are exploring agents and LLM-enhanced analytics. 52% say knowledge gaps slow them down. Plotly Studio (in early access) was built to close that gap.
• From Gartner, 2025 – Natural-language querying and AI-generated narratives are now must-haves. Plotly Studio allows you to move beyond the chat interface for natural language. It leverages autonomous data prep, data viz, and data app agents under the hood to allow you to go from dataset to data app in minutes without prompt and open-ended chat conversations. To iterate from there, the user interface is well-structured specification files in plain English. No YAML hell. No chat doom loops. 

Taken together, these data points underscore what we saw at Data + AI Summit: the shift from passive consumption of data to apps and actions is accelerating with the onset of AI, and teams need tools built for that future.

Why this matters for Plotly

Plotly is already the go-to stack for Python data apps in production. But with Plotly Studio and Plotly Cloud, we’re enabling a faster, smarter, and more user-directed way to build. Anyone can generate full Dash apps from their datasets, customize them in a visual editor, and publish them with one click. Every idea can now take flight, with its own dedicated app.

Plotly Studio automates structure, layout, and logic based on patterns we’ve seen across industries. Plotly Cloud simplifies deployment and sharing. Together, they form the heart of our agentic analytics platform where humans, data, and intelligent agents collaborate inside decision-ready apps.

What’s next

Dashboards are losing ground to flexible, AI-enabled data applications. With the leverage the agentic era brings, we envision millions of purpose-built data apps across industries that empower every data specialist to improve and accelerate outcomes for clinical trial operations, predictive maintenance, anomaly detection, trading analytics, financial risk management and hundreds of other use cases. Plotly is committed to being the platform that powers that shift.

Want to see it for yourself? We're offering an exclusive introduction to Plotly Studio: the AI-native platform that generates professional data applications in minutes from your datasets alone. Register for the webinar and let your data vibe.

See you at Data + AI Summit 2026.