Raycatch (DeepSolar) alternative for AI solar diagnostics (2026)

DeepSolar is a cloud-based automated AI diagnostics platform for PV plants. It ingests production data from a plant’s existing monitoring systems, runs daily analysis, and produces a diagnostic mapping plus a prioritized, ROI-ranked task list down to string level, explicitly with no new hardware, software, or site visits. Raycatch material claimed more than 4 GW of solar assets analyzed, an identified 4 to 5 percent of average recoverable yield, and up to 30 percent O&M cost savings, and the product won an Intersolar Award. Diagnostic coverage spans cleaning recommendations, MPPT and inverter efficiency, disconnections, PID, and underperforming panels under warranty.

These are public facts, not commentary. Raycatch was founded in Tel Aviv in 2015 and raised about 6.1 million dollars, backed among others by BayWa r.e. Energy Ventures. In November 2022 its assets were acquired by BladeRanger, an Israeli robotic panel-cleaning company, for 1.5 million dollars. In March 2025 BladeRanger sold DeepSolar to PainReform Ltd., a NASDAQ-listed company that was previously a clinical-stage pharmaceutical business, for 4.5 million dollars in shares and securities. DeepSolar now operates as PainReform’s solar unit; it joined the NVIDIA Connect program in August 2025 and announced a forecasting module, DeepSolar Predict, in pilots as of November 2025. As of August 2026, raycatch.com does not load, while deepsolar.ai is live but publishes no customer roster, track-record numbers, or pricing. For an operator signing a multi-year analytics contract, vendor continuity belongs in the diligence checklist alongside the feature list.

The core idea is genuinely good and the deployment model is frictionless: full compatibility claims across monitoring systems and hardware, no IT project, and a daily ROI-ranked action list that field teams can execute directly. The diagnostic breadth from production data alone is wide, and the claimed impact figures came from analysis across 75 utility-scale plants totaling 1.2 GW. A portfolio owner who wants automated daily diagnostics dropped on top of an existing monitoring stack, with minimal integration effort, is squarely in DeepSolar’s original design center.

Same software-only category, different construction. NuraVolt publishes its model accuracy as open benchmarks on public datasets, per class and per horizon, where DeepSolar’s current site publishes no verifiable accuracy evidence. Soiling is treated as an economic decision, estimated per inverter after digital-twin weather and temperature correction and fed into cleaning-schedule optimization, not only flagged as a cleaning recommendation. NuraVolt also covers what DeepSolar does not attempt: BESS health, warranty, and revenue analytics, an O&M ticketing workflow, contract-obligation tracking, and an MCP server that makes the fleet queryable from AI assistants. Pricing is public, and the company is the operating business itself rather than a unit of a listed company from another industry.