Utility-scale PV • Software-native delivery

One AI platform for the entire PV lifecycle.

From design automation to autonomous site supervision, from AI SCADA to predictive O&M—Xskylight unifies every stage into one auditable system. We replace fragmented tools and manual workflows with a single platform that connects design intent to construction reality to operational performance.

Design → Build → Operate
One modular system covering every phase
Edge + Cloud
Real-time automation on site, fleet intelligence in the cloud
PV + BESS + H2
Hybrid-ready architecture from day one

The numbers behind our platform. Every metric represents real engineering output—from design pipelines that compress weeks into hours, to anomaly detection that shrinks response time from days to minutes.

500+ MW
Design capacity automated through AI pipelines
< 3 min
From anomaly detection to full triage report
99.5%
Platform availability SLA for mission-critical operations
6 modules
One integrated platform covering the full lifecycle
Aerial overview of a fully operational utility-scale solar farm with central substation, showing precise panel alignment across hundreds of rows

Platform

Six modules that cover engineering, construction oversight, SCADA/EMS, and O&M—deployed together or independently.

The PV industry operates on fragmented tools—one for design, another for monitoring, a third for maintenance, a fourth for document management. None of them talk to each other. Xskylight replaces this patchwork with a unified platform where data flows continuously between modules. When a design decision changes, the construction verification module knows immediately. When SCADA detects an anomaly, it references the design model to diagnose root cause. When O&M generates a maintenance ticket, it traces the issue back to the component’s full history.

Each module is production-ready on its own, but the real value is in the connections between them. This is integration by architecture, not by stitching APIs together after the fact.

Engineer reviewing solar farm CAD layout on a high-resolution monitor with design automation software

Design Automation

DXF-aware pipelines that produce layouts, routing, string plans, BOQ, and full documentation packages—validated by engineering rules and constraints. What traditionally takes a team of engineers weeks of manual work, our system completes in hours with full traceability. Every output can be audited back to its design inputs, making it ready for EPC handoff and permitting without rework.

  • CAD + GIS ingestion with automatic coordinate alignment
  • Deterministic pipelines combined with AI design assistants
  • Auditable outputs formatted for EPC contractors & permitting authorities
  • Parametric design iteration for rapid site optimisation
Technical documents being digitally extracted and analysed on a tablet screen

Document Intelligence

Automated due diligence that transforms how projects handle documentation. Our system extracts entities from unstructured documents, detects missing or expired records, and generates structured project context and templates. Think of it as an AI analyst that has read every document in the project room—instantly finding what you need, flagging what’s missing, and generating what doesn’t exist yet.

  • RAG-powered extraction with citation-ready outputs
  • Template-driven document generation for compliance packages
  • Automatic version tracking and expiration monitoring
  • Cross-project knowledge retrieval for institutional learning
Professional surveying drone hovering above rows of solar panels during an autonomous inspection mission

Autonomous Site Supervision

Drone and ground-level capture feeds into AI models that produce orthomosaic maps, point clouds, and object-level progress reports. The system compares what is built against the design model and generates evidence packages automatically—no manual site walks, no subjective assessments. Construction deviations that traditionally went unnoticed until commissioning are now caught within 24 hours of occurrence.

  • Thermal + RGB analysis for defect and progress detection
  • Object-level construction progress tracking against design model
  • Visual reporting with DXF overlays and deviation heatmaps
  • Automated mission planning for systematic site coverage
SCADA control room with multiple monitoring dashboards showing real-time solar plant performance

AI SCADA + EMS

Unified telemetry ingestion, anomaly detection powered by physics-informed models, generation forecasting, and dispatch optimisation for PV plants and hybrid installations. Unlike traditional SCADA systems that rely on static threshold alarms, our AI models understand what the plant should be producing given current irradiance, temperature, and degradation profiles. For hybrid sites combining PV, BESS, and hydrogen, the EMS layer optimises across multiple asset types in real time.

  • Device-level observability from inverter to transformer
  • Forecast-driven dispatch and curtailment mitigation
  • Grid-code aligned control strategies for every market
  • Physics-informed anomaly detection that reduces false positives by 80%
Field technician reviewing predictive maintenance analytics on a ruggedised tablet at a solar site

O&M Analytics

Predictive maintenance and performance diagnosis that goes far beyond simple threshold alarms. Our system diagnoses root causes, prioritises actions based on financial impact, and generates complete evidence packages for warranty claims and insurance. From a single alarm to a fully contextualised triage report—including historical data, component specifications, and recommended actions—in under three minutes. Fleet operators gain portfolio-wide visibility into performance patterns that would be invisible at the single-site level.

  • Automated triage and prioritisation by financial impact
  • Complete ticket + evidence packages for field crews
  • Fleet-level benchmarking across sites and climate zones
  • Warranty claim automation with traceable evidence chains
3D digital twin visualisation of a utility-scale solar farm showing component-level data

Digital Twin

The connective tissue of the entire platform. A living model that links design intent to as-built reality to operational performance—so every decision is traceable across the full asset lifecycle. Every component is tracked from specification through installation to degradation. Every change is versioned. Every anomaly can be traced to its origin in the design or construction phase. This is how you turn a solar farm into a software-defined asset.

  • Design twin ↔ operational twin synchronisation
  • Change impact analysis across the full lifecycle
  • Component-level asset knowledge base with full provenance
  • Scenario modelling for repowering and extension decisions

Our modular architecture means you can start with the module that addresses your most urgent need—design automation for a new project, SCADA for an operational fleet, O&M analytics for a portfolio under warranty pressure—and expand as your requirements grow. The platform is designed to deliver value from day one, not after a twelve-month integration programme.

Solutions

From single PV sites to hybrid portfolios with storage and hydrogen—configured for your delivery model and operational reality.

Every energy project is different in scale, technology mix, and contractual structure. But the underlying workflows—design, verify, monitor, maintain—follow common patterns. Xskylight provides solution configurations tailored to the most demanding segments of the utility-scale market, each drawing on the platform’s six modules in combinations optimised for that use case.

Aerial orthomosaic of a solar farm construction site showing precise panel installation progress

Utility-scale PV delivery

Accelerate engineering and keep construction aligned with the design model—down to individual components and milestones. Our design automation module produces EPC-ready outputs in hours, not weeks. During construction, autonomous inspection compares every installed row against the design model, flagging deviations before they compound into costly rework. From first CAD import to final commissioning certificate, the platform provides a continuous thread of verified data.

This is how you deliver a 200 MW site on schedule: by replacing manual verification with automated, objective, design-referenced progress tracking at every stage of the construction programme.

EPC-ready outputsProgress verificationQuality assuranceCommissioning support
Battery energy storage containers installed alongside a utility-scale solar farm

Hybrid optimisation (PV + BESS + H2)

The energy market is moving rapidly from pure PV to hybrid installations that combine solar, battery storage, and increasingly hydrogen production. This transition multiplies the complexity of dispatch decisions, grid compliance, and asset coordination. Our EMS layer optimises across all asset types simultaneously, using forecast-driven strategies that account for irradiance predictions, market prices, grid constraints, and battery degradation curves.

Result: maximised revenue per MWh generated, minimised curtailment, and full compliance with grid codes across every market you operate in—from the UK to continental Europe and beyond.

EMSForecastingDispatch optimisationMulti-asset coordination
Construction site engineer with tablet overseeing solar farm installation progress

Autonomous construction oversight

Traditional construction oversight relies on manual site walks, photo logs, and subjective assessments. Our platform replaces this with drone-captured orthomosaics and point clouds, processed by AI models that detect objects, measure progress, and compare against the design model. The output is an objective progress report with visual evidence, DXF overlays, and deviation heatmaps—generated automatically, without requiring a site engineer to walk every row.

For multi-site developers managing construction across geographies, this means consistent quality assurance regardless of which contractor is on the ground. Every site is measured by the same standard, with the same rigour.

Drone missionsOrtho mapsDXF overlaysEvidence packages
Field maintenance crew with ruggedized laptop showing dispatch priorities at sunrise

Performance-driven O&M

Reduce downtime with proactive diagnostics, anomaly detection, and prioritised action pipelines across your entire fleet. Our O&M solution combines SCADA telemetry, thermal inspection data, and historical performance patterns to identify issues before they cause generation loss. When a fault is detected, the system generates a complete triage package—root cause analysis, financial impact estimate, recommended action, and all supporting evidence—and assigns it to the right team automatically.

For fleet operators managing hundreds of megawatts, this transforms O&M from a reactive cost centre into a data-driven performance optimisation function that directly protects revenue.

Anomaly detectionPredictive maintenanceSLA workflowsFleet analytics
Operations center video wall showing fleet management dashboard across multiple solar sites

Unified operations across your portfolio

Whether you operate five sites or fifty, Xskylight gives you a single pane of glass across your entire portfolio. Fleet-level dashboards aggregate performance, flag outliers, and benchmark sites against each other—revealing patterns that would be invisible at the individual site level.

Operators gain the ability to compare performance across different climate zones, inverter manufacturers, and O&M contractors. This is how you move from reactive operations to continuous, data-driven improvement across every asset you manage.

Portfolio viewCross-site benchmarkingUnified alerting

How it works

Four stages. One continuous data thread. Every output auditable from input to decision.

The Xskylight platform operates as a continuous pipeline: data enters from any source—CAD files, drone captures, SCADA telemetry, documents—and moves through analysis, action, and reporting stages. At every step, the system maintains full provenance: every output can be traced back to the data that produced it and the model that processed it. This is auditability by design, not by afterthought.

Below is the four-stage pipeline that powers every module in the platform. Whether you’re running design automation or fleet-level O&M analytics, the underlying pattern is the same: ingest, analyse, act, report.

Server room corridor with glass-fronted racks and fiber optic cables conveying data infrastructure
Step 1
Ingest
CAD files (DXF, DWG), GIS shapefiles, SCADA telemetry streams, drone captures (RGB + thermal), and project documents. The platform normalises every format into a unified data model, so downstream modules work with clean, structured data regardless of its source.
Step 2
Analyse
AI models process, validate, and cross-reference data against engineering rules, historical patterns, and physics-based expectations. Design layouts are checked for constraint violations. Drone imagery is compared against the design model. Telemetry is evaluated against forecast and degradation baselines.
Step 3
Act
Generate optimised designs, dispatch control commands to inverters and batteries, trigger prioritised alerts, create triage tickets with full evidence packages, and produce documentation—all automatically, with human approval gates where required by your operational procedures.
Step 4
Report
Auditable outputs with visual evidence, traceable decision logs, and complete provenance chains. Every report can be traced to the raw data that produced it. Formats designed for stakeholders: EPC contractors, asset managers, investors, and regulatory authorities.
Glass whiteboard with connected flow diagrams showing design-to-operate data traceability

One continuous data thread

The key architectural principle is traceability. Every piece of data that enters the platform—a CAD file, a drone image, a telemetry reading, a document—is tagged, versioned, and linked to everything it produces downstream.

When an O&M analyst reviews a maintenance recommendation, they can trace it back through the anomaly detection model, the SCADA data that triggered it, the design specifications of the affected component, and the construction records that documented its installation. This is the level of traceability that regulated infrastructure demands.

Split-screen workstation showing live SCADA telemetry and automated triage report generation

From data to decisions in minutes

Traditional workflows take days or weeks to move from raw data to actionable intelligence. A drone survey generates imagery that sits on a hard drive waiting for manual review. A SCADA alarm generates a notification that joins a queue of uncontextualised alerts.

Xskylight compresses this timeline. Drone imagery is processed and compared against the design model within hours of capture. SCADA anomalies are contextualised, diagnosed, and triaged in under three minutes. Documents are extracted and structured as they arrive. The gap between data and decision narrows to minutes, not days.

Innovation pillars

Four principles that define our engineering philosophy. These are not features bolted on—they are the architectural foundation everything else is built upon.

Building software for regulated infrastructure requires a different mindset. Every decision we make—from model selection to data architecture to API design—is governed by four non-negotiable principles. These are not aspirational values on a poster. They are engineering constraints that shape every line of code and every product decision we make. They are what make Xskylight suitable for assets where reliability, traceability, and continuous operation are not optional.

Thermal infrared drone image showing solar panel hotspot detection with AI-generated annotations highlighting defective cells
Drone on automated launch pad at edge of solar farm at dawn with propellers starting

Automation

Missions, analysis, and reporting run without constant human presence—while remaining supervised and fully auditable. Drone flight paths are pre-programmed and executed autonomously. Design pipelines process CAD inputs without manual intervention. SCADA anomaly detection operates 24/7 without an operator watching a dashboard. The platform handles routine complexity so your teams can focus on decisions that require human expertise and judgement.

Edge computing device mounted in weatherproof enclosure at a solar site with status LEDs

Intelligence

Real-time detection at the edge combined with model-based reasoning in the cloud. Our AI is not generic machine learning applied to energy data—it is domain-specific intelligence built by engineers who have operated PV assets. Edge inference handles time-critical decisions like anomaly detection and dispatch commands locally, even when cloud connectivity is intermittent. Cloud analytics aggregates fleet-wide learning and continuously improves detection accuracy across all deployments.

Composite aerial view of four solar farm sites across different terrains showing consistent quality

Continuity

Repeatable pipelines, automated checks, and consistent outputs—across sites, teams, and contractors. When you have twenty sites under construction with five different EPC contractors, you need confidence that quality assurance is uniform. Our pipelines produce the same standard of output whether it’s the first site or the hundredth. Knowledge learned on one project is automatically available to the next. This is how you scale operations without scaling headcount linearly.

Solar site central equipment hub showing transformer, weather station, camera and communications antenna

Integration

APIs that connect SCADA, VMS/CCTV, CMMS/ticketing systems, data lakes, and weather services—without reinventing your existing stack. Xskylight is designed to fit into your infrastructure, not replace it wholesale. We integrate with the SCADA systems you already have, the CMMS your O&M team already uses, and the data warehouse your analytics team already queries. Our platform adds intelligence on top of your existing investments, creating value from data that currently sits in disconnected silos.

These four pillars are not independent features. They reinforce each other: automation requires intelligence to be useful, intelligence requires continuity to improve over time, continuity requires integration to work across systems. Together they create a platform that is greater than the sum of its modules.

Proof

A startup with deep field DNA—built by engineers who have designed, built, and operated PV assets at utility scale.

We do not build software in isolation and hope it fits the field. Our platform is informed by years of hands-on experience with the exact workflows it automates: CAD layouts that need to work for EPC contractors, construction verification that needs to satisfy project finance requirements, SCADA systems that need to meet grid code compliance, and maintenance programmes that need to protect asset availability SLAs.

Every engineering decision we make is tested against a simple question: would we trust this output if it were our own asset? If the answer is not an unqualified yes, it does not ship.

Team meeting room viewed through glass wall with solar farm design layout on screen, collaborating on platform development with solar farm data on multiple screens
Split composition showing CAD drawing transitioning into real construction site
Turn design intent into an operational system. No more guessing what’s built—prove it with data. Every layout, every string plan, every component placement is verified against the design model and documented with evidence. The gap between what was designed and what was built becomes measurable, traceable, and auditable.
Platform principle — Design-to-field traceability
Monitor displaying structured audit trail interface with timestamped log entries
Automation is only valuable if the outputs are auditable. Every decision must be traceable to its inputs and the model that produced it. In a world where AI can generate impressive results, the critical differentiator is being able to explain why a particular output was produced—and to prove it was correct. That is the standard we hold ourselves to.
Engineering standard — Auditability by default
Data center server racks through glass door with blinking status LEDs
We treat PV sites like software systems: observability, versioning, and continuous improvement. Every component has state. Every change is logged. Every performance metric is tracked over time. This is how you run a fleet of energy assets at the reliability and efficiency levels that investors and regulators demand—by treating infrastructure with the same rigour as production software.
Operations philosophy — Software-defined infrastructure
Monitor showing thermal image analysis with AI bounding boxes and confidence scores

Engineering-grade AI

Our models are not black boxes. Every detection, classification, and recommendation produced by the platform includes the data it was based on, the confidence level, and the logic chain that led to the conclusion. For thermal hotspot detection, this means the raw IR image, the processing parameters, the classification model version, and the severity assessment criteria are all preserved alongside the result.

This level of transparency is not a nice-to-have. It is a requirement for any AI system that influences decisions on multi-million-pound energy assets. We build to the standard that project finance, insurance, and regulatory stakeholders demand.

Ready for a walkthrough of the full stack?

See how design automation, autonomous inspection, and AI SCADA/EMS connect end-to-end in a live demonstration tailored to your specific use case and operational context.

Book a demo

Company

Xskylight LTD is headquartered in London and serves utility-scale energy developers, EPC contractors, and asset operators worldwide.

We are not a generic AI company that discovered energy as a market. We are energy engineers who discovered that AI could solve the problems we had been living with for years. Our founding team has spent careers designing PV plants, supervising construction sites, debugging SCADA systems, and writing maintenance procedures. This domain depth shapes every product decision we make, from the data models we use to the workflows we automate to the outputs we produce.

Xskylight headquarters in London — a modern tech workspace where engineering and energy domain expertise meet
Sunrise between solar panel rows with golden rays and morning dew

Mission

Make PV and hybrid assets faster to deliver and simpler to operate—by turning deep domain expertise into software that is auditable by default. The energy transition needs more than hardware. It needs software infrastructure that can manage the complexity of designing, building, and operating gigawatts of renewable capacity efficiently, reliably, and transparently.

We exist to build that software. Not as a consulting engagement for each client, but as a product platform that codifies best practices and makes them accessible to every project, every site, every operator.

Multiple laptop screens showing different platform modules in a dark office

What we build

A modular AI platform that unifies engineering, autonomous inspection, and operational optimisation—so every data point flows from design to field to control room and back. Six modules, one data thread, every output auditable. We build for the requirements of regulated infrastructure: enterprise security, multi-tenant isolation, audit trails, and the performance reliability that 24/7 energy operations demand.

Our technology stack spans computer vision, physics-informed machine learning, geospatial processing, real-time telemetry pipelines, and large language models for document intelligence—all purpose-built for the PV domain.

Engineering workspace with code on one screen and edge computing prototype on desk

How we work

Engineering-led, performance-first, and transparent. We design for regulated infrastructure—every output is traceable, every model is explainable, every decision has a data trail. We do not ship features that we would not trust with our own assets. Our engineering culture prioritises correctness over speed, clarity over cleverness, and reliability over novelty.

Based in London, we benefit from the UK’s deep talent pool in AI and renewable energy, proximity to Europe’s largest concentration of energy investment firms, and access to a regulatory environment that demands the highest standards of transparency and accountability.

Aerial view of hybrid energy park with solar arrays, BESS containers and hydrogen electrolyzer

Built for the energy transition

The world needs to install more than 1,000 GW of new solar capacity every year to meet climate targets. The software that manages these assets has not kept pace with the hardware. Design tools have not changed fundamentally in a decade. SCADA systems still rely on threshold alarms. O&M is still largely reactive.

Xskylight is building the software layer that the energy transition needs. Not another monitoring dashboard, but a genuine operating system for utility-scale renewable energy—one that connects every stage of the asset lifecycle into a single, intelligent, auditable platform.

Investors

The AI operating system for utility-scale solar. Here is our story.

~$100M
Target company valuation

Xskylight is positioning to capture a category-defining opportunity in the $8–12B PV lifecycle software market—with a modular AI platform, deep domain moat, and a compounding data flywheel that grows with every deployment.

$8–12B
Addressable market (TAM)
18–22%
Market CAGR
SaaS
Recurring revenue model

The problem we saw

We came from the field. We spent years designing PV plants, supervising construction, and optimising operations for utility-scale solar assets across Europe and beyond. And everywhere we looked, the same pattern repeated itself.

A design engineer in one office produces a CAD layout. A construction manager on site compares it against reality using a clipboard and a camera. An operations team monitors performance through a SCADA system that doesn’t talk to the design model. When something goes wrong, a maintenance technician drives out, takes photos, writes a report by hand, and emails it to three different people.

Every stage of the PV lifecycle is disconnected from the next. Data is trapped in silos. Decisions are made without context. And the industry loses billions each year to design rework, construction delays, undetected faults, and suboptimal operations.

This is not a niche inefficiency. It is a structural gap across a $300+ billion annual market—and it grows wider with every gigawatt installed.

Aerial view of massive solar farm construction site in early stages

The scale of the disconnect

  • Design phase: Layouts are produced in CAD tools that have no awareness of construction sequencing, procurement constraints, or operational performance. Design engineers iterate manually, often producing outputs that need rework once they reach the EPC contractor.
  • Construction phase: Progress is tracked through manual site walks and photo logs. There is no systematic way to compare what was designed against what is actually built. Deviations are caught late—sometimes only during commissioning.
  • Operations phase: SCADA systems collect telemetry, but they lack intelligence. Anomalies are detected by threshold alarms, not by models that understand what the plant should be doing. Maintenance is reactive. Fleet-level insights are virtually non-existent.
  • Documentation: Every project generates thousands of documents—permits, technical specifications, compliance records, inspection reports. Finding the right document at the right time costs projects weeks of delays.

We realised that the PV industry doesn’t need another point solution. It needs a platform that connects design intent to construction reality to operational performance—with AI that understands the entire lifecycle.

What we are building

Xskylight is a modular AI platform that unifies the entire PV lifecycle into one system. Six modules, one data thread, every output auditable.

Our platform is not a collection of disconnected tools stitched together by APIs. It is built from the ground up as an integrated system where design data flows into construction verification, construction data feeds into operational baselines, and operational insights loop back into design optimisation.

The six modules

  • Design Automation—DXF-aware pipelines that produce layouts, routing, string plans, BOQ, and full documentation packages. Every output validated against engineering rules and constraints. What used to take a team of engineers weeks, our system produces in hours—with full traceability.
  • Document Intelligence—Automated due diligence and document management. Our system extracts entities from unstructured documents, detects missing or expired records, and generates structured project context. An AI analyst that has read every document in the project room.
  • Autonomous Site Supervision—Drone and ground-level capture feeds into AI models that produce orthomosaic maps, point clouds, and object-level progress reports. The system compares what is built against the design model and generates evidence packages automatically.
  • AI SCADA + EMS—Unified telemetry ingestion, anomaly detection powered by physics-informed models, generation forecasting, and dispatch optimisation. For hybrid sites (PV + BESS + H₂), the EMS layer optimises across multiple asset types in real time.
  • O&M Analytics—Predictive maintenance that goes beyond simple threshold alarms. Our system diagnoses root causes, prioritises actions based on financial impact, and generates complete evidence packages for warranty claims and insurance.
  • Digital Twin—The connective tissue. A living model that links design intent to as-built reality to operational performance. Every component is traceable. Every change is versioned. Every decision has a data trail.
State-of-the-art renewable energy operations center with curved video wall

Why integration matters

The value of our platform is not just in what each module does individually. It is in the connections between them. When the design automation module produces a layout, the autonomous inspection module knows exactly what to look for during construction. When the SCADA system detects an anomaly, it can trace it back to the design parameters and construction records. When the O&M system recommends maintenance, it references the digital twin to predict the impact on plant performance.

This is what we mean by “one AI platform for the entire PV lifecycle.” It is not a marketing slogan. It is an architectural decision that creates compounding value at every stage.

The market opportunity

Solar PV is no longer an alternative energy source. It is the dominant form of new electricity generation globally:

593 GW
New solar installed in 2024
$300B+
Annual PV investment globally
5.5 TW
Cumulative capacity by 2030

Yet the software that manages these assets has not kept pace. The PV software market is fragmented across dozens of point solutions: one tool for design, another for monitoring, a third for maintenance, a fourth for document management. None of them talk to each other. None of them use AI in a meaningful way.

The global market for PV lifecycle software—spanning design tools, SCADA/EMS, asset performance management, drone analytics, and document management—is estimated at $8–12 billion annually and growing at 18–22% per year, driven by volume, complexity, and regulation.

Hybrid energy park at golden hour with BESS containers and solar trackers

The hybrid transition

The energy market is moving rapidly from pure PV to hybrid installations that combine solar, battery storage, and increasingly hydrogen production. This transition multiplies the complexity of asset management—and it multiplies the value of an integrated platform. Xskylight is built for this complexity from day one.

Why now

1. AI has matured for industrial applications

Computer vision models can now reliably detect defects in thermal images, track construction progress from drone imagery, and extract structured data from engineering documents. Large language models can process unstructured documentation and generate reports that meet engineering standards. These capabilities were experimental three years ago. Today they are production-ready.

2. Edge computing enables autonomous operations

Modern edge devices can run inference models on site, enabling real-time detection and control without depending on cloud connectivity. This is critical for remote PV installations where network reliability cannot be guaranteed. Xskylight uses a hybrid edge-cloud architecture that processes time-critical decisions locally and aggregates fleet-level intelligence in the cloud.

3. The industry is ready

For the first time, utility-scale PV developers and operators are actively seeking AI-powered solutions. The combination of growing portfolio sizes, increasing regulatory requirements, and pressure on margins has created genuine demand for automation. We are not creating a market—we are entering one that is actively pulling for our product.

Industrial hexacopter drone with dual cameras flying over vast solar farm

Our competitive advantage

The four moats

  • Full lifecycle integration—No competitor covers design, construction oversight, SCADA/EMS, and O&M in a single platform.
  • Domain-specific AI—Our models are trained on PV-specific data: DXF layouts, thermal drone imagery, SCADA telemetry patterns, construction photography. Generic AI tools cannot match this accuracy.
  • Auditability by design—Every output our platform produces is traceable to its inputs and the models that generated it. In an industry where decisions affect multi-million-pound assets, this is not a feature—it is a requirement.
  • Compounding data advantage—Each project that runs through our platform improves our models. Design patterns, construction defect signatures, performance degradation profiles—all feed back into the system. A classic data flywheel.

What we are not

We are not a consulting firm that builds custom solutions for each client. We are not an AI research lab looking for a problem. We are a product company with deep domain expertise, building software that we ourselves would have wanted when we were designing and operating PV assets.

Traction and roadmap

~$100M
Target valuation
500+ MW
Design pipeline automated
6
Platform modules in development
2026
First commercial deployments

Where we are today

  • Core AI models for design automation, document intelligence, and autonomous inspection are functional and in testing
  • SCADA/EMS telemetry ingestion and anomaly detection pipeline architecture is complete
  • Active conversations with utility-scale developers, EPC contractors, and asset managers across Europe
  • Platform architecture designed for multi-tenant SaaS deployment with enterprise security

The next 18 months

  • Q2 2026: Pilot deployments with design automation and document intelligence modules
  • Q3 2026: Autonomous site supervision module enters field trials
  • Q4 2026: AI SCADA + EMS module enters beta with early adopters
  • Q1 2027: Full platform available for commercial deployment
  • 2027: Expansion into hybrid asset management (PV + BESS + H₂)
Wall display showing AI thermal analysis dashboard with detection overlays

The team

Xskylight was founded by engineers who have spent their careers in the PV industry. We are not outsiders applying generic technology to energy. We are insiders who understand the workflows, the pain points, and the regulatory landscape from first-hand experience.

Our founding team combines expertise across PV system design, construction management, SCADA engineering, computer vision, machine learning, and enterprise software architecture. We have designed PV plants, supervised construction on site, debugged SCADA systems at 3 AM, and written the maintenance procedures that technicians follow in the field.

This domain depth is our unfair advantage. We build software that solves real problems because we have lived those problems ourselves.

Modern London office building exterior at twilight with warm interior lighting

Our engineering culture

We believe that software for regulated infrastructure must meet a higher standard. Our code is tested, our models are validated, our outputs are traceable. We do not ship features that we would not trust with our own assets. This is not a cultural aspiration—it is how we work every day.

We are headquartered in London, where we benefit from access to the UK’s deep talent pool in AI, renewable energy, and financial technology—and proximity to Europe’s largest concentration of energy investment firms.

The investment thesis

Why invest in Xskylight

  • Massive addressable market—$8–12B annual TAM in PV lifecycle software, growing 18–22% annually, driven by record installation volumes and the hybrid energy transition.
  • No incumbent competitor—The market is fragmented across point solutions. There is no dominant platform that covers the full lifecycle. Xskylight is building that platform.
  • Deep domain moat—Our team’s PV industry experience, combined with domain-specific AI models and an integrated data architecture, creates a competitive advantage that is difficult to replicate.
  • SaaS business model—Recurring revenue from platform subscriptions, with expansion revenue as customers adopt additional modules. High switching costs once the platform is embedded in workflows.
  • Data flywheel—Each deployment improves our models, which improves our product, which attracts more customers. This compounding advantage accelerates over time.
  • Timing—The convergence of industrial AI maturity, edge computing, and genuine market pull creates a window that will not stay open indefinitely. The platform that captures this market in the next 2–3 years will define the category.

Use of funds

  • 60% Engineering: Complete and harden all six platform modules, including edge inference and hybrid asset support.
  • 20% Commercial: Fund pilot programmes with strategic early adopters, build sales engineering capability, establish partnerships.
  • 15% Operations: Scale the London team with senior hires in AI/ML, PV domain expertise, and product management.
  • 5% Legal & compliance: IP protection, ISO 27001 certification, and regulatory alignment across target markets.

Our ask

We are raising capital at a target valuation of approximately $100 million, reflecting the scale of the market opportunity ($8–12B TAM), the depth of our technology moat, and the category-defining potential of a full-lifecycle AI platform in the fastest-growing segment of energy infrastructure.

We are looking for investors who understand infrastructure technology, have patience for enterprise sales cycles, and recognise that the energy transition needs better software—not just more hardware. If you invest in deep-tech companies that solve real industrial problems, we should talk.

Join us in building the future of solar.

We are building the infrastructure that will power the energy transition. If you invest in deep-tech companies solving real industrial problems, let’s talk.

Talk to the founders

Insights & resources

Technical deep-dives, case studies, and engineering perspectives on AI in utility-scale energy.

We believe that the PV industry benefits from open knowledge sharing. Our engineering team regularly publishes perspectives on the technical challenges of applying AI to real-world energy infrastructure—from the practical difficulties of training computer vision models on thermal drone imagery, to the architectural decisions behind building a platform that serves both edge and cloud workloads. Below is a preview of topics we are exploring. Subscribe for early access as these resources become available.

Close-up of screen showing document extraction interface with highlighted regions and structured data

AI for PV document intelligence

How we built a RAG-powered system that can extract structured data from thousands of unstructured engineering documents—permits, technical specifications, compliance records—and make them searchable, cross-referenceable, and auditable in seconds instead of weeks.

Technical deep-dive
Computer screen showing ML training interface with grid of thermal images and classification markers

Thermal defect detection at scale

Training computer vision models on thermal infrared drone imagery for solar panel defect classification. We discuss the challenges of data labelling, model generalisation across different panel technologies, and achieving production-grade accuracy for warranty-grade evidence.

Computer vision
Split composition of outdoor edge computing cabinet at solar site connected to cloud data center

Edge-cloud architecture for energy AI

Why we chose a hybrid architecture that runs time-critical inference at the edge and aggregates fleet intelligence in the cloud. Covers the trade-offs between latency, connectivity reliability, model update deployment, and the specific requirements of remote PV installations.

Architecture
Xskylight platform overview showing the unified data pipeline connecting all six modules

Stay ahead of the curve

The intersection of AI and utility-scale energy is moving fast. New techniques in computer vision, physics-informed machine learning, and edge computing are unlocking capabilities that were not possible even two years ago.

Subscribe to receive early access to our technical publications, case studies from pilot deployments, and invitations to engineering webinars where our team shares what we are learning in the field.

Contact

Tell us what you’re building. We’ll respond with a technical path, relevant module recommendations, and next steps.

Whether you are a utility-scale developer looking to automate design pipelines, an EPC contractor seeking objective construction verification, an asset manager exploring AI-powered O&M, or an investor evaluating the platform opportunity—we want to hear from you.

Every inquiry receives a personalised technical response from our engineering team. No sales scripts, no generic brochures—just a direct conversation about how Xskylight can address your specific operational challenges.

Panoramic view of a utility-scale solar installation at golden hour, showcasing the scale and precision of modern PV infrastructure