SMR Race to Power Data Centers: Three Technologies, One Market
Three SMR technologies - lead-cooled (Blykalla), TRISO (X-energy), and sodium-cooled (Aalo) - are competing to power hyperscaler AI data centers. Meta's 6.6 GW nuclear commitment signals the nuclear race has begun.
TL;DR
Three small modular reactor (SMR) technologies - lead-cooled (Blykalla), TRISO-based (X-energy), and sodium-cooled (Aalo) - are racing to power hyperscaler AI data centers. Aalo targets 2029 commercial deployment, giving it first-mover advantage. Meta’s 6.6 GW nuclear commitment represents the largest corporate nuclear purchase in US history, signaling that the SMR market for data centers has moved from theoretical to urgent.
Executive Summary
The convergence of AI power demand and SMR technology readiness has created a competitive market that did not exist five years ago. Three technologies have emerged as frontrunners: Blykalla’s lead-cooled SEALER, X-energy’s TRISO-fueled XE-100, and Aalo’s sodium-cooled microreactor. Each offers distinct trade-offs in safety profile, scalability, and deployment timeline.
The stakes are substantial. Hyperscalers (Meta, Amazon, Google, Microsoft) have committed to nuclear power at unprecedented scale:
| Company | Nuclear Commitment | Timeline | Primary Technology |
|---|---|---|---|
| Meta | 6.6 GW | By 2035 | TerraPower Natrium, Oklo Aurora, Vistra PPA |
| Amazon | 5+ GW | By 2039 | X-energy XE-100, Talen partnership |
| 600+ MW (Duane Arnold) + SMR | 2029+ | Kairos Power, NextEra | |
| Microsoft | Three Mile Island restart + SMR | 2028+ | Constellation, Helion, Aalo |
This analysis compares the three SMR technologies competing for this market, evaluates their commercialization timelines, and identifies the decision factors that will determine market leadership.
Key Facts
- Who: Blykalla (Sweden), X-energy (US), Aalo Atomics (US) - three SMR vendors with distinct cooling technologies
- What: Competition to supply nuclear power to hyperscaler AI data centers demanding 1+ GW per campus
- When: Aalo targets 2029, X-energy and Blykalla target 2030s; DOE deadline July 4, 2026 for test reactors
- Impact: Meta’s 6.6 GW commitment is the largest corporate nuclear purchase in US history; total hyperscaler demand exceeds 12 GW by 2040
Background & Context
The AI Power Problem
AI data centers have fundamentally different power requirements than traditional computing facilities. Training large language models requires sustained, high-power computation for weeks or months. Inference workloads demand instant response with minimal latency. Both require 24/7 power that renewable sources cannot reliably provide.
“The AI data center power demand has created an energy procurement challenge that nuclear is uniquely positioned to solve.” — World Nuclear News, Meta Nuclear Agreement Analysis, March 2026
A single hyperscaler campus now demands 1+ GW of continuous power. Meta’s Prometheus supercluster in New Albany, Ohio required multiple GW-scale commitments. Microsoft’s Three Mile Island restart secures 835 MW for a single data center complex. These numbers exceed the output of many utility-scale power plants.
Why SMRs for Data Centers
Small modular reactors offer three advantages that align with hyperscaler needs:
- Scale matching: SMR modules can be added incrementally as data centers expand, unlike traditional 1+ GW reactors that require massive upfront commitment
- Location flexibility: Factory-built modules can be transported to sites near demand centers, reducing transmission losses
- Carbon-free baseload: Nuclear provides 24/7 carbon-free power that matches AI workload requirements
The market opportunity has attracted over a dozen SMR developers. Three have emerged with clear pathways to commercialization: Blykalla, X-energy, and Aalo.
Analysis Dimension 1: Technology Comparison
Blykalla SEALER - Lead-Cooled Reactor
Blykalla, a spin-off from KTH Royal Institute of Technology with research dating to 1996, uses lead as the primary coolant. This approach eliminates the hydrogen explosion risk inherent in water-cooled reactors, as lead remains chemically inert at operating temperatures.
Technical Specifications:
| Parameter | Value |
|---|---|
| Coolant | Lead (chemically inert) |
| Single module capacity | ~50 MWt thermal |
| Deployment configuration | 6 reactors = 300 MW total |
| Steam temperature | Higher thermal efficiency than water-cooled |
| Commercialization target | Early 2030s |
| Primary market | Industrial heat, data centers |
The lead coolant operates at atmospheric pressure, eliminating the need for high-pressure containment systems. This simplifies construction and reduces costs. The SEALER-55 commercial reactor is rated at 140 MWt thermal output.
Blykalla announced in March 2026 that it is proceeding with planning for a 6-unit SEALER plant at Norrsundet, Gavle, Sweden. The project requires approvals from the Swedish Radiation Safety Authority, Land and Environment Court, Swedish Government, and Gavle Municipality. Permitting is expected to initiate later in 2026.
Key differentiator: Lead cooling eliminates the hydrogen explosion risk that contributed to the Fukushima accident, providing a fundamentally different safety profile than water-cooled designs.
X-energy XE-100 - TRISO Fuel, Helium-Cooled
X-energy takes a different approach with TRISO (TRIstructural-ISOtropic) particle fuel. Each fuel particle is a poppy-seed-sized uranium kernel coated with multiple ceramic layers. The US Department of Energy describes TRISO as “the most robust nuclear fuel on earth” - it can withstand extreme temperatures without melting.
Technical Specifications:
| Parameter | Value |
|---|---|
| Coolant | Helium (inert gas) |
| Single module capacity | 80 MWe / 200 MWt |
| Standard deployment | 4-unit plant = 320 MWe |
| Steam temperature | 565C |
| Design life | 60 years |
| Load-following | 40-100% in 12 minutes |
| Fuel type | TRISO-X particle fuel |
| Commercialization target | 2030s |
The XE-100 uses a pebble-bed design where fuel spheres continuously circulate through the reactor core. Helium coolant remains chemically inert in single-phase operation.
X-energy has built the largest SMR pipeline in the industry: 11+ GW across US and UK projects including:
- Amazon/Energy Northwest Cascade Center (12 XE-100 modules)
- Dow/Texas industrial deployment (first ARDP project)
- Talen/PJM market deployment (960+ MW LOI signed March 2026)
- Centrica/UK advanced reactor fleet (6 GW)
Key differentiator: TRISO fuel cannot melt at any temperature the reactor can achieve, providing inherent safety without complex active systems.
Aalo Pod - Sodium-Cooled Microreactor
Aalo Atomics, founded in 2023 in Austin, Texas, has moved fastest toward commercialization. The Aalo Pod is a 50 MWe sodium-cooled fast reactor designed specifically for data center deployment.
Technical Specifications:
| Parameter | Value |
|---|---|
| Coolant | Sodium (liquid metal) |
| Single module capacity | 10 MWe per Aalo-1 reactor |
| Standard deployment | 5 reactors = 50 MWe (Aalo Pod) |
| Fuel type | Low-enriched UO2 |
| Commercialization target | 2029 (fastest timeline) |
| Primary market | Data centers (purpose-built) |
In March 2026, Aalo completed assembly of its Critical Test Reactor at Idaho National Laboratory. The company expects to achieve criticality before the July 4, 2026 DOE deadline for the Nuclear Reactor Pilot Program - a key milestone toward commercial licensing.
Aalo’s design emphasizes factory-built modularity. Each reactor is manufactured in a factory and transported to the site, enabling rapid deployment. The company explicitly targets data centers as its primary market.
Key differentiator: Aalo is the only SMR designed from the ground up for data center deployment, with 50 MW granularity matching typical data center power increments.
Technology Comparison Matrix
| Dimension | Blykalla SEALER | X-energy XE-100 | Aalo Pod |
|---|---|---|---|
| Coolant | Lead (inert) | Helium (inert gas) | Sodium (reactive metal) |
| Fuel | Not disclosed | TRISO-X particles | Low-enriched UO2 |
| Module size | ~50 MWt | 80 MWe / 200 MWt | 10 MWe |
| Deployable unit | 300 MW (6 units) | 320 MWe (4-pack) | 50 MWe (5 reactors) |
| Key safety feature | No hydrogen risk | Fuel cannot melt | Fast-neutron stability |
| Load-following | Not disclosed | 40-100% in 12 min | Not disclosed |
| Design life | Not disclosed | 60 years | Not disclosed |
| Commercialization | Early 2030s | 2030s | 2029 |
| Pipeline scale | 300 MW demo | 11+ GW | Not disclosed |
| Data center fit | Medium scale | Hyperscale | Purpose-built |
Analysis Dimension 2: Commercialization Timeline Race
The 2029 Window
Aalo’s 2029 commercial target gives it first-mover advantage. If successful, Aalo will deliver the first SMR designed specifically for data centers three to five years before competitors.
The timeline depends on DOE Nuclear Reactor Pilot Program success. Aalo must demonstrate criticality by July 4, 2026 to maintain program momentum. The completed Critical Test Reactor at INL positions the company to meet this milestone.
“Aalo expects criticality before July 4, with commercial Aalo Pod deployment targeted for 2029.” — World Nuclear News, March 2026
X-energy’s Pipeline Advantage
X-energy may trail Aalo in first deployment but leads in total pipeline commitment. The company has secured:
- ARDP funding for Dow/Texas first deployment
- Amazon partnership for 5 GW by 2039
- Talen LOI for 960+ MW in PJM market
- Centrica agreement for 6 GW UK fleet
The breadth of X-energy’s partnerships provides validation that TRISO technology meets hyperscaler requirements. Amazon’s commitment to 12 XE-100 modules at the Cascade Advanced Energy Center demonstrates confidence in the technology at scale.
Blykalla’s European Position
Blykalla represents the European SMR option, with Sweden as the lead market. The Norrsundet plant will require multi-year permitting, pushing commercial operation to the early 2030s.
European data center operators (currently relying on grid power) may find Blykalla attractive for:
- Lower regulatory barriers in Sweden compared to US NRC licensing
- Proximity to Nordic data center clusters
- Industrial heat applications beyond electricity
Critical Path Milestones
| Milestone | Blykalla | X-energy | Aalo |
|---|---|---|---|
| Test reactor criticality | Not applicable | Not applicable | July 2026 target |
| First commercial deployment | Early 2030s | ~2030 | 2029 |
| Hyperscaler partnership | None announced | Amazon (5 GW), Talen, Centrica | Microsoft (permitting AI) |
| Regulatory approval pathway | Swedish 4-agency process | NRC design certification | DOE Pilot Program + NRC |
| Manufacturing facility | Not disclosed | Under development | Factory model |
Analysis Dimension 3: Hyperscaler Requirements Analysis
What Data Centers Actually Need
Hyperscaler power requirements differ from traditional baseload nuclear customers in four dimensions:
1. Power Density
AI training clusters consume power at densities of 50-100 MW per acre. A single hyperscaler campus can require 1+ GW continuous. This exceeds the output of most single nuclear units and approaches the scale of multi-reactor plants.
| Hyperscaler Project | Power Requirement | Technology |
|---|---|---|
| Meta Prometheus (Ohio) | Multiple GW | Natrium SMR, existing PPA |
| Amazon Cascade Center | 320-960 MW | X-energy XE-100 |
| Microsoft Three Mile Island | 835 MW | Existing reactor restart |
| Google Duane Arnold | 600+ MW | Existing reactor restart |
2. Carbon-Free Requirement
All four hyperscalers (Meta, Amazon, Google, Microsoft) have publicly committed to carbon-free energy targets:
- Meta: Net zero by 2030
- Amazon: Net zero by 2040
- Google: Carbon-free energy 24/7 by 2030
- Microsoft: Carbon negative by 2030
Nuclear provides the only scalable baseload carbon-free option. Solar and wind require massive overbuild plus storage, with land requirements that many data center locations cannot accommodate.
3. Load-Following Capability
AI workloads fluctuate based on training schedules and inference demand. X-energy’s XE-100 can ramp from 40% to 100% power in 12 minutes - significantly faster than Gen III reactors that require hours for load changes.
This capability matters for data centers that may vary power consumption:
- Training runs: sustained 100% load for weeks
- Inference peaks: rapid scaling based on user demand
- Maintenance windows: reduced load during equipment changes
4. Location Flexibility
Data centers locate based on connectivity, climate, and tax incentives - not proximity to power plants. SMRs that can be transported to demand centers have a fundamental advantage over traditional nuclear.
X-energy emphasizes “road-shippable” design for the XE-100. Aalo emphasizes factory-built modularity. Both approaches enable siting near data center clusters.
Hyperscaler Partnership Landscape
The partnerships formed in 2024-2026 reveal hyperscaler technology preferences:
| Hyperscaler | SMR Technology | Capacity | Status |
|---|---|---|---|
| Meta | TerraPower Natrium | 2.8 GW (8 units) | Agreement signed |
| Meta | Oklo Aurora | 1.2 GW Ohio campus | Agreement signed |
| Meta | Vistra (existing plants) | 2,176 MW | PPA signed |
| Amazon | X-energy XE-100 | 5+ GW target | Partnership |
| Kairos Power | SMR | Agreement signed | |
| NextEra/Duane Arnold | 600+ MW | 25-year PPA | |
| Microsoft | Constellation/TMI | 835 MW | Restart agreement |
| Microsoft | Helion Fusion | Not disclosed | PPA signed |
| Microsoft | Aalo Atomics | Not disclosed | AI permitting collaboration |
Meta’s approach is notable for its breadth: existing plant restarts (Vistra), sodium-cooled SMRs (TerraPower Natrium), and fast reactors (Oklo Aurora). This diversification suggests hedging across technologies rather than committing to a single approach.
Amazon has concentrated on X-energy’s TRISO technology, with a 5 GW commitment by 2039. This represents a stronger technology bet than Meta’s portfolio approach.
Microsoft’s strategy combines near-term restarts (Three Mile Island), fusion (Helion), and advanced permitting (Aalo). The Aalo collaboration on AI-powered permitting acceleration may indicate deeper technology interest.
Google’s primary commitment is the Duane Arnold restart (600+ MW by 2029) with Kairos Power SMR agreements for longer-term power.
Analysis Dimension 4: Regulatory and Market Outlook
US Regulatory Environment
The DOE Nuclear Reactor Pilot Program has created a race to criticality. The July 4, 2026 deadline has concentrated industry attention on demonstrating advanced reactor viability. Aalo’s test reactor completion positions it well for this milestone.
The NRC design certification process remains the primary regulatory hurdle for commercial deployment. X-energy has invested heavily in pre-application engagement with NRC. Aalo benefits from the DOE Pilot Program’s streamlined pathway.
Key regulatory indicators to watch:
- NRC design certification progress for each technology
- State-level nuclear development incentives (Wyoming, Texas, Ohio)
- Federal loan guarantee availability
European Regulatory Environment
Blykalla faces a four-agency approval process in Sweden:
- Swedish Radiation Safety Authority
- Land and Environment Court
- Swedish Government
- Gavle Municipality
This multi-layered process will extend permitting timelines compared to US projects, but Sweden’s supportive nuclear policy (reversing previous phase-out plans) creates a favorable environment.
Market Size Projections
The hyperscaler nuclear market has grown from negligible to substantial in 18 months:
| Year | Estimated Hyperscaler Nuclear Demand |
|---|---|
| 2024 | < 1 GW committed |
| 2026 | ~15 GW committed (Meta 6.6, Amazon 5+, Google, Microsoft) |
| 2030 | First SMR deployments operational |
| 2035 | 20+ GW cumulative SMR capacity projected |
| 2040 | 40+ GW potential if technology proves reliable |
The transformation from near-zero to 15+ GW committed capacity in 24 months indicates that the market has crossed the adoption threshold from experimental to operational planning.
Key Data Points
| Metric | Value | Source | Date |
|---|---|---|---|
| Meta nuclear commitment | 6.6 GW | World Nuclear News | 2026-03 |
| Amazon nuclear target | 5+ GW by 2039 | World Nuclear News | 2026-03 |
| Google Duane Arnold capacity | 600+ MW | World Nuclear News | 2026-03 |
| X-energy total pipeline | 11+ GW | World Nuclear News | 2026-03 |
| DOE test reactor deadline | July 4, 2026 | World Nuclear News | 2026-03 |
| Aalo commercial target | 2029 | World Nuclear News | 2026-03 |
| XE-100 load-following | 40-100% in 12 min | X-energy Official | 2026-03 |
| TerraPower Natrium capacity | 690 MW + 500 MWh storage | World Nuclear News | 2026-03 |
| Blykalla SEALER plant | 300 MW (6 reactors) | World Nuclear News | 2026-03 |
| Aalo Pod capacity | 50 MWe (5 reactors) | World Nuclear News | 2026-03 |
🔺 Scout Intel: What Others Missed
Confidence: high | Novelty Score: 85/100
Coverage of SMR announcements treats each technology as an isolated story. The strategic pattern is missed: hyperscalers are not merely buying power - they are locking in manufacturing capacity for SMR technologies that have not yet proven commercial viability.
Meta’s 6.6 GW commitment spreads risk across three technology bets (Natrium, Oklo, Vistra PPA). Amazon’s concentrated X-energy commitment represents a stronger technology conviction. Microsoft’s Aalo collaboration on AI-powered permitting signals recognition that regulatory bottlenecks may be the primary constraint, not technology readiness.
The 2029 deployment window for Aalo is underestimated in its significance. If Aalo achieves commercial operation before competitors, it captures first-mover advantage in a market that will define hyperscaler power procurement for the next decade. However, sodium’s chemical reactivity (unlike lead or helium) introduces operational risks that TRISO-based designs avoid.
Key Implication for Hyperscalers: Technology diversification across cooling approaches (lead, TRISO/helium, sodium) hedges against single-technology failure. Early movers like Aalo may capture premium pricing and deployment slots if technology proves reliable.
Outlook & Predictions
Near-term (0-6 months)
- Aalo criticality demonstration (July 2026): Success validates sodium-cooled microreactor pathway; failure creates uncertainty for entire SMR market
- Blykalla permitting initiation (late 2026): Swedish regulatory process begins; watch for timeline signals
- X-energy/Talen detailed planning (mid-2026): Site selection and permitting milestones for PJM deployment
Medium-term (6-18 months)
- Hyperscaler SMR commitments may double from ~15 GW to 25+ GW as technology risk decreases
- NRC design certification progress for XE-100 and Aalo will determine 2029-2030 deployment feasibility
- Manufacturing facility announcements from leading SMR vendors will indicate production capacity
Long-term (18+ months)
- 2029 as inflection point: First Aalo deployment could validate SMR economics for data centers; failure would set back industry 3-5 years
- Technology convergence likely: If one cooling approach demonstrates clear cost/safety advantages, expect hyperscalers to concentrate commitments
- Regulatory acceleration: Microsoft/Aalo AI permitting collaboration could reduce licensing timelines by 50%+ if successful
Key Trigger to Watch
Aalo criticality by July 4, 2026: This single milestone will validate or undermine the entire SMR-for-data-centers thesis. Success enables 2029 commercial deployment; extended delays push the market toward X-energy’s 2030s timeline.
Sources
- Planning for Swedish SMR plant proceeds — World Nuclear News, March 2026
- X-energy, Talen to assess deployment of multiple SMR plants — World Nuclear News, March 2026
- Aalo completes assembly of experimental reactor — World Nuclear News, March 2026
- Meta announces landmark agreements for new nuclear — World Nuclear News, March 2026
- Amazon updates SMR progress with new images of proposed plant — World Nuclear News, March 2026
- Duane Arnold restart underpins NextEra Energy and Google collaboration — World Nuclear News, March 2026
- Aalo Atomics and Microsoft target AI permitting gains — World Nuclear News, March 2026
- X-energy XE-100 Technical Specifications — X-energy Official, 2026
SMR Race to Power Data Centers: Three Technologies, One Market
Three SMR technologies - lead-cooled (Blykalla), TRISO (X-energy), and sodium-cooled (Aalo) - are competing to power hyperscaler AI data centers. Meta's 6.6 GW nuclear commitment signals the nuclear race has begun.
TL;DR
Three small modular reactor (SMR) technologies - lead-cooled (Blykalla), TRISO-based (X-energy), and sodium-cooled (Aalo) - are racing to power hyperscaler AI data centers. Aalo targets 2029 commercial deployment, giving it first-mover advantage. Meta’s 6.6 GW nuclear commitment represents the largest corporate nuclear purchase in US history, signaling that the SMR market for data centers has moved from theoretical to urgent.
Executive Summary
The convergence of AI power demand and SMR technology readiness has created a competitive market that did not exist five years ago. Three technologies have emerged as frontrunners: Blykalla’s lead-cooled SEALER, X-energy’s TRISO-fueled XE-100, and Aalo’s sodium-cooled microreactor. Each offers distinct trade-offs in safety profile, scalability, and deployment timeline.
The stakes are substantial. Hyperscalers (Meta, Amazon, Google, Microsoft) have committed to nuclear power at unprecedented scale:
| Company | Nuclear Commitment | Timeline | Primary Technology |
|---|---|---|---|
| Meta | 6.6 GW | By 2035 | TerraPower Natrium, Oklo Aurora, Vistra PPA |
| Amazon | 5+ GW | By 2039 | X-energy XE-100, Talen partnership |
| 600+ MW (Duane Arnold) + SMR | 2029+ | Kairos Power, NextEra | |
| Microsoft | Three Mile Island restart + SMR | 2028+ | Constellation, Helion, Aalo |
This analysis compares the three SMR technologies competing for this market, evaluates their commercialization timelines, and identifies the decision factors that will determine market leadership.
Key Facts
- Who: Blykalla (Sweden), X-energy (US), Aalo Atomics (US) - three SMR vendors with distinct cooling technologies
- What: Competition to supply nuclear power to hyperscaler AI data centers demanding 1+ GW per campus
- When: Aalo targets 2029, X-energy and Blykalla target 2030s; DOE deadline July 4, 2026 for test reactors
- Impact: Meta’s 6.6 GW commitment is the largest corporate nuclear purchase in US history; total hyperscaler demand exceeds 12 GW by 2040
Background & Context
The AI Power Problem
AI data centers have fundamentally different power requirements than traditional computing facilities. Training large language models requires sustained, high-power computation for weeks or months. Inference workloads demand instant response with minimal latency. Both require 24/7 power that renewable sources cannot reliably provide.
“The AI data center power demand has created an energy procurement challenge that nuclear is uniquely positioned to solve.” — World Nuclear News, Meta Nuclear Agreement Analysis, March 2026
A single hyperscaler campus now demands 1+ GW of continuous power. Meta’s Prometheus supercluster in New Albany, Ohio required multiple GW-scale commitments. Microsoft’s Three Mile Island restart secures 835 MW for a single data center complex. These numbers exceed the output of many utility-scale power plants.
Why SMRs for Data Centers
Small modular reactors offer three advantages that align with hyperscaler needs:
- Scale matching: SMR modules can be added incrementally as data centers expand, unlike traditional 1+ GW reactors that require massive upfront commitment
- Location flexibility: Factory-built modules can be transported to sites near demand centers, reducing transmission losses
- Carbon-free baseload: Nuclear provides 24/7 carbon-free power that matches AI workload requirements
The market opportunity has attracted over a dozen SMR developers. Three have emerged with clear pathways to commercialization: Blykalla, X-energy, and Aalo.
Analysis Dimension 1: Technology Comparison
Blykalla SEALER - Lead-Cooled Reactor
Blykalla, a spin-off from KTH Royal Institute of Technology with research dating to 1996, uses lead as the primary coolant. This approach eliminates the hydrogen explosion risk inherent in water-cooled reactors, as lead remains chemically inert at operating temperatures.
Technical Specifications:
| Parameter | Value |
|---|---|
| Coolant | Lead (chemically inert) |
| Single module capacity | ~50 MWt thermal |
| Deployment configuration | 6 reactors = 300 MW total |
| Steam temperature | Higher thermal efficiency than water-cooled |
| Commercialization target | Early 2030s |
| Primary market | Industrial heat, data centers |
The lead coolant operates at atmospheric pressure, eliminating the need for high-pressure containment systems. This simplifies construction and reduces costs. The SEALER-55 commercial reactor is rated at 140 MWt thermal output.
Blykalla announced in March 2026 that it is proceeding with planning for a 6-unit SEALER plant at Norrsundet, Gavle, Sweden. The project requires approvals from the Swedish Radiation Safety Authority, Land and Environment Court, Swedish Government, and Gavle Municipality. Permitting is expected to initiate later in 2026.
Key differentiator: Lead cooling eliminates the hydrogen explosion risk that contributed to the Fukushima accident, providing a fundamentally different safety profile than water-cooled designs.
X-energy XE-100 - TRISO Fuel, Helium-Cooled
X-energy takes a different approach with TRISO (TRIstructural-ISOtropic) particle fuel. Each fuel particle is a poppy-seed-sized uranium kernel coated with multiple ceramic layers. The US Department of Energy describes TRISO as “the most robust nuclear fuel on earth” - it can withstand extreme temperatures without melting.
Technical Specifications:
| Parameter | Value |
|---|---|
| Coolant | Helium (inert gas) |
| Single module capacity | 80 MWe / 200 MWt |
| Standard deployment | 4-unit plant = 320 MWe |
| Steam temperature | 565C |
| Design life | 60 years |
| Load-following | 40-100% in 12 minutes |
| Fuel type | TRISO-X particle fuel |
| Commercialization target | 2030s |
The XE-100 uses a pebble-bed design where fuel spheres continuously circulate through the reactor core. Helium coolant remains chemically inert in single-phase operation.
X-energy has built the largest SMR pipeline in the industry: 11+ GW across US and UK projects including:
- Amazon/Energy Northwest Cascade Center (12 XE-100 modules)
- Dow/Texas industrial deployment (first ARDP project)
- Talen/PJM market deployment (960+ MW LOI signed March 2026)
- Centrica/UK advanced reactor fleet (6 GW)
Key differentiator: TRISO fuel cannot melt at any temperature the reactor can achieve, providing inherent safety without complex active systems.
Aalo Pod - Sodium-Cooled Microreactor
Aalo Atomics, founded in 2023 in Austin, Texas, has moved fastest toward commercialization. The Aalo Pod is a 50 MWe sodium-cooled fast reactor designed specifically for data center deployment.
Technical Specifications:
| Parameter | Value |
|---|---|
| Coolant | Sodium (liquid metal) |
| Single module capacity | 10 MWe per Aalo-1 reactor |
| Standard deployment | 5 reactors = 50 MWe (Aalo Pod) |
| Fuel type | Low-enriched UO2 |
| Commercialization target | 2029 (fastest timeline) |
| Primary market | Data centers (purpose-built) |
In March 2026, Aalo completed assembly of its Critical Test Reactor at Idaho National Laboratory. The company expects to achieve criticality before the July 4, 2026 DOE deadline for the Nuclear Reactor Pilot Program - a key milestone toward commercial licensing.
Aalo’s design emphasizes factory-built modularity. Each reactor is manufactured in a factory and transported to the site, enabling rapid deployment. The company explicitly targets data centers as its primary market.
Key differentiator: Aalo is the only SMR designed from the ground up for data center deployment, with 50 MW granularity matching typical data center power increments.
Technology Comparison Matrix
| Dimension | Blykalla SEALER | X-energy XE-100 | Aalo Pod |
|---|---|---|---|
| Coolant | Lead (inert) | Helium (inert gas) | Sodium (reactive metal) |
| Fuel | Not disclosed | TRISO-X particles | Low-enriched UO2 |
| Module size | ~50 MWt | 80 MWe / 200 MWt | 10 MWe |
| Deployable unit | 300 MW (6 units) | 320 MWe (4-pack) | 50 MWe (5 reactors) |
| Key safety feature | No hydrogen risk | Fuel cannot melt | Fast-neutron stability |
| Load-following | Not disclosed | 40-100% in 12 min | Not disclosed |
| Design life | Not disclosed | 60 years | Not disclosed |
| Commercialization | Early 2030s | 2030s | 2029 |
| Pipeline scale | 300 MW demo | 11+ GW | Not disclosed |
| Data center fit | Medium scale | Hyperscale | Purpose-built |
Analysis Dimension 2: Commercialization Timeline Race
The 2029 Window
Aalo’s 2029 commercial target gives it first-mover advantage. If successful, Aalo will deliver the first SMR designed specifically for data centers three to five years before competitors.
The timeline depends on DOE Nuclear Reactor Pilot Program success. Aalo must demonstrate criticality by July 4, 2026 to maintain program momentum. The completed Critical Test Reactor at INL positions the company to meet this milestone.
“Aalo expects criticality before July 4, with commercial Aalo Pod deployment targeted for 2029.” — World Nuclear News, March 2026
X-energy’s Pipeline Advantage
X-energy may trail Aalo in first deployment but leads in total pipeline commitment. The company has secured:
- ARDP funding for Dow/Texas first deployment
- Amazon partnership for 5 GW by 2039
- Talen LOI for 960+ MW in PJM market
- Centrica agreement for 6 GW UK fleet
The breadth of X-energy’s partnerships provides validation that TRISO technology meets hyperscaler requirements. Amazon’s commitment to 12 XE-100 modules at the Cascade Advanced Energy Center demonstrates confidence in the technology at scale.
Blykalla’s European Position
Blykalla represents the European SMR option, with Sweden as the lead market. The Norrsundet plant will require multi-year permitting, pushing commercial operation to the early 2030s.
European data center operators (currently relying on grid power) may find Blykalla attractive for:
- Lower regulatory barriers in Sweden compared to US NRC licensing
- Proximity to Nordic data center clusters
- Industrial heat applications beyond electricity
Critical Path Milestones
| Milestone | Blykalla | X-energy | Aalo |
|---|---|---|---|
| Test reactor criticality | Not applicable | Not applicable | July 2026 target |
| First commercial deployment | Early 2030s | ~2030 | 2029 |
| Hyperscaler partnership | None announced | Amazon (5 GW), Talen, Centrica | Microsoft (permitting AI) |
| Regulatory approval pathway | Swedish 4-agency process | NRC design certification | DOE Pilot Program + NRC |
| Manufacturing facility | Not disclosed | Under development | Factory model |
Analysis Dimension 3: Hyperscaler Requirements Analysis
What Data Centers Actually Need
Hyperscaler power requirements differ from traditional baseload nuclear customers in four dimensions:
1. Power Density
AI training clusters consume power at densities of 50-100 MW per acre. A single hyperscaler campus can require 1+ GW continuous. This exceeds the output of most single nuclear units and approaches the scale of multi-reactor plants.
| Hyperscaler Project | Power Requirement | Technology |
|---|---|---|
| Meta Prometheus (Ohio) | Multiple GW | Natrium SMR, existing PPA |
| Amazon Cascade Center | 320-960 MW | X-energy XE-100 |
| Microsoft Three Mile Island | 835 MW | Existing reactor restart |
| Google Duane Arnold | 600+ MW | Existing reactor restart |
2. Carbon-Free Requirement
All four hyperscalers (Meta, Amazon, Google, Microsoft) have publicly committed to carbon-free energy targets:
- Meta: Net zero by 2030
- Amazon: Net zero by 2040
- Google: Carbon-free energy 24/7 by 2030
- Microsoft: Carbon negative by 2030
Nuclear provides the only scalable baseload carbon-free option. Solar and wind require massive overbuild plus storage, with land requirements that many data center locations cannot accommodate.
3. Load-Following Capability
AI workloads fluctuate based on training schedules and inference demand. X-energy’s XE-100 can ramp from 40% to 100% power in 12 minutes - significantly faster than Gen III reactors that require hours for load changes.
This capability matters for data centers that may vary power consumption:
- Training runs: sustained 100% load for weeks
- Inference peaks: rapid scaling based on user demand
- Maintenance windows: reduced load during equipment changes
4. Location Flexibility
Data centers locate based on connectivity, climate, and tax incentives - not proximity to power plants. SMRs that can be transported to demand centers have a fundamental advantage over traditional nuclear.
X-energy emphasizes “road-shippable” design for the XE-100. Aalo emphasizes factory-built modularity. Both approaches enable siting near data center clusters.
Hyperscaler Partnership Landscape
The partnerships formed in 2024-2026 reveal hyperscaler technology preferences:
| Hyperscaler | SMR Technology | Capacity | Status |
|---|---|---|---|
| Meta | TerraPower Natrium | 2.8 GW (8 units) | Agreement signed |
| Meta | Oklo Aurora | 1.2 GW Ohio campus | Agreement signed |
| Meta | Vistra (existing plants) | 2,176 MW | PPA signed |
| Amazon | X-energy XE-100 | 5+ GW target | Partnership |
| Kairos Power | SMR | Agreement signed | |
| NextEra/Duane Arnold | 600+ MW | 25-year PPA | |
| Microsoft | Constellation/TMI | 835 MW | Restart agreement |
| Microsoft | Helion Fusion | Not disclosed | PPA signed |
| Microsoft | Aalo Atomics | Not disclosed | AI permitting collaboration |
Meta’s approach is notable for its breadth: existing plant restarts (Vistra), sodium-cooled SMRs (TerraPower Natrium), and fast reactors (Oklo Aurora). This diversification suggests hedging across technologies rather than committing to a single approach.
Amazon has concentrated on X-energy’s TRISO technology, with a 5 GW commitment by 2039. This represents a stronger technology bet than Meta’s portfolio approach.
Microsoft’s strategy combines near-term restarts (Three Mile Island), fusion (Helion), and advanced permitting (Aalo). The Aalo collaboration on AI-powered permitting acceleration may indicate deeper technology interest.
Google’s primary commitment is the Duane Arnold restart (600+ MW by 2029) with Kairos Power SMR agreements for longer-term power.
Analysis Dimension 4: Regulatory and Market Outlook
US Regulatory Environment
The DOE Nuclear Reactor Pilot Program has created a race to criticality. The July 4, 2026 deadline has concentrated industry attention on demonstrating advanced reactor viability. Aalo’s test reactor completion positions it well for this milestone.
The NRC design certification process remains the primary regulatory hurdle for commercial deployment. X-energy has invested heavily in pre-application engagement with NRC. Aalo benefits from the DOE Pilot Program’s streamlined pathway.
Key regulatory indicators to watch:
- NRC design certification progress for each technology
- State-level nuclear development incentives (Wyoming, Texas, Ohio)
- Federal loan guarantee availability
European Regulatory Environment
Blykalla faces a four-agency approval process in Sweden:
- Swedish Radiation Safety Authority
- Land and Environment Court
- Swedish Government
- Gavle Municipality
This multi-layered process will extend permitting timelines compared to US projects, but Sweden’s supportive nuclear policy (reversing previous phase-out plans) creates a favorable environment.
Market Size Projections
The hyperscaler nuclear market has grown from negligible to substantial in 18 months:
| Year | Estimated Hyperscaler Nuclear Demand |
|---|---|
| 2024 | < 1 GW committed |
| 2026 | ~15 GW committed (Meta 6.6, Amazon 5+, Google, Microsoft) |
| 2030 | First SMR deployments operational |
| 2035 | 20+ GW cumulative SMR capacity projected |
| 2040 | 40+ GW potential if technology proves reliable |
The transformation from near-zero to 15+ GW committed capacity in 24 months indicates that the market has crossed the adoption threshold from experimental to operational planning.
Key Data Points
| Metric | Value | Source | Date |
|---|---|---|---|
| Meta nuclear commitment | 6.6 GW | World Nuclear News | 2026-03 |
| Amazon nuclear target | 5+ GW by 2039 | World Nuclear News | 2026-03 |
| Google Duane Arnold capacity | 600+ MW | World Nuclear News | 2026-03 |
| X-energy total pipeline | 11+ GW | World Nuclear News | 2026-03 |
| DOE test reactor deadline | July 4, 2026 | World Nuclear News | 2026-03 |
| Aalo commercial target | 2029 | World Nuclear News | 2026-03 |
| XE-100 load-following | 40-100% in 12 min | X-energy Official | 2026-03 |
| TerraPower Natrium capacity | 690 MW + 500 MWh storage | World Nuclear News | 2026-03 |
| Blykalla SEALER plant | 300 MW (6 reactors) | World Nuclear News | 2026-03 |
| Aalo Pod capacity | 50 MWe (5 reactors) | World Nuclear News | 2026-03 |
🔺 Scout Intel: What Others Missed
Confidence: high | Novelty Score: 85/100
Coverage of SMR announcements treats each technology as an isolated story. The strategic pattern is missed: hyperscalers are not merely buying power - they are locking in manufacturing capacity for SMR technologies that have not yet proven commercial viability.
Meta’s 6.6 GW commitment spreads risk across three technology bets (Natrium, Oklo, Vistra PPA). Amazon’s concentrated X-energy commitment represents a stronger technology conviction. Microsoft’s Aalo collaboration on AI-powered permitting signals recognition that regulatory bottlenecks may be the primary constraint, not technology readiness.
The 2029 deployment window for Aalo is underestimated in its significance. If Aalo achieves commercial operation before competitors, it captures first-mover advantage in a market that will define hyperscaler power procurement for the next decade. However, sodium’s chemical reactivity (unlike lead or helium) introduces operational risks that TRISO-based designs avoid.
Key Implication for Hyperscalers: Technology diversification across cooling approaches (lead, TRISO/helium, sodium) hedges against single-technology failure. Early movers like Aalo may capture premium pricing and deployment slots if technology proves reliable.
Outlook & Predictions
Near-term (0-6 months)
- Aalo criticality demonstration (July 2026): Success validates sodium-cooled microreactor pathway; failure creates uncertainty for entire SMR market
- Blykalla permitting initiation (late 2026): Swedish regulatory process begins; watch for timeline signals
- X-energy/Talen detailed planning (mid-2026): Site selection and permitting milestones for PJM deployment
Medium-term (6-18 months)
- Hyperscaler SMR commitments may double from ~15 GW to 25+ GW as technology risk decreases
- NRC design certification progress for XE-100 and Aalo will determine 2029-2030 deployment feasibility
- Manufacturing facility announcements from leading SMR vendors will indicate production capacity
Long-term (18+ months)
- 2029 as inflection point: First Aalo deployment could validate SMR economics for data centers; failure would set back industry 3-5 years
- Technology convergence likely: If one cooling approach demonstrates clear cost/safety advantages, expect hyperscalers to concentrate commitments
- Regulatory acceleration: Microsoft/Aalo AI permitting collaboration could reduce licensing timelines by 50%+ if successful
Key Trigger to Watch
Aalo criticality by July 4, 2026: This single milestone will validate or undermine the entire SMR-for-data-centers thesis. Success enables 2029 commercial deployment; extended delays push the market toward X-energy’s 2030s timeline.
Sources
- Planning for Swedish SMR plant proceeds — World Nuclear News, March 2026
- X-energy, Talen to assess deployment of multiple SMR plants — World Nuclear News, March 2026
- Aalo completes assembly of experimental reactor — World Nuclear News, March 2026
- Meta announces landmark agreements for new nuclear — World Nuclear News, March 2026
- Amazon updates SMR progress with new images of proposed plant — World Nuclear News, March 2026
- Duane Arnold restart underpins NextEra Energy and Google collaboration — World Nuclear News, March 2026
- Aalo Atomics and Microsoft target AI permitting gains — World Nuclear News, March 2026
- X-energy XE-100 Technical Specifications — X-energy Official, 2026
Related Intel
Aalo Atomics Completes Microreactor Assembly, Eyes Data Center Market
Aalo Atomics completes assembly of sodium-cooled Critical Test Reactor at Idaho National Laboratory, targeting July criticality and 2029 commercial operation for 50 MWe Aalo Pod data center power.
Vietnam-Russia Sign VVER-1200 Nuclear Plant Cooperation Agreement
Vietnam and Russia signed an intergovernmental agreement to build Ninh Thuan 1 Nuclear Power Plant with two VVER-1200 reactors, marking Vietnam's return to nuclear after 2016 suspension.
Small Modular Reactors (SMRs): A Comprehensive Guide to Next-Generation Nuclear Power
A decision-maker's guide to SMR technology, economics, and deployment. Compare 6 leading designs, understand regulatory pathways, and evaluate the business case for nuclear's future.