Colin VanderSmith

🌟 California is facing new barriers on its journey toward a 100% clean energy future. Flexible, future-ready interconnection could offer the state a faster route. We’ll discuss just that at this year’s California Power Summit, where I’ll join a powerhouse panel on Flexible Interconnections and Grid Innovation moderated by Eric Little and featuring Karin L Burns, CEO of San Diego Community Power, and Matthew Langer, COO at Clean Power Alliance. We’ll unpack how smart interconnection, grid-edge tech, and real-time capacity tools like DLR can help California bring on renewables more efficiently—while boosting reliability and affordability along the way. I’m proud to share Heimdall Power’s solution and look forward to learning from the bold strategies others are deploying across the West Coast grid. ⚡ 🔗 Full agenda: https://lnkd.in/gCmHaBUw #GridInnovation #CleanEnergy #Interconnection #DLR #SmartGrid #CaliforniaEnergy #HeimdallPower #CaliforniaPowerSummit2025

101

2 Comments

Prometheus Hyperscale will power 1.2 GW hyperscale facility in Wyoming in deal with PureWest Energy. Flexible, Dispatchable Energy Model: The natural gas-based power solution is designed to provide flexible and dispatchable energy, ensuring reliable data center operations. The integration with carbon capture aims to deliver a zero-carbon footprint with full traceability. Phase One Underway: Construction of the first phase of the data center is already in progress, with facilities expected to come online by 2026. Prometheus Hyperscale plans to power its 1.2GW data center campus in Evanston, Wyoming, using natural gas supplied by PureWest Energy. To achieve a net-zero carbon footprint, the company will offset emissions by purchasing 500,000 tons of carbon removal credits annually from Frontier Carbon Solutions "This partnership represents an unprecedented opportunity to bridge the gap to a net-zero future using local resources," said Trenton Thornock , Prometheus Hyperscale founder and CEO. "Wyoming wins together when we leverage our natural resources and innovative spirit. PureWest's verified low-carbon natural gas and Frontier's advanced carbon capture and storage solutions enable us to build sustainable, next-generation data centers for the age of AI." Prometheus, formerly Wyoming Hyperscale Whitebox, relaunched in September with an expanded vision. Originally planning a 120MW campus in 2020, the company revised its goal to 1GW at relaunch. With the latest partnership, the planned capacity has increased to 1.2GW. Are you hiring for your energy infrastructure projects? DM me on LinkedIn or email tom@futuraenergygroup.com for a free 30-min consultation on how Futura Energy Group can help. #energyinfrastructure #powerinfrastructure #hyperscale #datacenters #digitalinfrastructure

324

22 Comments

📰 🔥 BREAKING: Blockware adds 50% more at Project Dorothy 2 -- Following their initial 5 megawatt (MW) deployment in April, Blockware is adding 3 MW of hosting capacity at Soluna’s Project Dorothy 2 data center in Texas, bringing their total contracted footprint at the site to 8 MW. (We Keep Pushin' even as the world is in chaos.) __ “Blockware is proud to continue building alongside Soluna, expanding from our initial 5MW deployment with an additional 3MW of capacity,” says Kentaro Masuda, Chief Operating Officer, Blockware. “The first phase has consistently delivered reliable uptime and operational efficiency, validating both the site and the partnership. This partnership with Soluna enhances the Blockware Marketplace by increasing the supply of next-generation ASICs, making profitable and efficient Bitcoin mining accessible to everyone.” $SLNH @SolunaHoldings @BlockwareTeam @MitchellHODL Anthony Power ACMA CGMA @powermining_BTC https://lnkd.in/ejtfETqE

46

Equinix and Pacific Gas and Electric Company partner to bring new hyperscale capacity online in Silicon Valley Equinix has just energized its new SV12x data center in San Jose a 20MW, 82,000 sq ft facility and the first powered under the City of San José + PG&E partnership. What stands out isn’t just another hyperscale launch, it’s the model. Guaranteed power delivery, substation expansion funded by the operator, and accelerated infrastructure timelines. This is what real public-private collaboration looks like when AI, cloud, and hyperscale demand are colliding with grid constraints. Equinix Americas President Arquelle Shaw noted the opening this week saying the collaboration with the City and PG&E matters “so much” because reliable energy access is “a community investment as much as it is a technological one.” SV12x also marks Equinix’s first xScale facility in California and a clear signal that Silicon Valley capacity expansion is back.

187

6 Comments

Exciting news to announce as Distributech 2026 kicks off. WattCarbon and Piclo are launching Accelerated Community Energy (ACE) as the first stand-alone BYOC market for utilities and communities around the country. ACE allows utilities and regulators to require data center investments into local community energy resources as part of their interconnection agreements. Piclo's DER marketplace allows utilities to set the rules for procurement - location, project types, etc. WattCarbon's Aristotle+WEATS platform provides automated M&V and a system of record for auditability and to prevent double counting. To kick things off, we've jointly published a white paper (https://lnkd.in/gU7BzYqR) that dives deeply into how it works. We're also hosting an informal happy hour Tuesday night in San Diego at the East Village Brewing Company (https://luma.com/47af5am9). Stop by and say hi.

131

8 Comments

PG&E is transforming grid planning with IFS Copperleaf. By shifting to circuit-based, bundled investments, they’ve cut planned outages by 50%, saved $18M in execution costs, and unlocked up to $3.7B in potential long-term efficiencies. It’s a smarter, risk-optimized approach to planning that reduces customer impact and maximizes value across programs. Learn how PG&E is leading the way: https://ifs.link/5Ej28S #UtilityTransformation #Copperleaf #GridPlanning #EnergyEfficiency

16

1 Comment

As part of a large scale modernization project, I had a long conversation with the director of engineering around the long term “best” architecture for multiple plants he’s overseeing. We both agree that networks are a critical component of the project as it will unlock a variety of benefits: connectivity, distributed SCADA systems, data collection, etc. What we had a debate about is the way you’d create the network at every plant. In the ideal world, you’d standardize and use the same “template” for each machine, line, etc. However, in practice it’s not as simple. There are two “main” options we agree can make sense - the first one is deploying a private network which is tied to a PLC rack; most brands of PLCs (Rockwell Automation, Siemens, Opto 22, Phoenix Contact USA, etc.) provide a way to set two network cards: one for the plant, the other for the devices | the second approach is to create a distributed network and assign VLANs which incorporate specific devices and route traffic based on which VLAN the device belongs to. There are obvious trade-offs to each scheme and I’m weighing both options as we convert from legacy control systems onto new platforms. The first dilemma is the cost of managed hardware - setting up VLANs isn’t complex, but every device is around $5,000. With a limited amount of ports, it’s common to see panels requiring multiple switches if we decide to go this route. On the flip side, most PLCs come with dual ports at nearly no extra cost and a good unmanaged switch is nearly free compared to the managed ones. The downside is obviously that routing complex layouts (Ex: multiple VLANs giving access to distributed HMIs vs control hardware) is going to be more complicated. The second dilemma is the ability to remote into these “sub-networks.” With properly routed VLANs you can have access to every single piece of hardware on the plant floor. Although you can drill down through the backplane of a PLC of the same brand, it’s rarely possible to do so with 3rd party devices. For example, the plant wouldn’t be able to remote into a load cell on a private network if it’s not reachable over a VLAN through a managed switch. I’m curious to hear from my network about similar challenges / decisions they’ve faced. Do you see the value in having a fully managed network at a much higher pricepoint, or do you prefer to have segmented private networks via PLCs? On a similar note - I’m looking for alternatives to the Rockwell hardware when it comes to larger systems; how are you handling more than the “standard” 20 ports? Have you found ways to increase the number of devices via a marginal increase in cost? (Please don’t say DLR…) I'm interested in evaluating and hearing opinions on a few brands; let me know if you're able to put me in touch with someone that can ship out test networking hardware - Moxa, Siemens, ??? Note: some details on the diagram have been ommited... It's an oversimplification of what some of the machines...

328

73 Comments

Today marks a historic day for Meta and for the future of clean energy in the US: We announced agreements with three companies for up to 6.6 GW of clean, reliable nuclear energy. Scaling advanced nuclear power is no easy feat, however the announcements we made today mark a pivotal step forward. Through agreements with Vistra Corp., Oklo Inc and TerraPower, we will extend the life of three nuclear power plants, accelerate development of new reactor technologies, and support thousands of skilled jobs. These agreements amount to one of the most significant corporate purchases of nuclear energy in America. As we look forward toward our AI ambitions and the amount of compute we will bring online over the next few years, innovation at this scale requires a robust energy pipeline. Nuclear offers a clean, reliable carbon-free solution. The agreements we announced today will help create a more reliable electric grid and power American communities and our industry-leading data centers - all while creating American jobs and advancing our country's leadership in AI. I'm proud of the nuclear RFP process that our team ran over the past year, soliciting submissions from developers across the country in an effort to bring this energy online. We look forward to working closely with Vistra, Oklo, and Terrapower to advance these projects and bring more nuclear energy onto the grid. Congratulations to Meta's Energy team and many cross-functional partners for leading these milestone projects. https://lnkd.in/didZ9FZm

903

19 Comments

Tomorrow’s power grid will be defined by digital innovation. A new IBM study outlines how utilities can balance demand, manage renewable intermittency, and leverage digital tools for a more secure and efficient grid. Learn more about the future of energy. https://ibm.biz/BdnBPZ

43

Excited to share that Amazon Web Services (AWS) opened Project Rainier, our new $11B data center project in Indiana. This facility features our custom-designed AWS #Trainium chips, purpose-built to help enterprises optimize their AI/ML training workloads and accelerate innovation.   Organizations can now access infrastructure powered by Trainium to: 🔹 Train foundation models more cost-effectively 🔹 Scale ML workloads efficiently 🔹 Process complex datasets faster 🔹 Deploy AI solutions with improved latency 🔹 Scale AI workloads based on business needs   The combination of Trainium chips and our expanded infrastructure helps organizations drive their AI initiatives forward securely and responsibly, while maintaining control of their computing resources and data. Read more here: https://lnkd.in/g4Sxt327   #AWS #EnterpriseAI #CloudComputing #Innovation #AIInfrastructure #DigitalTransformation 

353

1 Comment

PG&E reports a 40% increase in service requests from data center developers. Much of this is driven by AI and cloud computing. The challenge is that grid capacity is limited in many regions, and the average California industrial electricity rate is already nearly $0.20 per kWh.   Meeting demand without adding emissions or increasing strain on the system will require new strategies...   https://lnkd.in/dhtgaSua

15

We're excited to announce new capabilities in Google Cloud #VertexAI Training designed to simplify and accelerate large-scale model development. Key highlights from our latest update: 🔹 Flexible, Self-Healing Infrastructure: Leverage fully managed, resilient Slurm environments with automated failure detection and performance-optimized checkpointing. 🔹 Cost-Effective Scheduling: Utilize Dynamic Workload Scheduler (DWS) for fixed future reservations or flexible on-demand capacity. 🔹 Integrated Frameworks: Access optimized recipes for the full model lifecycle (including SFT and DPO) and seamless integration with NVIDIA NeMo. Whether you're fine-tuning standard models or training massive custom foundational models from scratch, these new features are built to get you to production faster. Read the full announcement to see how organizations like Salesforce and AI Singapore are already leveraging these tools to elevate their models. 👇 https://lnkd.in/gVVq3WCw

185

2 Comments

😯 Energy demand for AI is only heading in one direction! "Data centers now consume 5.7% of Washington’s total electricity production, and that share is expected to grow as AI applications expand." https://lnkd.in/eAa-EqEn

10

2 Comments