Hacker News Reader

Comment by ChuckMcM | Hacker News Reader

Always over until it starts again. Note that the party is over (in this case) because it is threatening smaller businesses, not because there is a lack of capacity or that the "big guys" couldn't still make money at the reduced rate. In every industry where I have been aware of this this situation coming up (voluntary reduction in production in order to shore up prices), it works for a while and then people "cheat" and ramp up their production to make a bit more money with their now dormant capacity. And then more people cheat, and the party is back on again. The Chinese are notoriously fierce in their desire to make a buck and so I give this one about a 5% chance of lasting through the end of 2019.

View on HN

The solar industry is evolving fast enough that most dormant capacity left behind by bankrupt smaller producers probably isn't worth reactivating even if module prices rebound. These producers don't just have smaller annual capacities. They also have less sophisticated quality control, less automation, and older, slower generations of the automated tools they do have.

Module assembly is simple enough that some of those lines could be reactivated in response to demand, but module assembly has always been the easiest part. Smaller wafer and cell manufacturers that go bust probably won't see their equipment resurrected anywhere. Newer capacity from the likes of JinkoSolar, LONGi, and GCL yields more wafers per ingot and more efficient cells from wafers, more consistently, at lower costs.

Jinko's polycrystalline cells are so thin as to be almost impossible to keep from breaking, something on the range of 150um. I've got a few sitting here on my desk. Even an ultra-low pressure vacuum pick and place breaks them half of the time.

The mechanical fragility of thinner silicon has broadly stalled the progression toward even thinner wafers/cells. It's a pity since the optically/electronically optimum thickness would be thinner than is common now, and thinner wafers would yield more wafers per ingot.

Do you know how Jinko is handling such thin wafers when they assemble their own modules? If they're making cells of such thinness they presumably have found a way to do it; nobody can afford to accidentally break half of their cells on the way to modules.

Probably use an Electrostatic wafer chuck.

This is a well-studied phenomenon known as Cartel Theory: https://www.cambridge.org/core/journals/business-history-rev...

"The pain will mostly be felt by smaller Chinese producers, which lack international supply chains, triggering industry consolidation or forcing them to close, he added"

So it looks like capacity is being reduced.

Will capacity increase again? Probably. But your mention of cheating seems to be suggesting it's a cartel, I don't think that view is supported by the article, so have you got any other evidence?

Why would consolidation result in a lower capacity? All that's happening is bigger companies becoming bigger while a mix of gov policy and market forces during downturns squeezes out the little guys. Which is standard practice in modern state-capitalist societies.

Bigger companies aren't going to buy up smaller production that is not efficient. The smaller production just goes offline and gets scrapped. Solar panel production is at the point where you need to be doing economies of scale to really churn a profit at it.

You guys need to read up on what happens when a government subsidies something. The subsidies allows inefficient producers to stay in business and it also results in over production....which is what pushes the price artificially low. Once the subsidies end. The weakest go out of business and only the most efficient producers stay in business.

If I run a firm with lots of capacity and a thriving overseas trade, why would I buy up a competitor with no overseas business just to shut down their plant?

Because you will be buying it for less than the cost of the equipment in that plant + some of the best talent + their domestic customer list thrown in for free.

predict: China will increase their cap on solar production in 2019 to at least 20 GW. https://predictionbook.com/predictions/194304

Congratulation on the prefiction, they've already done since they started the year with a capacity of ~125GW.

200GW would make sense however.

Often, one sees industries making changes because market forces have made the transaction cost of manufacturing too onerous. Some industries in that case will consolidate and vertically integrate. If the pain from uncertain prices and supply gets too great, could the solar industry start with mergers and vertical integration? Or is it far too early for that?

Yes, mergers have been common in solar for a while.

This is one of the top manufacturers today:

https://www.hanwha-qcells.com/

A Chinese cell and module manufacturer acquired a Chinese ingot manufacturer. Korea's Hanwha Group subsequently acquired this merged Chinese entity as well as the German solar manufacturer Q-Cells. The merged Hanwha Q-Cells is vertically integrated from cells to modules.

A handful of manufacturers like GCL have vertically integrated even further:

- Refined silicon production

- Ingot production from purified silicon

- Wafers from ingots

- Cells from wafers

- Modules from cells

I don't know if that depth of integration makes sense long-term or not. Some companies have been burned in the past as they integrated with producers that weren't close enough to the leading edge. All of these steps are evolving rapidly in small increments; a vertically integrated producer has to keep pretty close to the leading edge at every stage if they don't want to see their integration advantages reverse.

I suspect vertical integration works when your key intermediate manufacturing inputs aren't commodities. And you have enough scale to overcome production granularity issues.

I suspect vertical integration works when your key intermediate manufacturing inputs aren't commodities.

What if your manufacturing outputs are also getting to be commodities? The example I was thinking of was refrigerated cargo containers. The inputs are obviously commodities. The outputs are also getting to the point where the cost of financing to build stock for sale or lease is getting to be onerous for smaller players.

Vertical integration to the point where they could acquire cheap land songwriters to optimistically cover with surplus production? This could be an interesting difference to other markets with an overcapacity cycle, e.g. you can't just hook up a month's output of DRAM to the network and have an income stream.

Reply on news.ycombinator.com