Following yesterday’s perspective, capital expenditures for 2024 may not be formally finalized until next year.
If there is a manufacturing recession in 2024, capital expenditures (in the polysilicon sphere, especially for ingots, wafers, cells and modules) could be a leading indicator of this recession. Capital expenditures are often directly correlated with economic downturns, which the PV industry has seen before.
The chart shows industry-wide capital expenditures (excluding polysilicon) for PV manufacturing since 2005. Aside from gross margins, manufacturers’ upcycles and downcycles tend to closely correlate with capital expenditure activity, with a higher correlation than perhaps any other manufacturing or financial metric.
By going back to 2005, I have successfully tracked the two major cycles the PV industry has gone through over the past 20 years; essentially the entire commercial lifecycle of the PV industry thus far. I divide the timeline into different periods, similar to how the first chart in this article described the upcycles/downcycles from a cost and average selling price perspective for polysilicon and modules.
Referring to the chart above, from left-to-right. Firstly, prior to 2009, the PV industry was largely still in a growth phase, having emerged from an R&D-based cottage industry beforehand. Capital expenditures were almost entirely about improving production levels for a small handful of Western firms. Polysilicon was borrowed from the semiconductor industry. Overcapacity did not exist, and while manufacturing margins were not amazing, there was some order to things with investment and growing market demand.
By 2009, the whole industry was on the cusp of a downturn, or at least a flatlining of polysilicon capital expenditures. At the time, capital expenditure surges for thin-film technologies like amorphous silicon (a-Si) and copper-indium-gallium-selenium (CIGS), which were seen as potential mainstream alternatives to incumbent silicon-based manufacturing techniques, cushioned the transition from upcycle to downcycle to some extent.
The 2010-2011 capital expenditure spike was primarily driven by display equipment suppliers (likes of Applied Materials, Oerlikon and ULVAC) selling turnkey a-Si production lines to start-up firms. These efforts largely failed to achieve their aims, and the capital expenditure surge during this period is a legacy issue, as shown on the chart.
Ultimately, the downturn did of course materialize, as outlined earlier in this article. Indeed, the subsequent industry evolutionary phases (as outlined in the first chart here) correlate precisely with the capital expenditure cycles shown in the chart above: the doldrums of 2012-2013; the uptick from 2014-2020; and the looming downturn, delayed until 2023.
The 2020-2023 capital expenditures make no sense from any perspective, and are on a completely unsustainable trajectory. If this article was just about capital expenditures, alarm bells would be ringing for 2024. This is when capital expenditures will go negative.
However, predicting capital expenditures is extremely tricky. Ultimately we can only track funds in real-time. And with most equipment orders now landing at Chinese firms, metrics linking orders to shipments are impossible to ascertain. During the first industry downturn, the order-shipment ratio was the final (capex-driven) leading indicator, forewarning of the difficult 2012-2014 period for manufacturing.
This could be tracked, since Western public equipment suppliers were under heavy scrutiny, and these suppliers were the lynchpins of PV manufacturing at the time, with their (actual) order levels (and backlog) closely audited. A decade ago, order-shipment ratio analysis was feasible: today, it is not.
Who will be the ‘winners’ during a PV manufacturing downturn?
In a general sense, PV manufacturers do not build, operate or own solar projects. Some firms partake at the final development and early construction stages, but even these companies soon flip or need to cash-out to long-term asset owners. So we can rule out PV manufacturers benefiting from industry-adjacent activities, thereby gaining secondary revenue streams with module prices (and the entire value-chain pricing levels) depressed.
Of course, the winners are those buying modules, at least in the short term. However, buying modules (even at low prices) is not without its issues, as there are increased risks from the financial health of the suppliers themselves compared to business-as-usual. What’s the value of a 25-year warranty? Will quality be compromised as everyone looks to cut corners?
If half of the components in the modules are coming from new entrant firms presently losing money and with questionable long-term survival, what incremental risks might exist? And so on. So yes, low module pricing is great, but the risks from bankruptcies across manufacturing outweigh the opportunities for anyone in the sector.
Potential winners also include module suppliers that are owned by cash-rich conglomerates. These firms see enough strategic value in having PV manufacturing to effectively bankroll (more than during good times) their PV manufacturing vision. Truth be told, there are very few firms that fit into this category, and these firms are usually not found in the top 10 PV module suppliers by shipment volumes. The top 10 module suppliers dictate pricing and market-share terms for everyone.
Who else potentially wins during the recession? Stock market plays? Notably, several module sales firms that are public entities still have ‘Buy’ recommendations against them. It will be interesting to see if this remains the case come March 2024; at least before my articles scare some folks! (disclaimer)
Technology disrupters. Yes, indeed – I contend that it may take an extended PV manufacturing recession from 2024, lasting throughout 2025, for truly disruptive technologies to enter the market, even more urgently than is the case today. There is a concept that companies innovating during down periods end up in leadership after markets recover.
Of course, this is vague and non-specific, but I hope readers get the gist. For most other technology spheres, time-to-market with product innovation and roll-out is everything. Innovating in down periods is not just an option but an imperative. Even so, there are absolutely no guarantees recovering markets will need the innovative products.
But for PV specifically, can the manufacturing recession (from 2024, lasting throughout 2025), finally end up being the catalyst to drive technology change and therefore set the industry firmly on course for a strong capital expenditure phase again in 2026: and not just be about building more polysilicon plants in China and more ingot pullers?
In this respect – and somewhat provocatively in line with my ‘innovate-or-die’ mantra – the single biggest opportunity arising from a 2024-2025 PV manufacturing recession (centred on China more than anywhere else), is for the entire Western PV community to take stock and get behind start-ups, technology-leaders and manufacturing eco-system plays, with an Apollo-type mandate to get them over the line before 2026.
Brace yourself for a bumpy 2024, anyway! Let’s see which firms reset expenditure goals and market-growth visions first.
