Dan Rapp

We are excited to be among the select companies invited to the Southern California Edison (SCE) Grid Digitization Summit. From day one, we have been laying the foundation for a truly horizontal infrastructure company at the forefront of embedding controlled AI in some of the most critical paths. An opportunity to collaborate on America’s grid resilience, turning data and models into actionable frameworks that empower field and technical teams to deliver reliable, clean energy, is paramount. Our foundational platform, Lemma, is designed to cut through the complexity of legacy systems, transitioning infrastructure from static, disjointed assets to dynamic, autonomous systems. We believe that securing our nation's future requires a new and aggressive but very intentional technological push, one that demands we build secure, AI-native physical infrastructure from the ground up. #CriticalAI #ControlledAI #AIinfrastructure  #TechforCleanEnergy

75

3 Comments

The Energy Department’s Pacific Northwest National Laboratory is pushing forward on its project to use better data and artificial intelligence to speed up infrastructure permitting. Last week, the team released an updated dataset of over 120,000 permitting documents, and they’re continuing to test AI tools to help users make sense of this data.

8

Rimba turned 6 months of hidden data into real-time insights for a Renewable Natural Gas (RNG) facility. The results? Visibility on the right data and faster decisions. At an RNG facility, the operations team was stuck. This was because their SCADA system was clunky, their tools were inefficient, and their data was siloed. Every decision was manual, slow, and reactive. Tracking KPIs like uptime required endless spreadsheets. Troubleshooting anomalies meant digging through limited data points. And the interfaces were far from intuitive; only the most technical team members could navigate them. The result was missed patterns, delayed decisions, and unnecessary downtime. That’s when Rimba’s analytics tool stepped in. Here’s how we turned things around: - Unlocked data access: The team gained access to six months of historical data, not just fragments. Suddenly, they could correlate past anomalies with current trends and benchmark performance across multiple RNG plants. -Simplified analysis: Instead of relying on SQL experts or Power BI, every team member, from plant operators to the CEO, could query data directly. No bottlenecks, no delays. -Predictive maintenance: With historical data at their fingertips, the team identified subtle declines in key parameters. This allowed them to schedule maintenance proactively, preventing breakdowns before they happened. - Custom dashboards & alerts: Tailored dashboards visualised the metrics that mattered most. Intelligent alerts flagged issues in real time, enabling faster responses and smoother operations. -Process optimisation: By analysing larger datasets, hidden inefficiencies came to light. Small tweaks in the process led to big wins: increased output and lower operating costs. The transformation was undeniable. What was once hidden in the noise became actionable insights. The team went from reactive firefighting to proactive decision-making. Fewer breakdowns. Faster decisions. More confidence. If your team is stuck in the same cycle, drowning in data but starving for insights, maybe it’s time to rethink the system. What would your operations look like with all the right data at your fingertips?

36

2 Comments

Resonance has joined a $25 million initiative led by Roadrunner Venture Studios to build a quantum campus and venture studio in Albuquerque’s Innovation District, aiming to accelerate quantum commercialization in New Mexico. The coalition brings together national labs, the University of New Mexico, startups such as QuEra Computing Inc., Maybell Quantum, and Qunnect, and investors including DCVC, Playground Global, and Quantonation to launch companies, attract capital, and strengthen the region’s leadership in quantum technology. Planned infrastructure includes a multi-node quantum network, testbeds, prototyping labs, and a dedicated venture studio, with measurable goals through 2026 such as creating startups, deploying private capital, and increasing institutional collaboration. https://lnkd.in/eZwwJcbX

11

The real energy revolution isn’t coming from Washington. It’s happening locally. We recently interviewed Karin L Burns, CEO of San Diego Community Power, on the #RenewableRides podcast. In only three years, SDCP has grown from a dozen people to 85 staff, $1B in revenue, and nearly a million customer accounts, making it California’s second-largest Community Choice Aggregator. CCAs like SDCP put power procurement in the hands of communities, not utilities, delivering cleaner, cheaper, locally-produced energy. And their strategy goes further: building a 150 MW virtual power plant by connecting batteries to “stranded” rooftop solar, coordinating thousands of smaller systems to act as one. This is the grid of the future: decentralized, intelligent, resilient. For businesses, the takeaway is clear: while federal policy debates drag on, those deploying onsite energy today are cutting costs, strengthening resilience, and easing pressure on the broader grid. The revolution is local and its onsite. The technology and capital is available today and it is happening now for those that want profitable, sustainable and secure energy now and for decades to come. Dan Roberts VECKTA

24

3 Comments

Went deep with Heatmap News about the latest tariff fun times -- it's super important to realize that the US manufacturing ecosystem has atrophied enough that medium+ volume, non-defense startups are presently wise to start offshore until something changes. Have heard horror stories about cost structures simply not working unless you have sufficient volume to fully vertically integrate -- and that's before we increased tariffs on manufacturing inputs.

70

6 Comments

The Brookings Institution has identified five distinct AI clusters in US metros: 🌟 Superstars 🌟 Star Hubs 🌟 Emerging Centers 🌟 Focused Movers 🌟 Nascent Adopters Standing out as an AI Star Hub, the Phoenix-Mesa-Chandler metro area ranks 17th nationally. 🚀 Anticipating positive spillover effects from TSMC's expansion in Arizona, benefiting both the semiconductor ecosystem and AI sector. #AI #Innovation #aidatacenters #semiconductors State of Arizona City of Chandler, Arizona City of Phoenix City of Mesa City of Tempe

10

When I worked in HPC we used to describe computational capacity in FLOPS. More recently that has shifted to GigaWatts. And now people talk about it in $AUM. I find this shifting vocabulary supremely interesting and quite telling of the underlying drivers, scarcities and business models

58

6 Comments

Rising power bills are often blamed on AI companies, but PJM's forecast methodology is playing a major role too.   The ISO's load forecast still calls for roughly 11 GW of new capacity annually after 2030 to maintain reliability. Enverus Intelligence® Research sees a more moderate path, closer to 6 GW. That gap highlights the uncertainty around how much new demand will actually materialize and how much new generation is truly required.   In response, PJM is proposing to extend its capacity price collar through 2030, keeping the range between $175/MW and $325/MW to maintain strong investment signals for new generation. The mechanism was originally scheduled to expire in 2028 but would remain in place as the ISO continues to plan for strong load growth. PJM is also introducing an Expedited Interconnect Track for shovel-ready projects, allowing up to 10 projects of at least 250 MW each to move through a faster pathway before the program sunsets in 2027.   These measures may help in the near term, but they do not resolve the structural challenge of aligning supply with uncertain demand growth. Until forecasts converge with realized load and substantial new generation comes online, temporary guardrails like price collars and expedited interconnection are likely to remain part of PJM’s market design.   AI demand may receive the headlines. But market design will shape the bill. Enverus Intelligence® Research

46

As systems with single-node ephemeral engines over #S3 become mainstream (hello bauplan!), a new wave of "format aware" caching tools is inevitable. Our friends at University of Wisconsin-Madison and InfluxData just released LiquidCache, a cloud-native cache built on Apache DataFusion, a disaggregated cache based on logical data representation with incredible performances: go and add a star to Xiangpeng Hao repo now! While bauplan cache works without custom formats (it's specifically built to be native for Apache Iceberg and Apache Arrow), it shares many of the same ideas: extracting #S3 scans from queries, re-using entries on overlapping requests and more importantly, representing scans with *logical* filters, supporting relational algebra inference across languages. Our paper from last year with Ryan Curtin is available here (and perhaps some new papers are brewing as well, who knows 😎): https://lnkd.in/eCmedBgf! Congrats to our friends Xiangpeng Hao, Andrew Lamb and Remzi Arpaci-Dusseau for the incredible work: look forward to geeking out on caching the next time we all chat! #iceberg #s3 #arrow #cache

282

5 Comments

One of the best videos I’ve seen for truly understanding transformers and attention: https://lnkd.in/ggWEZDTS If you want intuition beyond the hype, this is it. Transformers work because words are represented as real-valued vectors, allowing us to use linear algebra and calculus to optimize meaning. In high-dimensional space, many vectors are nearly orthogonal, which lets the model encode an enormous number of ideas. As dimensions grow, representational capacity scales in surprisingly powerful ways. The explanation of query and key matrices is especially clear. Adjectives relate to nouns because their query vectors align with the relevant key vectors, producing larger dot products. Attention is geometry. Words issue queries, other words expose keys, and alignment determines relevance. It also covers causal self attention and why context windows scale quadratically, while showing how clever key value caching reduces redundancy at inference time. The value matrix is what actually carries the information forward once attention weights are computed. What I appreciate most is the shift in thinking. Transformers do not process language sequentially like humans. They take in the full context and build relationships simultaneously.

18

4 Comments

What if the “perfect” grid never arrives—but the energy transition succeeds anyway? Today, companies can’t afford to wait five to eight years for utility upgrades. Manufacturing is exploding, data centers are multiplying, and the EV boom is here. The result? Grid connection delays are the new choke point for America’s industrial future. So why did DCVC back Mainspring Energy $290M Series F? Because their modular generators flip the script—delivering dispatchable, clean-capable power in months, not years. General Catalyst Tom Linebarger Genevieve Kinney Shannon Miller https://lnkd.in/ezMsK_Y4

134

13 Comments

HISTORY OF EXPANSO (Part 5): E_OVERFLOWING_PIPES They needed one data point every 10 minutes. Instead, they shipped raw sensor data every second. It worked perfectly. But it still failed. Why? They were moving 600x the necessary data over thin, unreliable cell links. The network choked, and what should have been a real-time signal became a week-old history lesson. This isn't just a tech problem. It's a business problem that shows up as concrete costs: - Failed predictions: Your failure models are useless when they're running on stale data. - Wasted resources: You're paying to move and store data that provides zero immediate value. - Stalled decisions: Your teams are missing critical market opportunities waiting for data that's already irrelevant. It’s slowing your teams down and costing your business real money. What you actually want is to process the data where it's created. Filter, aggregate, and move only the valuable signal, not the raw noise. This is what we're building at Expanso: a way to run your logic right next to your data, no matter where it is. You collapse latency and get the answers you need in minutes, not weeks. When did the speed of light stop being a textbook number and become a real bottleneck for your business? #DataEngineering #EdgeComputing #DistributedSystems #DataPipelines #Latency

8