South, east-west or vertical: which solar park layout pays off for investors

Anyone reading an offering document for a solar park direct investment usually finds the layout in a single line: south, 20 degrees tilt, or east-west, 15 degrees. Few investors ask about it, yet that line fixes the hours in which the park delivers its electricity, and since 2025 also how many of its kilowatt hours fall into hours with a negative exchange price. The market has two standard answers: south for maximum yield, east-west for a better market value. The second answer is only half right. The calculation shows why, step by step.

What sets south, east-west and vertical apart?

Three layouts, three daily profiles. The classic ground-mounted park points its modules south, today mostly at 20 to 25 degrees with tight row spacing; it harvests the most irradiation and generates half of its annual output between 11 a.m. and 2 p.m. The east-west park splits its capacity across two rows tilted flatly against each other, usually 10 to 15 degrees; row shading disappears, the land can be packed more densely, and the profile widens slightly in the morning and evening. The third layout stands upright: bifacial modules that generate on both faces, in rows facing east and west. They produce in the morning and afternoon while the midday sun strikes the module edge at a flat angle. Fraunhofer ISE describes exactly this arrangement as a complement to the predominantly south-facing fleet; as agrivoltaics it leaves room for farming between the rows, the basics of which Investing in agrivoltaics: dual land use, revenues and tax leverage at a glance explains.

How much yield does leaving south cost?

Around 13 percent for flat east-west and, on our model assumption, around 3 percent for vertical bifacial modules. For the calculation we pulled hourly profiles from PVGIS, the photovoltaic calculator of the European Commission's Joint Research Centre, for a site in central Germany, the weather year 2023 and 14 percent system losses. The south-facing park at 25 degrees comes to 1,011 kilowatt hours per kilowatt; that matches the roughly 1,080 full-load hours Fraunhofer ISE quotes for a new, flat south-facing park without curtailment. The east-west park at 15 degrees reaches 876 kilowatt hours. For the vertical park PVGIS only calculates the front face; we add the rear face with a bifaciality factor of 0.9, because modules with values above 90 percent are available, and arrive at 984 kilowatt hours. That is an assumption, not a measurement: the rear-side yield depends on ground reflection, row spacing and mounting height, and row shading is not modelled.

Two points of context belong here. The site weighs more than the tilt: the same south-facing park delivers 10 percent more in southern Germany than in the centre of the country, more than the difference between 25 and 35 degrees. And anyone who wants to raise the yield rather than shift it finds the lever in tracking: single- or dual-axis trackers lift the specific annual yield by 15 to 30 percent according to Fraunhofer ISE; they are not modelled in this article. Which of these assumptions applies to a specific project is written in the yield assessment, which Reading a yield assessment: what P50 and P90 mean for a solar park direct investment explains how to read.

Does east-west really earn more per kilowatt hour?

At a flat tilt, practically not. For the market value we multiplied every hour of the profile by the day-ahead price of the same hour from SMARD, the Bundesnetzagentur's market data portal, and divided by the electricity generated; the result is the capture price the park earns on average per kilowatt hour. With 2023 prices the south-facing park reaches 72.9 euros per megawatt hour, the east-west park 72.6. With 2025 prices it is 47.0 against 46.7 euros. The difference is below one percent, in both years to the disadvantage of east-west. The reason sits in the table above: at 15 degrees both module halves see the midday sun almost as well as the south-facing park, and the midday share only drops from 50 to 48 percent. The profile is not shifted, merely compressed by 13 percent.

The vertical park is different: its midday share is 30 percent, and its capture price is 84.7 euros per megawatt hour with 2023 prices and 63.1 euros with 2025 prices, 16 and 34 percent above south respectively. The gap grows with every year in which the midday hours get cheaper. The yardstick is the market value factor, the ratio of the price achieved to the unweighted annual average price: for the entire German solar fleet it has fallen, by our own calculation from SMARD data, from 0.76 in 2023 to 0.52 in 2025; the resulting figure, the Marktwert Solar, is also the second input of the market premium, as Direct marketing and the market premium: how a solar park earns its money shows. Our model value for 2025, 4.61 cents, sits close to the official annual Marktwert Solar of 4.51 cents per kilowatt hour on netztransparenz.de; the difference comes from hourly instead of quarter-hourly prices and from the way generation is extrapolated.

There is a historical reason why east-west counts as the market value layout. In a short study for Agora Energiewende in 2014, Fraunhofer ISE modelled an east/west scenario with around 20 percent higher market value; but the single plant assumed there delivered 19 to 22 percent less yield than south, so it was tilted far more steeply than today's usual 10 to 15 degrees. The study also noted that, under pure EEG marketing at the time, an investor had no incentive for east-west. Both still hold: the market value advantage comes from the tilt angle, not the compass direction, and it only reaches the investor if the marketing route reflects it. In a power purchase agreement with an industrial buyer the profile is in any case part of the negotiation; what matters there is set out in Understanding PPAs: how power purchase agreements make solar park revenues predictable.

What happens in the hours with negative prices?

They hit the south-facing and the flat east-west park equally hard and the vertical park much less. In 2025 SMARD counted 573 hours with a negative day-ahead price, almost twice the roughly 300 hours of 2023. Laid over those hours, the south-facing park has 25 percent of its annual output in negative-price hours, east-west likewise 25 percent, the vertical park 17 percent. As a plausibility anchor: for the entire German solar fleet the same calculation from SMARD generation data gives 24 percent. Because our profile comes from the weather year 2023 and the prices from 2025, the model tends to understate the hit rate: negative prices arise on sunny days, and the correlation with the actual weather of the price year is missing from the calculation.

For plants commissioned since 25 February 2025 there is no market premium in those hours under §51 EEG; the electricity is sold at the negative exchange price or curtailed. What that means for solar park revenues, and how the extension of the support period under §51a EEG partly offsets it, is covered in Negative electricity prices: what they mean for solar and storage investors. For the layout a simple rule follows: every kilowatt hour that moves from midday into the morning or afternoon is one kilowatt hour less exposed to negative-price hours. The vertical arrangement achieves that without a battery; a battery at the same grid connection achieves it with south-facing modules, as Co-located vs. stand-alone: which gives the better risk structure shows.

What does that mean per euro invested?

A clear revenue lead for vertical, a surcharge on the cost side, and the sign of the difference depends on the price year. With the applicable value of 4.79 cents per kilowatt hour from the July 2026 auction, the market premium on the self-calculated monthly market value and no premium in negative-price hours, the south-facing park earns around 54 euros per kilowatt and year with 2025 prices, the east-west park 47 euros and the vertical park 71 euros. With 2023 prices the Marktwert Solar was above the applicable value, the market premium fell to zero, and the three values were 74, 64 and 83 euros. The vertical park's lead is therefore 13 percent in the old and 30 percent in the new price structure; the east-west park trails by its 13 percent yield shortfall in both cases.

Against that stand the construction costs. In its levelised cost study of July 2024, Fraunhofer ISE puts ground-mounted parks above one megawatt at 700 to 900 euros per kilowatt. For the upright design there is no equally broad market statistic, but two robust reference points from the same institute's agrivoltaics guide: the mounting structure of ground-level agri-PV systems costs 97 to 167 euros per kilowatt against 76 euros for the ground-mounted park, and bifacial glass-glass modules are put at 326 euros per kilowatt in the guide's example, noticeably above the standard module. Translated into levelised cost, the guide gives 6.0 cents per kilowatt hour for ground-level agri-PV against 5.4 cents for the ground-mounted park, a surcharge of around 11 percent. Then there is the land: a vertical park as agrivoltaics occupies, on Fraunhofer's figures of 0.25 megawatts per hectare, four times as much land per kilowatt as a dense south-facing park; rent per kilowatt rises accordingly but, according to the guide, is shared between farming and plant operation. For east-west the 2014 Agora study estimated around 5 percent lower system costs than for south, among other things for inverters, cables, fencing and mounting; no more recent independent figure exists.

Roughly, that means: a revenue lead of 30 percent against a cost surcharge of around 10 to 15 percent carries the vertical design in the 2025 price structure on paper; in the 2023 price structure, with a 13 percent revenue lead, it would have been a zero-sum game. Anyone deciding for 20 years is therefore betting on whether the midday hours keep losing value or whether battery build-out and flexible demand lift them again; both forces are real, and how batteries currently earn from the midday dip is worked through in 2-hour or 4-hour battery: which storage duration pays off for investors. Flat east-west does not justify itself through revenue but only through land and grid connection: more kilowatts per hectare, no row shading, a flatter profile at the feed-in point.

Why does a high DC/AC ratio cost kilowatt hours but hardly any revenue?

Because the clipped hours are precisely the cheapest ones. The DC/AC ratio describes by how much module capacity exceeds inverter and thus grid connection capacity; Fraunhofer ISE calls values above one typical and describes the limitation as permanent internal curtailment that raises the full-load hours of the grid connection. In our model the south-facing park at a ratio of 1.4 loses around 0.8 percent of its yield but only 0.3 percent of its revenue with 2023 prices and practically nothing with 2025 prices, because the clipped midday peaks in 2025 fell almost exclusively into hours with prices around or below zero. At 1.6 it is 3.1 percent of yield against 1.7 and 0.4 percent of revenue. The east-west park loses only half a percent of yield even at 1.6, the vertical park almost nothing, because its peaks are lower.

For investors that is good news with a checking task attached. The good news: a project that oversizes its grid connection with modules gives away less than the yield losses suggest, and scarce grid capacity is used better. The checking task: the yield assessment must include the clipping, and the provider's revenue model must not value the clipped hours at average prices. If the assessment shows a yield before clipping and the revenue model an annual average price, both figures are too high.

What does the EEG 2027 change about the calculation?

According to the government draft, little about the layout question and a lot about the support logic. The draft EEG 2027 (Bundestag printed paper 21/7867 of 7 September 2026, first reading from the week of 21 September, entry into force planned for 1 January 2027) replaces the one-sided market premium for new plants from 100 kilowatts with a two-sided mechanism: if the annual market value is below the applicable value, the market premium is paid as the difference; if it is above, the operator pays a refinancing contribution to the grid operator under §21d. What matters for the layout is the basis of both: under Anlage 1 it is the technology-specific annual Marktwert Solar, the average of the entire fleet, not the revenue of the individual plant. A plant whose profile earns more than the solar average keeps that difference in both directions. The vertical design's profile advantage therefore stays with the operator under the new system too; it even becomes the only lever through which a project can still influence its revenue beyond the applicable value.

For projects under way, the transition under §100 of the draft applies: plants commissioned before 1 January 2027, or whose applicable value was determined in an auction round before that date, remain under the EEG as in force on 31 December 2026. Everything else is draft status; figures, sections and dates can still change during the parliamentary process.

Which questions should you ask the provider about the layout?

• Which orientation and tilt are in the yield assessment, and do they match the detailed design and the construction drawings?
• Does the revenue model use the Marktwert Solar, an annual average price or the capture price of the plant's own profile, and for which price year?
• What is the DC/AC ratio, and is clipping reflected in both the yield and the revenue?
• For bifacial modules: which bifaciality factor and which rear-side yield are assumed, and is row shading included in the assessment?
• How many kilowatts per hectare are installed, and how is rent per kilowatt calculated?
• Are the negative-price hours under §51 EEG explicitly deducted, and with how many hours?
• If east-west was chosen: for which reason, land, grid connection or market value, and can that reason be found in the revenue model?

How we check projects for their layout

For every solar park we present to investors, we take orientation, tilt and DC/AC ratio from the detailed design and run the plant's hourly profile against the price series of recent years, with the same disclosed model as in this article. We reconcile the provider's market value assumption with the capture price of the plant's own profile, deduct the negative-price hours explicitly in the economic calculation, and require the clipping to be in the yield assessment. We give no return promises; we show which figure rests on which assumption. What makes a solar park a direct investment overall is set out in Investing in solar parks: revenues, costs and tax leverage at a glance. That is exactly what we check with you in a no-obligation initial consultation, gladly using a specific offering document.