作为半导体的分支 ，光伏行业也呈现专业化和一体化两种道路的分野 。从方向来看 ，光伏行业从上游到下游的产业链各环节具有相当程度的差异性 ，所需要的核心能力不尽相同。此外由于硅片 、电池 、组件等环节的技术更新仍旧没有停止 ，全产业链承受的技术路线风险会大大高于专注于某一环节的专业化厂商 ，因此未来光伏行业大概率会走半导体的专业化道路 。且按下两种道路不谈 ，一体化道路也分为两种路径 ：一是晶科为代表的从组件往硅片做 ，二是隆基股份为代表i的从硅片往组件做 。从产能分布来看 ，晶科一开始会呈现组件大硅片小的金字塔型结构 ；隆基则呈现硅片大于组件小的倒金字塔型分布 。两种路径的区别在于 ，晶科是通过自建硅片和电池减少供应商的供货 ，逻辑是依据组件的产能和销售情况来建上游产能 ，相对来说整条循环最终是 “以销定产 ” ，组件出货量决定了上游产能的建设情况 ，是一条完整通顺的链条 。因为终端客户在手中 ，所以整个链条是有出口的 ，不容易积压库存 。隆基则是硅片产能最大 ，产量减去外销部分决定了自建电池和组件的产能 。

        然而 ，随着组件和电池产能的扩大 ，隆基越来越威胁到下游客户的市场份额 ，这对下游客户来说面临一个问题 ，就是采购隆基的硅片就是给隆基提供子弹来在电池和组件端与自己竞争 。这在单晶硅片紧缺时期危害尚不明显 ，一旦需求低迷或供给过剩 ，客户将不会优先采购隆基的产品也不会支持隆基所推广的166标准 “也就是不会进入你的体系 ，只会买市场标准产品 ” 。

因此 ，隆基只能继续扩大电池组件产能来消化自己的无法外销的硅片产能 ，而继续扩大组件产能会加剧上述矛盾的进一步不可调和 。如此循环往复 ，隆基的组件产能会出现一个巨大的问题 ，那就是产能建设不是根据销售能力来定 ，而是根据硅片的冗余度来定 。从而在市场低迷的时候出现巨大的库存 ，这在531和近期的市场表现中已经反应的淋漓尽致 。原因在于david所述的 ，组件的销售需要公司品牌的多年积累以及渠道方面一城一地的艰难争夺 ，不可能跟得上产能的巨大扩张 ，也就是 “以产定销 ”是不现实的 。

        那么结论不难推定 ，销售跟不上产能会导致巨大的库存堰塞湖 ，结合一年来行业技术的日新月异 ，而在这个过程中隆基已成了那个守成而不是颠覆的角色。因此 ，隆基是否有可能是即将被高库存和技术落后双重挤压而爆炸的巨人呢 ？

       As a branch of semiconductor, the photovoltaic industry also shows specialization and integration of the two roads.From the perspective of direction, the PV industry from the upstream to the downstream of the industrial chain has a considerable degree of difference, the required core capabilities are not the same. In addition, because the technological update of silicon wafers, batteries, modules and other links has not stopped, the risk of the whole industrial chain will be much higher than that of specialized manufacturers focusing on a certain link. Therefore, in the future, the photovoltaic industry is likely to take the path of semiconductor specialization.In addition, the integration path is divided into two paths: one is from components to silicon wafers represented by Jinko, and the other is from silicon wafers to components represented by Longi.From the perspective of capacity distribution, Jinko will start with a pyramid-shaped structure with larger components and smaller wafers;Long groups show an inverted pyramid distribution in which the wafer is larger than the component. The difference between the two paths is that Jinko reduces the supply of suppliers by building silicon wafers and batteries by itself. The logic is to build upstream production capacity according to the capacity and sales of components. Relatively speaking, the whole cycle is ultimately "production based on sales", and the shipment of components determines the construction of upstream production capacity.Because the end customer is in hand, the whole chain is exported, and it is not easy to accumulate inventory.Longi, on the other hand, is the biggest producer of silicon wafers. Production minus exports determines the capacity of its own cells and modules.

However, with the expansion of module and battery capacity, Longi is increasingly threatening the market share of downstream customers, which is facing the problem that purchasing Longi's silicon wafers is providing Longi with bullets to compete with them on the battery and module side. This is not obvious in the period of shortage of monocrystal silicon, once the demand is depressed or oversupply, customers will not give priority to purchase Longi's products and will not support the 166 standard promoted by Longi "that is, will not enter your system, will only buy market standard products".

As a result, Longe will have to continue to expand its battery module capacity to absorb its unsaleable silicon wafer capacity, which will exacerbate the contradictions. The big problem with this cycle is that Longi's component capacity is built not by selling capacity, but by the redundancy of the wafer.This has resulted in large inventories when the market is depressed, as has been evident in the 531 and recent market performance. The reason is that, as stated by David, the sales of components require years of accumulation of the company's brand and the tough competition from one city to another in terms of channels, so it is impossible to keep up with the huge expansion of production capacity. That is to say, it is unrealistic to "set sales by production".

It's not hard to conclude, then, that sales can't keep up with production capacity, which will create a huge inventory barrier lake. Combined with a year of technological change in the industry, Longi has become an established rather than a disruptive player in this process. So, is it possible that Longi is a giant about to explode, squeezed by both high inventory and backward technology?