For the first time in a long time, what one can see looks like good financial management, perhaps due to clarity of actions taken, and certainly to factors that matter to investors and markets. At the end of the day, those will enable the company to stay in the running more than the mode of selling modules at prices below the cost. More so, some of the companys statements suggest that tier-1 companies can no longer support liquidation sales, and there is an active pursuit of price stability, even if it means having to refuse orders.
We note that Trina came in short by 70 MW toward the lower end of its original Q2 call guidance, and we were almost right predicting shipments around 378 MW, versus the actual 380.3MW. What we got wrong was the 13% Chinese aspect of the revenue shown in the supplemental earnings presentation. Approximately $20M of the total $298M revenue was a sale of cells into the Chinese market, something which has not often been done by Trina in the past.
Those sales were made at break-even gross margins, and they were done to reduce inventory levels and to increase cell capacity utilization. Taking this into account, we estimate the revenue was in the $277M range for approximately 380MW modules, or around $0.72 per watt. Not bad, considering that Jinko Solar sold their modules at $0.68. Still, Jinkos costs more easily accommodated such an ASP, and the price difference indicates that China dominated Jinkos sales, versus large majority of shipments overseas by Trina.
Certainly, fewer dollars are tied to Jinkos inventory, shielding it off from fluctuations due to market pricing. Jinkos capacity utilization, during lower demand this period, saved additional amortization costs. Trinas non-polysilicon cost of $0.54 per watt increased 2 cents this time due to underutilization. When it comes to polysilicon, Trina was able to reduce it to $0.13 per watt, while Jinko paid $0.12.
From the perspective of cost of goods sold for the period, Trinas numbers had a bit more than the usual complexity. First there was a $25.8M reversal of provisions made in preparation for AD and CVD duties during Q1 of 2012 and Q4, 2011. This benefit was reduced by the 13.3M provision for the non-cash inventory write off. The remaining provision reversal of $12.5M reduced COGS to $295M. Taking those effects out, the cost was $308M. Adjusting for the cost of sales of cells, the cost of revenue ended up being in the region of 288M or $0.75 per watt. The difference of $0.08 per watt from declared processing costs of $0.67 per watt, have been described by the company as inventory effects.
During the Q3 conference call, Trinas management had disclosed that in Q2 the company had 85% of the inventory in finished goods, then having approximate cost of $395M. At the end of the third quarter, Trina had approximately 75% of inventory in finished goods, according to management, worth $275M out of total of $367M. The estimated dollar reduction for finished goods was in the range of $120M. The companys September 30, balance sheet indicated $96M decrease from Q2, implying that approximately $20M worth of materials were added during the quarter. On its recent Q3 call, the company acknowledged its utilization rate was about 50%.