INNOVATIONTerraSky.ai / CLIMATE TECH
InvestmentDrones and proprietary AI reveal hidden solar-farm losses, turning analysis into actionable repair tasks.
Official website
Photograph / illustration: TerraSky.ai
IDENTIFIED PRODUCTION OPPORTUNITY
5–17%We typically identify at least 5% and up to 17% production-improvement potential, including several percent at regularly inspected farms.
An experience-based range, not an independently audited performance guarantee. Realised output depends on the faults found and repairs completed.What could this mean nationwide? ↓TerraSky combines proprietary image processing with engineering analysis to identify lost production and the work needed to recover it.
RGB and thermal images, panel-level identification and site data work together to reveal losses that a basic visual inspection cannot explain. Deeper analysis brings entire additional categories of fault into view.
We use proprietary image-processing AI, developed specifically for solar-farm inspections and substantially more effective than language models available to everyone.
Outputs locate the fault, describe it and estimate the associated loss. That supports priorities and an actionable task list, backed by maps and panel-level documentation for the field team.
Autonomous flight and repeatable data collection provide a consistent survey across the full site. This supports later analysis and comparisons over time.
A survey every two years. Repairs completed. Generally at least 5% recovered production. We used the full 2025 calendar year to estimate the annual value across Hungary’s industrial solar farms. Household systems are excluded.
4,746.1 MW · year-end installed capacity. MW measures power; MWh measures energy produced.
≈ 8.44 billion HUF / year
The benefits of these repairs recur every year over the next decade, as they bring lost production capacity back into operation.
≈ 42.20 billion HUF / 5 years
Five times the annual value, assuming unchanged annual output and prices.
5% assumption. The monthly table uses the same additional output rate.
We valued MAVIR’s 2025 monthly industrial output at generation-weighted market prices and the selected improvement rate, then converted the annual total to HUF at the ECB annual average exchange rate. The weighting uses a quarter-hour generation profile, so cheap and negative summer midday prices reduce the result in proportion to actual production. [1–4]
This is an illustrative gross value opportunity, not a guaranteed return or net national saving. Survey, repair and financing costs have not been deducted.
Electricity is already cheap in summer and at midday
Summer midday prices can be zero or negative, but not every day. We valued MAVIR’s monthly output using 15-minute generation profiles and the day-ahead price for each interval. Daytime energy is not valued at expensive evening prices. [2–3]
Swipe sideways for more data ↔
13:00–14:00 → 21.16 EUR/MWh. 20:00–21:00 → 198.37 EUR/MWh. June–August averages in Hungarian local time. All 281 negative-price hours in the year are included; they are not replaced with zero.
Swipe sideways for more data ↔
| Month | Installed capacityMW | Monthly outputGWh | Daily averageMWh |
|---|---|---|---|
| Jan | 4,094.9 | 195.81 | 6,317 |
| Feb | 4,204.3 | 379.64 | 13,558 |
| Mar | 4,389.2 | 539.82 | 17,414 |
| Apr | 4,441.9 | 728.86 | 24,295 |
| May | 4,582.4 | 749.60 | 24,181 |
| Jun | 4,608.6 | 981.24 | 32,708 |
| Jul | 4,641.7 | 894.24 | 28,846 |
| Aug | 4,657.7 | 852.54 | 27,501 |
| Sep | 4,676.8 | 618.01 | 20,600 |
| Oct | 4,680.7 | 484.14 | 15,618 |
| Nov | 4,692.7 | 238.31 | 7,944 |
| Dec | 4,746.1 | 139.85 | 4,511 |
| Full year | 4,746.1* | 6,802.06 | 18,636 |
* Year-end capacity, not an annual mean. Daily output is the monthly total divided by calendar days. Average monthly output across the year is 566.84 GWh. This is one completed calendar year, not a long-term weather average or a 2026 forecast.
Swipe sideways for more data ↔
| Month | Market averageEUR/MWh | Solar-weighted priceEUR/MWh | Gross value of 5%thousand EUR |
|---|---|---|---|
| Jan | 140.19 | 123.29 | 1,207.1 |
| Feb | 158.88 | 127.99 | 2,429.4 |
| Mar | 109.02 | 67.13 | 1,812.0 |
| Apr | 85.46 | 40.31 | 1,469.0 |
| May | 80.89 | 35.17 | 1,318.2 |
| Jun | 84.10 | 39.05 | 1,915.7 |
| Jul | 102.56 | 71.67 | 3,204.4 |
| Aug | 80.51 | 41.55 | 1,771.0 |
| Sep | 101.93 | 61.66 | 1,905.3 |
| Oct | 122.03 | 91.40 | 2,212.4 |
| Nov | 124.35 | 104.03 | 1,239.6 |
| Dec | 115.88 | 105.10 | 734.9 |
| Full year | 108.51 | 62.39 | 21,219.0 |
The market mean is duration-weighted. The solar-weighted price is an estimate: each interval is weighted by its output. It is not an official HUPX index or a plant’s contractual selling price. Values are calculated before rounding.
Monthly weighted price = Σ (MWh × EUR/MWh) / Σ MWh
We matched 35,040 quarter-hour generation observations with the effective hourly or quarter-hourly price. Each monthly profile was scaled to MAVIR’s revised monthly energy total. This is an estimate anchored to official monthly data, not a new set of interval measurements.
Annual value = Σ (monthly MWh × weighted price) × 5%
Conversion: 1 EUR = 397.7675 HUF · ECB, 2025 annual average [4].
We assume that identified faults are repaired and the production benefit persists between surveys. The value then accrues annually; it is not halved merely because surveys occur every two years. Initial nationwide rollout timing is not modelled.
The assumed uplift follows the same generation profile. The model adds no new capacity and does not store energy for the evening.
Only MAVIR’s industrial PV series is included. Household (HMKE) and own-use (SCTE) systems are excluded. 5% is the lower end of TerraSky’s stated 5–17% experience range, not a verified nationwide result. The public sources below do not establish this recovery rate.
Costs, grid constraints, price effects of extra supply and changes in imports are not modelled. Market value is not the same as subsidised or contractual revenue, or net national savings.
CHECKABLE EVIDENCE
Data checked on 21 September 2026. Output and capacity are official data; weighted prices and values are our calculations. The studies provide context for the method.
2025 monthly industrial capacity: PDF page 3. Monthly measured output: PDF page 5. Additional check: Hungarian electricity system data for 2025 ↗.
Source of the 15-minute intramonth profile. Monthly totals are aligned to MAVIR; the two series differ slightly.
Hourly, then quarter-hourly prices matched to production times. Negative values are retained. CC BY 4.0 ↗.
A consistent conversion basis for HUF values, not a current exchange rate.
Price check: the report’s March daily-index mean of EUR 108.90/MWh is reproduced by the data. Our duration-weighted mean is 109.02; the difference comes from the weight of the 23-hour daylight-saving transition day.
Independent regulator cross-check: annual HUPX DAM average of EUR 108.51/MWh, matching our annual market mean.
Government presentation on battery storage. Slides 4–5 illustrate low or negative daytime prices and evening imports. Context, not the monthly calculator dataset.
Statisztikai Szemle, 103(7), 648–673, 2025. DOI: 10.20311/stat2025.07.hu0648. Research context for daily profiles and the distinction between household, own-use and industrial generation.
Research on the market value of morning and late-afternoon output, using a German example. Related developer perspective: aream · east–west orientation, 2025 ↗. No revenue premium from these sources is applied to Hungary.
Hungarian example of a new market-based project: the developer states that electricity will be sold without a support scheme or corporate power purchase agreement. It includes storage and is not part of the 2025 calculation.
Contact
I’m ready to build with you and your company—whether that starts with a conversation over coffee or takes shape through mentoring, innovation advisory or investment.
Looking for investment