This article is the first of two parts discussing some of the business trends and metrics regarding Chinese solar companies listed in the US markets. In this first part, I will talk about the upstream providers which are generally less integrated than their downstream counterparts (TSL, STP, YGE, SOLF) who sell modules to end demand markets. Two primarily cell producers, JASO and CSUN, along with two primarily wafer producers, LDK and SOL, are discussed here.
The second quarter, which ended last week, was very chaotic for the Solar Industry. On one end, global demand for solar products is exploding. Many solar companies who can produce and offer low cost products are over booked and sold out. However, since most of the demand came from Europe, it left many companies vulnerable as the euro depreciated over 9% during the quarter. This is in part why I wanted to talk about the upstream names listed above first, since they have less end demand exposure and thus have less of their earnings directly impacted by the decline in the euro during the quarter. Another main advantage of being a provider within the value chain are lower operating costs, most notably lower sales and marketing expenses which generally does not scale linearly with shipment volume. These factors generally apply to all four companies discussed here.
The numbers used are metrics directly stated by the companies, or numbers that I have derived by a series of statements companies have said. They may be correct, partially correct, or completely incorrect. As an outsider without reference to their books, I can only make estimates much like anyone else. Use these examples as a general reference only for understanding each company and some of their near term business trends, and not for speculation purposes. Also keep in mind I make a general assumption - companies will meet high end shipment estimates. Because demand is so strong such that all these companies are stating they could ship more if they could produce it, I'm going on the assumption that if they state they could ship a certain volume, it's likely to be achieved in this demand environment. I also cross reference this number with what I estimate actual capacity during the quarter is likely to be at from previous capacity statements. Thus, these estimates may be on the high side, but still very reasonable.
The cell producers: JASO & CSUN
In the past, both jaso and csun were single verticals in the crystalline solar food chain, offering solar cells produced from purchased wafers, for consumption by end module production. Since prices across the board dropped after the credit crisis, gross margins on a per watt basis compressed, causing many single verticals to seek further integration which benefits from economies of scale. Csun has recently entered into an agreement to purchase module capacity to become a two vertical cell/module producer. Jaso has expanded both upstream by adding wafer capacity, as well as downstream by adding module capacity; as a result, they are partially fully integrated. However since much of this process is new, the total benefits won't be realized for another few quarters. For now, their primary source of gross profits will be from cell production and sales.
Jaso is one of the largest cell producers in the world and based on their processing costs, among the lowest if not the lowest cost producers. As a result despite falling prices, they have been able to generate a decent level of per watt gross margins once input costs normalized and selling prices stabilized. They should also see minor incremental gross profits from some of the integration steps they have recently taken. They also benefit from very good procurement costs due to scale and relationships. Unlike many peers who might see foreign exchange losses, jaso might actually post a gain. In part, this is because they have extremely little exposure to the euro, offset by gains from the usd depreciating against their functional currency the rmb, on their usd denominated convertible debt. They should also see a small gain from mark to market accounting on derivatives linked to their convertible bonds. For operational purposes, these numbers can be factored out, but they will affect US GAAP numbers.
JASO Q2 2010:
Shipments: 210mw cell, 50mw oem, 35mw module
Asps: 1.28/watt cell, .35/watt oem, 1.66/watt module
Unit Costs: 1.0/watt cell, .18/watt oem
Gross Profit: 210 x .28 = 58.8m cell, 50 x .18 = 9m oem
Incremental Gross Profit: 35 x .05 = 1.5m module, 30 x .15 = 4.5m wafer
Gross Margin: 58.8m + 9m + 6m = 73.8m / 345m = 21.4%
Operating Expenses: 14m
Net Interest Expense: 7m
Forex Gain: 1m
Derivative Gain: 5m
Net Income: 49.8m
Share Count: 165m
Csun is the smaller cell producer with a more spotty record. At face value, they should report quite nice numbers, but there have always been factors that have caused overall numbers to be lower. For example, in the prior quarter, they appeared to have negative margins from other revenues outside their core cell business. This may also be the case for the second quarter, but for the example I used, I assume a neutral effect from non-core cell revenues. They also have an outstanding legal case with REC which may cause legal expense in any given quarter to be higher than expected. As a result, I would view these estimates as less reliable, although very achievable without unforeseen costs.
CSUN Q2 2010:
Shipments: 80m cell, 2m module
Asps: 1.28/watt cell, 1.65/watt module
Unit Costs: 1.03/watt cell, 1.65/watt module
Gross Profit: 80 x .25 = 20m cell
Gross Margin: 20m / 105.7m = 18.9%
Operating Expenses: 7m
Net Interest Expense: 2m
Net Income: 9m
Share Count: 44m
The wafer producers: LDK and SOL
Much like their cell producing cousins, both sol and ldk were single vertical providers in the solar value chain. Both have taken very aggressive steps in integrating their business by adding cell and module capacity. In part, it's because some of their customers have requested those services as the trend to outsource intensified. As with jaso and csun, the other half of their integration plan is to maintain a higher level of per watt gross margins within the value chain as prices fell. After the credit crisis hit, wafer producers were among the hardest hit, as upstream pricing fell much faster and to a much higher magnitude than downstream pricing. It was not until very recently that gross margins within the wafer vertical started to normalize after high inventory costs had been blended down and asps stabilized. Both companies have polysilicon capacity but for the examples I used, I assume a neutral earnings effect from their polysilicon division. One last factor to note is both companies have high debt levels which may result in secondaries that could dilute future earnings.
Sol and ldk are very similar in structure, but differ in scale. Their operational metrics are very similar as well. In terms of core wafer capacity, sol has half the capacity as ldk. In terms of polysilicon capacity, at a nameplate level and at full ramp, sol has less than a fifth of ldk's capacity. Since debt comes with expansion in the case for many of these companies, sol also has a third of ldk's debt load. Sol recently updated second quarter guidance, so not much of this is a surprise unless some undisclosed expense came up along with the final numbers. As with the cell producers discussed above, being primarily a single vertical results in very lean operating expenses which should result in very high operating margins.
SOL Q2 2010
Revenues: 255m: 172m wafer, 83m module
Shipments: 260mw: 210mw wafer, 50mw module
Asps: .82/watt wafer, 1.66/watt module
Unit Costs: .58/watt wafer, 1.15/watt in house module, 1.33/watt cell outsourced module
Gross Profit: 210 x .24 = 50.4m wafer, 30 x .51 + 20 x .33 = 21.9m module
Gross Margin: 50.4 + 21.9 = 72.3m / 255m = 28.4%
Operating Expenses: 15m
Net Interest Expense: 5.5m
Forex Loss: 2m
Net Income: 38.1m
Share Count: 87m
Again, ldk is very similar to sol except on a larger scale. As a result, they have higher interest costs which will be a weight on earnings until they can further scale polysilicon production to higher levels. In terms of integration, they are slightly behind sol in that they did not have cell capacity up until the end of the prior quarter. Instead they use their sister company's cell capacity for outsourcing. This results in lower margins on their module volumes. While the numbers may look different from sol's metrics, they are actually almost identical if one adjust for efficiency assumptions. Sol recently did a recalibration on their efficiency assumptions, while ldk hasn't. In effect, it raises sol's capacity on a megawatt basis, but lowers their costs, selling price, and per watt gross margins. This doesn't do anything on a dollar basis, but does bump up stated gross margins on a percentage level up a couple hundred basis points. On a non-operating level, ldk may continue to receive government subsidies as they have in the past, which could take the bottom line slightly higher. They also have a high usd denominated convertible bond balance like jaso which along with the rmb appreciation against the usd, will result in minor foreign exchange gains. This is offset by losses from their euro exposure however.
LDK Q2 2010
Shipments: 550mw: 380mw wafer, 100mw wafer oem, 70mw module
Asps: .87/watt wafer, .43/watt wafer oem, 1.66/watt module
Unit Costs: .67/watt wafer, .25/watt wafer oem, 1.36/watt module
Gross Profit: 380 x .20 = 76m wafer, 100 x .18 = 18m oem, 70 x .30 = 21m module
Gross Margin: 76 + 18 + 21 = 115m / 490m = 23.5%
Operating Expenses: 24m
Net Interest Expense: 30m
Forex Loss: 2m
Net Income: 50m
Share Count: 131m
As we look forward into the third quarter, I think the results should be somewhat similar if not slightly better. The companies in question have recently stated pricing has been stable while input costs on a blended level should be stable if not slightly lower. This should result in similar per watt gross margins in the third quarter for their core business. In addition, much of their capacity expansion will be activated by the end of the second quarter, so shipments should increase sequentially since they are still stating they are operating at sold out conditions. Sol should have about 300mw wafer capacity for Q3, so shipments sequentially could increase by 20%. Ldk's wafer capacity should only increase modestly by about 10%. In both cases however, they should have higher integrated capacity which will result in incremental gross profits. Jaso could also see sequential shipment gains of about 20% as well as higher integrated capacities. In addition, jaso will also ship a new higher efficiency product that could increase selling prices and margins. Csun should be shipping on a two vertical basis in the second half. As a result, the outlook should be pretty good, and Q3 should be the high quarter of the year for these companies. As prices may retrace around 10% in the fourth quarter, earnings may drop back to levels seen in the second quarter with higher volumes making up for lower per watt margins.
Disclosure: Long JASO, LDK, SOL