Power to Compute 01 / The new grid
Power
Compute

Enabling existing power infrastructure to compute

Power electronics can compute, let's unlock it!

Book a demo

Compute capacity where power already exists.

02 Two routes to P2C

Create compute where power already exists.

For power infrastructure owners

Own power infrastructure? You are only a software update away from increased revenue!

Your power infrastructure can compute - enable it now! Zero effect on operations, but with an increased revenue stream.

For compute operators

Need more compute, NOW? Let existing power infrastructure compute - no new GPUs required!

Outsource part of your workload to grow your customer base while still meeting peak demand.

03 Additional revenue streams

Unlock computing - a cherry on top

Complementing the ongoing efforts on workload orchestration engines and carbon-aware computing

Additional revenue before and after P2C enablement Before P2C, a grid asset and data center generate revenue only through flexibility services. When P2C is enabled, the same two assets retain flexibility-service revenue and gain an additional route through P2C. BEFORE P2C ENABLED GRID ASSET EXISTING POWER INFRASTRUCTURE FLEXIBILITY SERVICES GRID BALANCING REVENUE DATA CENTER COMPUTE OPERATOR P2C COMPUTE REVENUE ADDITIONAL REVENUE STREAM
04 Software + marketplace

We provide the software
enabling compute
and the marketplace to
access and sell it.

05 / The P2C shift 01 — 03

01 / Reverse the flow

Power can’t reach compute. So compute moves to power.

Grid constraints stop new data-center capacity from reaching available generation. P2C reverses the equation by bringing modular compute directly to renewable power.

  • 01 Grid capacity blocks expansion
  • 02 Compute relocates to generation
  • 03 Power becomes computing capacity

02 / Orchestrate workloads

Find capacity. Route workloads. Compute where power exists.

Data centers use the P2C marketplace to discover available computing resources. The marketplace matches each workload and dispatches it to the relevant physical grid component.

  • 01 Discover available computing resources
  • 02 Match workloads to grid assets
  • 03 Run on existing infrastructure

03 / Scale intelligently

From years of waiting to months of deployment.

Modular compute follows renewable power—unlocking scalable AI capacity with fewer permits, fewer barriers, and less waste.

  • 01 No new land permissions
  • 02 No major grid investments
  • 03 No hardware deployment
  • 04 Sustainable by design
06 P2C methodology

Unlocking hidden compute

Power electronics before and after P2C software enablement A solar farm connects to the DC side of a PV power-electronic converter, while a three-unit battery bank connects to the DC side of a second converter. The converters share an AC connection. Before P2C, only this white physical power system is shown. When P2C is enabled, the topology turns green and AI workload, encoder, decoder, output, and moving compute signals appear. BEFORE P2C ENABLED SOLAR FARM PV POWER ELECTRONIC CONVERTER BATTERY STORAGE BATTERY POWER ELECTRONIC CONVERTER SHARED AC CONNECTION AI WORKLOADS IMAGES VIDEOS SPEECH DOCUMENTS SENSOR DATA ENCODER P2C ENCODER DECODER P2C DECODER OUTPUT ACCURACY R-SCORE PRECISION
07 One system

Same power.
No cooling.
Existing infrastructure.

08 Data center impact

Benefits per datacenter

Expand computing capacity by using power and infrastructure that already exist. Let P2C handle your delay-tolerant workloads.

10%*

Compute capacity expansion

GPU utilizationOffload baseline inference; reserve GPUs for heavy workloads.

5%*

Peak demand reduction

Peak-to-average ratioTail-load offload flattens demand.

Peak-to-idle ratioLower bursts; standby baseline stays stable.

€1 million*

Saving per year

Carbon Usage EffectivenessNear-zero marginal carbon dioxide per outsourced floating-point operation.

€5–15 million*

Infrastructure bottleneck savings

Rate of power changeControlled slew cuts uninterruptible power supply, power distribution unit and grid stress.

10–20%*

Reduction in cooling budget and demand

Power Usage EffectivenessP2C workloads run at approximately 1.0 effective Power Usage Effectiveness.

*approximate figures for a 100 megawatt data center — numbers vary based on data center capacity.

09 The team

The team

Subham Sahoo

Subham Sahoo

Associate Professor

AAU Energy
Freja Frost Kjeldsen

Freja Frost Kjeldsen

Data Scientist

AAU Energy
Jonas Lindegaard Mikkelsen

Jonas Lindegaard Mikkelsen

Power Engineer

AAU Energy
Yubo Song

Yubo Song

Postdoc

AAU Energy
10 Latest updates

From the field

P2C updates, research coverage, and perspectives from the field.

AAU Tech Transfer 14 Aug 2026

Power electronics can do more than deliver electricity

AAU Technology Transfer Office introduces P2C: using hidden computing capacity in grid-connected power electronics to support artificial intelligence workloads.

Read on LinkedIn
Electricity transmission lines crossing a field under a blue sky
DI Energi 03 Aug 2026

New research: AI computation can relieve pressure on the grid

Danish Industry highlights P2C research from Aalborg University and its potential to turn existing energy infrastructure into distributed computing capacity.

Read the article
PCIM Expo 2026 speaker announcement for Subham Sahoo's P2C presentation
LinkedIn PCIM Expo 2026

Plug-and-play P2C heads to PCIM Expo 2026

Subham previews a live P2C demonstration for data centers and power-converter OEMs—showing how AI can scale using existing power infrastructure.

Read on LinkedIn
Subham Sahoo presenting P2C at the AI and Data Centers Stage
LinkedIn Event recap

Industry demand for P2C comes into focus at PCIM

Conversations at the AI & Data Centers Stage put grid bottlenecks, long equipment lead times, and the value of reusing installed infrastructure in focus.

Read on LinkedIn

11 / Start the conversation

Compute is hidden.
P2C finds it.

Book a conversation about your power site, compute needs, or partnership.