Developed within Carbon13 · Cambridge University Venture Builder

Infrastructure that knows itself

A corporate bond is re-priced daily. A data centre portfolio is scored once, at construction, and again at acquisition. Sono keeps a live, agent-read score on every asset you hold and every site you're evaluating, so risk is priced the day it moves, not the year it's discovered.

See how it works Request a pilot
Revenue
SLA breach avoidance, stranded capacity
Capex cycles
Equipment lifecycle, refresh timing
Opex
Energy, predictive maintenance, downtime
Compliance
Regulatory and permit status, tracked live
Insurance
Fire, water and climate hazard exposure
Site selection
Grid access, permitting, climate zoning
The problem

Risk moves continuously. Assessment doesn't.

A corporate bond is re-priced daily. A €5 billion data centre portfolio is scored once, at construction, and again at acquisition, by a consultant. Equipment wears down, contracts renew, insurance terms shift, climate and grid conditions move — the risk keeps moving, and nobody is looking.

Investors

Price assets on data that's already out of date by the time it reaches the investment committee.

Operators

Find out about failure risk after it happens, not before.

Insurers

Underwrite on a survey that's eleven months old, so exposure builds between renewals and prices at the wrong risk.

Sustainability teams

Unplanned failure is the more carbon-intensive outcome — emergency replacement instead of planned end-of-life.

The cost of the gap

Undetected risk becomes a stranded asset

The stranded-asset situation in Ireland wasn't a failure of ambition or capital. It was a failure to track one risk factor — grid capacity — continuously, before it became a binding constraint. When the grid operator froze new connections, projects that looked viable at acquisition had no path to power.

€176bn
Committed to European data centres, 2026–31. None of it is continuously scored.
€5.8bn
Already stranded in Ireland alone — land bought and permitted, but no grid connection.
Product

Two modules, one output.

Sono runs a network of agents that read contract terms, insurance terms, manufacturer documentation and external conditions for every asset, continuously. Portfolio Management covers what you already hold; Portfolio Expansion covers what you're evaluating.

Portfolio Expansion

The same plot can look very different depending on what grid capacity is actually available, whether the water basin is under stress, or how many competing facilities are already queued for permits nearby. Sono combines those signals into a single score, calibrated against a growing reference population of European sites, so you can compare candidates on a common basis before committing capital.

No instrumentation required. One category — grid capacity, climate exposure, regulatory trajectory — because the question at this stage is narrow: should you commit capital here at all.

Existing DCs
5
monitored
Candidates
3
under eval.
Avg score
74
sono
Layers
9
active
Data Layers
Your DCs
Existing DCs
5
Candidates
3
Heat island
est.
IX points
Environment
Rivers
1240
Reservoirs
44
Power plants
62
Grid substations
187
HV power lines
341
Selected site
FRA-North-03
Score 87 · High
Eschborn, Hesse
87 FRA-North-03 71 RHN-East-01 58 HAN-West-06 sono. Frankfurt Rhine-Main + - 15 km
FRA-North-03
Eschborn, Hesse
50.14°N, 8.56°E · 6.8 ha
87
Resilience score
Power 91
380kV line 0.9km · dual feed
Climate 82
3,540 free cooling hrs/yr
Compliance 78
EnEfG compliant · 1 open item
Infra 90
Low seismic · no flood zone
Adaptability 71
Liquid cooling ready · partial
Top ranked
Best grid access in portfolio. Taunus foothills reduce UHI exposure. Minor EnEfG gap addressable pre-build.

Management asks one question: what is happening to the assets we already hold. The agents cross-reference contract, insurance and manufacturer documentation against each asset's actual condition, continuously, not at review time.

Live risk tracking

Five categories, one live state per asset: revenue, capex cycles, opex, compliance and insurance, each scored and updated as the inputs move. Financial impact is projected forward in euros today, and the year it crosses your action threshold.

Portfolio-level. No site visits. The agent states what it found, what it costs, and the next step to take.

sono · Portfolio Management · MXP-Edge-07
Agent summary
Two categories need attention this cycle: Capex cycles and Insurance.
Chiller 1 is past manufacturer end-of-life with no replacement scheduled. Current insurance terms leave a coverage gap on the climate-hazard rider renewed last quarter.
Revenue Renewal 2027 contract escalation clause reviewed
Capex cycles Chiller at EOL flagged before failure, not after
Opex +8% vs. plan tracked against actual spend
Compliance Permit current no open regulatory items
Insurance Coverage gap climate-hazard rider under-covers current risk

Operational Efficiency

The largest of the five categories: predictive maintenance, anomaly detection, and setpoint optimisation delivered to operations teams in plain language. Sono reads your existing historian, detects failure modes before alarms fire, traces root cause to specific equipment, and surfaces ranked recommendations with a confidence score.

No new hardware. No integration project. Historian-native. Physics-constrained so no physically impossible setpoint can ever be suggested.

sono · MXP-Edge-07 · 06:04
Anomaly detected · Chiller 1 · 06:00
Recommendation · 94% confidence
Raise Chiller 1 setpoint: 6.0°C → 8.5°C
Valve response lag detected. Current setpoint is below thermal floor given ambient conditions and IT load. Adjustment remains within safe operating envelope. Estimated compressor load reduction: 3.2% on this circuit.
Confidence 94%
Est. saving 3.2%
Reversible Yes
Value created

What both modules are worth to the owner

Ranges are Sono's modelled estimates for a 60 MW colocation asset, to be replaced with pilot-measured figures once the first deployment closes.

20%+
Value preservation
Insurability, compliance and equipment condition tracked continuously, so the asset holds its investment grade instead of drifting toward a stranded position.
~5%
Opex optimisation
Predictive maintenance and setpoint discipline instead of reactive callouts and over-cooling.
+3–5 yrs
Asset life extension
Replacement planned at end-of-life rather than forced by failure, spreading capex across the hold period.
Lower
Carbon footprint
Planned end-of-life and correct cooling avoid the emergency, carbon-intensive outcome regulators now treat as a risk category.
Coverage

21 scored layers across six domains

Internal and external data — 30+ sources in — mapped through category models onto the levers an owner actually pulls.

Revenue

2

SLA and uptime breach avoidance, stranded capacity monetisation.

Model · Uptime & capacity

Capex cycles

1

Asset refresh and end-of-life timing.

Model · Equipment lifecycle & maintenance

Opex

10

Energy, predictive maintenance, labour, downtime and SLA penalties, and six more.

Model · Energy & resource consumption · equipment lifecycle

Compliance

1

Contractual and SLA compliance.

Model · Infrastructure & resource availability

Insurance

2

Fire and water damage, climate and catastrophe hazard exposure.

Model · Physical & climate hazard

Site selection

5

Grid and power diligence, climate hazard zoning, permitting risk, and two more.

Model · Site viability
Team

Founding team

Three co-founders combining ML, energy infrastructure and commercial execution.

Harit Soni

Harit Soni

Co-founder, Product Strategy

Two-time founder. Raised $22M. 15+ years building and scaling cleantech, IoT and AI across Asia, the UK and North America.

Johannes Parzonka

Johannes Parzonka

Co-founder, Commercial

12+ years commercial leader. Former CEO of a green-gas spin-off. Led M&A across wind, PV and biomethane.

Caterina Giacomelli

Caterina Giacomelli

Co-founder, Technology

Physicist and Data Scientist. 10+ years experienced in signal processing & ML, IoT monitoring and industrial processes.

Contact

Work with us on an early pilot

Data centres carry risk across five dimensions simultaneously. The data to track most of it already exists. Sono reads it, structures it, and gives your team something to act on.

We work with colocation operators, infrastructure fund managers, and OEM partners across Europe.

Supported by