Jeff Tao

First blog post up on Robotics Simulation Infrastructure! I give a high-level overview, followed by an elementary example of better infrastructure for pose management. Link in thread below

Hannover Messe is underway, and it’s always interesting to see what everyone is building. Walking through the announcements this year, one thought keeps coming back to me. For decades, industrial software has been built around applications such as SCADA, MES, historians, dashboards, and analytics tools. Each system comes with its own data model, interface, and workflow. This model has worked, but it also makes systems rigid, difficult to evolve, and expensive to adapt when requirements change. AI is starting to reshape this model. Instead of navigating predefined applications, users can express intent directly, and systems can translate that into queries, analysis, and actions. Applications themselves become easier to create, modify, and even discard when they are no longer needed. This makes one thing much more important than before: the data foundation. If applications are no longer the stable layer, then data becomes the long-term asset that everything else depends on. I shared a few thoughts on this here: In the end, applications will continue to change, but the data foundation is what persists. The real question is whether that foundation is truly designed for the AI era. tdengine.com/the-future-of-… #IndustrialAI #DigitalTransformation #DataHistorian #TDengine

@Stone_Tao Great!

my former lab will now forever have a 100% NSF GRFP success rate :) Be on the lookout from work by Arth, by far one of the best programmers/researchers I’ve had the pleasure to mentor and work with

The video of my talk “Instagram for Factory Metrics” is now available. In this presentation, I explored a simple idea: what if interacting with industrial data felt as natural as scrolling through a feed? Today, most industrial systems rely on dashboards and predefined queries. But in the AI era, we can rethink this model — moving toward systems that generate insights automatically and allow users to explore operations in a more intuitive way like Instagram or TikTok. The goal is not just better visualization, but a new way to discover, understand, and act on industrial data. A big thank you to Michael Finocchiaro for the invitation and for organizing a great event. youtu.be/oH0zM_IlNPI?si… #TDengine #Database #IndustrialAI

Day 2 at Threaded @ ACE 2026 in Miami. I had the chance to present “Instagram for Factory Metrics” — sharing how we can rethink industrial data interaction in the AI era. The feedback was very positive, which is always encouraging. More importantly, I met many new friends across PLM, simulation, and CAD. It’s great to connect with people from different parts of the engineering and industrial software world and exchange ideas. For me, this has been a very valuable conference — not just for sharing, but for learning and building new connections. Looking forward to continuing the conversations.

Made it to Miami, and today I'm at Threaded @ ACE 2026 — a startup gathering space co-located with the Aras Community Event. It's a small, intimate setup — not a massive trade show floor. Just founders, engineers, and enterprise teams sitting together in a lounge, talking honestly about where industrial and engineering software is heading. Exactly the kind of environment where real conversations happen. Big thanks to Michael Finocchiaro for pulling this together. The curation and warmth of this event says a lot about the community being built around the AI Across the Product Lifecycle podcast. Tuesday afternoon I'll be presenting "Instagram for Factory Metrics" — the idea that the future of industrial data isn't better dashboards, it's a completely different interaction model. Instead of engineers querying systems, the system finds you — surfacing what matters, when it matters, the way Instagram surfaces content you care about. Looking forward to the conversations. If you're here at ACE, come find me. #ACE2026 #IndustrialAI #TDengine #IDMP #Manufacturing #PLM

We’re excited to announce a new partnership with Arabian Digital Solutions to expand TDengine’s presence in the Middle East. As industrial enterprises in the region accelerate digital transformation, there is a growing need for modern data infrastructure that can handle real-time data, provide context, and support AI-driven insights. Through this partnership, we aim to bring TDengine’s AI-powered industrial data platform to more customers across Saudi Arabia and beyond. Arabian Digital Solutions brings strong local expertise and deep understanding of the market, which is critical for delivering real value to customers. Looking forward to what we will build together. 🚀 tdengine.com/tdengine-expan…

For years, industrial data historians have done a great job collecting and storing time-series data. But when it comes to advanced analytics, most engineers still need to step outside the system. Tools like Seeq or TrendMiner exist for a reason. They don’t duplicate data—they sit on top of historians to provide capabilities that historians were never designed to deliver. That separation made sense before. It doesn’t anymore. In the AI era, analytics shouldn’t be something external. It should be part of the data foundation itself—applied consistently across real-time data, historical data, events, and asset context. In this blog, I share why: traditional historians fall short on advanced analytics. why Python has become the standard interface for innovation and what it means to make analytics truly “native” Most importantly, I argue that native analytics doesn’t mean a closed system. Openness still matters—because AI is evolving too fast for any vendor to keep up alone. Openness ensures evolution. Native capabilities ensure usability. tdengine.com/advanced-analy… #IndustrialAI #DataHistorian #DigitalTransformation

We’re excited to share that the latest release of TDengine IDMP now supports Spanish and Korean. As we continue to work with customers and partners around the world, making the platform more accessible across different regions is an important step. Language should never be a barrier to understanding industrial data or gaining operational insights. With this update, more users can interact with IDMP in their native language — from data exploration and visualization to analytics and AI-driven insights. This is just one step in our ongoing effort to make industrial data more accessible, intuitive, and valuable globally.

Check out my latest article: AI Is Not a Nice-to-Have — It’s a Requirement linkedin.com/pulse/ai-nice-… via @LinkedIn

For decades, industrial data historians have relied on Data Archive as the core engine. It was a brilliant design for its time—optimized for storage efficiency under tight resource constraints. But the assumptions have changed. Today, industrial systems generate far more data, and the expectation is no longer just to store it, but to analyze it, integrate it, and use it for AI-driven decisions. This is where the limitations start to show: . Lossy compression . Closed interfaces . Limited scalability . Difficult integration with modern data systems Modern time-series databases (TSDB) were built for a different world—one of scale, openness, and real-time analytics. In this article, I share why the shift from Data Archive to TSDB is not just an upgrade, but a fundamental change in how we think about industrial data foundations. Would love to hear how others are approaching this transition. #IndustrialAI #TimeSeriesDatabase #IIoT #Manufacturing #DataHistorian tdengine.com/from-data-arch…

Over the past few decades, data historians have been the foundation of industrial operational systems. They solved a critical problem: how to reliably collect and store time-series data from the plant floor. But the world has changed. Today, the question is no longer just how to store data, but how to understand it, generate insights, and support decisions in real time—especially in the age of AI. In this new context, traditional approaches are starting to show their limits: . Data is often locked in closed systems . Context is lost when moving into modern platforms . AI struggles without asset and event awareness In this article, I share a perspective on where industrial data infrastructure is heading: Historian → Data Platform → AI-Native Data Foundation And why the future of industrial software may look like: Agent Interface + Data Foundation. AI becomes the interface, and the data foundation becomes the system. Would love to hear your thoughts, especially from those working at the intersection of OT, IT, and AI. Read the full article here: #IndustrialAI #IIoT #Manufacturing #DataPlatform #DigitalTransformation tdengine.com/from-data-hist…

Check out my latest article: Asset-Centric Data Modeling Is Even More Important in the AI Era for Industrial Operation linkedin.com/pulse/asset-ce… via @LinkedIn

AI is changing how software is built. Features can be generated instantly. Dashboards can be created by agents. Entire applications can appear from a prompt. But AI cannot generate the industrial data foundation behind those systems. In my latest article I share why I believe the future of industrial software will be built on AI-native data platforms, and why the combination of TSDB + IDMP is becoming the foundation layer for industrial AI. The real competition in the AI era will not be about features. It will be about who owns the data foundation. Check: Next Frontier：Building the Industrial Data Foundation for the AI Era linkedin.com/pulse/next-fro…

A milestone worth celebrating. The number of TDengine TSDB installations worldwide has just surpassed 1,000,000. When we started TDengine 9 years ago, the goal was simple: build a high-performance, scalable time-series database designed for real-world data — especially the massive streams coming from industrial systems and IoT devices. Reaching one million installations is not just a number. It reflects the trust from developers, engineers, partners, and customers around the world who chose TDengine to power their data infrastructure. I want to sincerely thank our amazing team, our open-source community, and our global partners who helped make this possible. But the journey does not stop at a time-series database. Over the past few years, we have been building something much bigger on top of TDengine: an AI-native industrial data platform. By combining high-performance TSDB with industrial data management, contextualization, analytics, visualization, and AI capabilities, we aim to turn raw operational data into real-time insights and intelligence. The goal is simple but ambitious: make industrial data accessible, affordable, and truly valuable, and allow intelligence to emerge directly from operational data. One million installations is a milestone. The real journey is just getting started. #TDengine #IndustrialAI #TimeSeriesDatabase

Check out my latest article: Event Frames: One of the Most Brilliant Ideas in Industrial Data — And Why They Matter Even More in the AI Era linkedin.com/pulse/event-fr… via @LinkedIn

Started the morning with the Peninsula BURN Running Club, logging 13 km along the Bay Trail under the crisp sky. Nothing beats the energy of running with a great group to kick off the day. After the run, we gathered in the park for a potluck to celebrate the Chinese New Year—great food, great conversations, and a strong sense of community.

Chinese New Year lasts for 15 days, and the celebration isn’t over yet. According to tradition, it’s a time for reunion — and for sharing a great meal together. Today at noon, I cooked a small Chinese New Year feast for our team here in the bay area: Beef Back Ribs, Spicy Braised Pork Trotters, Pickled Cabbage, and Long Beans. Nothing fancy — just home-style dishes meant to bring people together. What made it special wasn’t the food, but the moment: a group of colleagues from different backgrounds sitting around the table, laughing, sharing stories, and celebrating a tradition that travels far beyond China. Building a company is a lot like this holiday — it’s ultimately about people, trust, and the bonds we create along the way. Wishing everyone a joyful and prosperous Year of the Horse. #ChineseNewYear #TeamCulture #TDengine

Proud of you, Stone! Wishing you a great talk and inspiring discussions.

Join us TODAY for the GRASP SFI seminar featuring Ph.D. candidate, Stone Tao @Stone_Tao , who will be presenting “Accelerating and Driving Robotics Progress with Simulation Integrated Robot Learning” from 3PM-4PM. More info: grasp.upenn.edu/events/spring-… #GRASP #GRASPLab #GRASPSFI