Rosa Del Mar

Daily Brief

Issue 90 2026-03-31

Portable Microreactors For Off Grid Firm Power

Issue 90 Edition 2026-03-31 8 min read
General
Sources: 1 • Confidence: Medium • Updated: 2026-04-11 17:54

Key takeaways

  • Radiant is building its first nuclear reactor in a roughly 70,000-square-foot facility and added a second building after increasing vertical integration.
  • Ferrite for high-frequency transformers and thin-film power capacitors are identified as key power-electronics supply-chain vulnerabilities, and Drew Baglino says he is working to onshore or nearshore production including reviving a former U.S. ferrite facility in Georgia.
  • A future grid architecture could rely on many intelligent generators and microgrids that dynamically mesh rather than only a traditional centralized structure.
  • In the U.S. grid today, transmission and distribution capacity is a bigger bottleneck than building new generation capacity.
  • Gigawatt-scale data centers can create grid stability risk by rapidly isolating from the grid during disturbances; Drew Baglino cites a report of roughly 2 gigawatts shutting off in Virginia and a similar event in Washington State.

Sections

Portable Microreactors For Off Grid Firm Power

  • Radiant is building its first nuclear reactor in a roughly 70,000-square-foot facility and added a second building after increasing vertical integration.
  • Radiant plans to fuel reactors at an 80-acre Tennessee site, then transport fueled reactors to customer sites, and is pursuing the needed regulatory permits for that approach.
  • Radiant is building a mass-producible, portable microreactor with a current product focus on a trailer-sized reactor.
  • Radiant claims it is permitted to go to full power rather than only to criticality, distinguishing it from other new reactor efforts.
  • Radiant claims a five-year microreactor energy output is equivalent to roughly two million gallons of diesel and that the reactor can be removed when operations end, shifting waste handling away from the customer site.
  • Radiant frames its primary customer decision as choosing between diesel generators and a microreactor and claims it becomes economic when diesel is roughly $6.50 per gallon or higher.

Grid Edge Power Electronics As Enabler

  • Ferrite for high-frequency transformers and thin-film power capacitors are identified as key power-electronics supply-chain vulnerabilities, and Drew Baglino says he is working to onshore or nearshore production including reviving a former U.S. ferrite facility in Georgia.
  • Heron Link is a five-megawatt bidirectional solid-state transformer that converts 800–1500V DC to 34kV AC for distribution and sub-transmission interconnection use cases including data centers, solar, and batteries.
  • Replacing conventional oil-filled, line-frequency transformers with high-frequency switched power electronics and software can improve power density and control responsiveness for industrial-scale conversion.
  • Heron Link uses a modular, fail-operational design with thirty 165-kW modules per five-megawatt unit so it can continue operating if a module fails.
  • Heron is initially targeting a roughly 500-gigawatt DC market segment across data centers, solar, and batteries, with future products planned for AC-to-AC conversion.
  • Drew Baglino expects Heron to build about 10 prototype systems this year largely by hand and then 30–50 prototype systems in a factory location to be announced next quarter.

Architecture Shift Toward Microgrids And Dc

  • A future grid architecture could rely on many intelligent generators and microgrids that dynamically mesh rather than only a traditional centralized structure.
  • A decentralized, software-defined grid is presented as necessary because future loads are spiky and complex and it is impractical to build the grid for maximum edge-case capacity everywhere.
  • A DC microgrid architecture is argued to reduce conversion layers because compute, batteries, solar, and Radiant’s reactor output are fundamentally DC.
  • A modular, removable power installation can reduce local community opposition compared with large permanent infrastructure such as visible nuclear cooling towers.
  • Modular power installations are claimed to simplify logistics, enable quick installation, and reduce permitting complexity to deliver clean power on short timelines.

Grid Bottleneck Shifts To Delivery

  • In the U.S. grid today, transmission and distribution capacity is a bigger bottleneck than building new generation capacity.
  • From roughly the 1980s through the 2010s, broad energy-efficiency gains largely offset growth in energy services, keeping net electricity delivered roughly flat.
  • U.S. electricity demand is rising again because data centers, electrified transport, and reshoring are outpacing prior efficiency gains.

Data Centers As Grid Stability And Policy Flashpoint

  • Gigawatt-scale data centers can create grid stability risk by rapidly isolating from the grid during disturbances; Drew Baglino cites a report of roughly 2 gigawatts shutting off in Virginia and a similar event in Washington State.
  • Data center grid instability can be mitigated with modern power electronics using grid-forming controls plus modest energy storage so data centers ride through disturbances and can stabilize the grid.
  • Electricity demand growth is expected to accelerate because AI enables more applications even though compute efficiency has improved dramatically.

Watchlist

  • Ferrite for high-frequency transformers and thin-film power capacitors are identified as key power-electronics supply-chain vulnerabilities, and Drew Baglino says he is working to onshore or nearshore production including reviving a former U.S. ferrite facility in Georgia.
  • A future grid architecture could rely on many intelligent generators and microgrids that dynamically mesh rather than only a traditional centralized structure.

Unknowns

  • What empirical evidence (interconnection queues, congestion metrics, project lead times) supports transmission/distribution as the dominant bottleneck versus generation in the specific target regions discussed?
  • Will Radiant achieve criticality and full-power operation in 2026, and what are the concrete regulatory milestones required to do so?
  • What is the all-in delivered cost of electricity for Radiant’s microreactor in early deployments, and how sensitive is it to fuel logistics, security, and regulatory compliance?
  • What is the verified performance of Heron Link (efficiency, fault response, grid-code compliance, thermal limits) in pilot environments and how quickly can it be certified and accepted by utilities?
  • Are ferrite and thin-film capacitor supply constraints binding at the volumes implied by Heron’s factory scale targets, and what concrete onshoring capacity and timelines exist?

Investor overlay

Read-throughs

  • Onshoring or nearshoring of ferrite and thin-film capacitor production could become a gating factor and beneficiary for scaling grid-edge power electronics like bidirectional solid-state transformers and grid-forming gear.
  • If transmission and distribution capacity is the dominant bottleneck, demand may shift toward technologies that increase deliverability or enable microgrids to reduce dependence on constrained delivery networks.
  • If large data centers create stability risk via rapid load isolation, there may be accelerated interest in grid-forming power electronics and modest storage designed to manage fault response and ride-through.

What would confirm

  • Concrete timelines and capacity announcements for domestic or nearshore ferrite and thin-film capacitor output, plus evidence these inputs are binding at targeted factory scales for power electronics.
  • Region-specific congestion and interconnection evidence showing delivery constraints dominate, alongside procurement or utility programs prioritizing solutions that increase local deliverability or microgrid capability.
  • Pilot results for bidirectional solid-state transformer performance including efficiency, fault response, grid-code compliance, thermal limits, and utility acceptance milestones such as certification and interconnection approvals.

What would kill

  • Ferrite and thin-film capacitor constraints prove non-binding at relevant volumes, or onshoring efforts stall, undermining the thesis that these inputs are key scaling bottlenecks or catalysts.
  • Empirical grid data shows generation capacity, not transmission and distribution, is the primary bottleneck in target regions, reducing urgency for microgrids and deliverability-focused solutions.
  • Repeated failures or delays in certifying and accepting grid-edge power electronics, or inability to demonstrate stable fault response and compliance in pilots, limiting adoption for data center stability mitigation.

Sources