After more than a decade of ineffective protectionist “sticks”, the IRA’s pivot to “carrots” is a step in the right direction. — by Joe Song


For many years now, US solar manufacturing has been largely unable to compete with international manufacturers, despite the numerous tariff-driven policies aimed at strengthening the competitive landscape for US based manufacturers. All the while, demand in the US market has grown rapidly, largely satisfied by international manufacturers, as the tariffs failed to increase domestic supply. In a highly welcome tactical deviation, the Inflation Reduction Act (IRA) utilizes incentives – “carrots” if you will – aimed at growing domestic manufacturing by providing meaningful economic motive for all stakeholders. We will discuss how we arrived at the current state of supply-demand conditions and how the IRA provisions may fuel increased domestic production. Ultimately, this paper’s primary focus is on determining the approximate amount of domestically manufactured crystalline silicon (c-Si) modules that a spectrum of solar projects may need to qualify for the Domestic Content incentive.

How did we get here?

In the decade leading up to 2010, solar PV manufacturing was the greatest strength of the US renewable energy industry. It was during this time when historically important technologies and companies first arrived and whose contributions were critical in enabling the opportunity we find today. Not only was production capacity sufficient to supply the US market, but the US was also a consistent net-exporter of those goods. The following decade, however, proved to be a much different story. The emergence of Chinese manufacturers who entered the US market, impeccably timed, was too great a force and resulted in monumental and beneficial (for most) shifts. The speed and scale at which Chinese manufacturers were operating led to incredible cost and price reduction. In 2011, US module pricing started the year around $1.80/W and ended the year in the $0.90/W range, as one example of how disruptive the conditions were at this time.

In an attempt to protect and grow US manufacturing, the first of multiple countervailing (CVD) and anti-dumping (AD) petition were filed, beginning in 2011, led by the German company, SolarWorld. The use of tariff measures has been the singular method utilized by the multitude of events ever since, including the current pending petition filed by the consortium led by Auxin. Each tariff event caused significant damage and heightened risk – arguably more so during the time leading up to the final ruling as the industry is paralyzed by uncertainty. Predictably, the US market adjusts and moves towards stability following each event, only to be punched in the face with another threat – creating an ongoing cycle of stability and instability since 2011. And despite efforts to bolster US manufacturing, these measures failed sidesplittingly to improve manufacturing conditions[1].

Figure 1: US Manufacturing Under Tariff Conditions

Section 201 did manage to produce some results, as seen in Figure 1, where modest gains were achieved by the end of 2020. These gains are largely attributed to new Hanwha and Jinko facilities, expected to produce up to 5 GW per year. It’s important to note that these new capacity gains are confined to module assembly only, meaning that only a fraction of costs are influenced by US conditions, namely labor costs. These new facilities are valuable additions[2], especially for the residential market; however, the potential impact of Section 201 has been realized at this point.

Figure 2: Example Module Cost Comparison

For most of the US market, there remains a void of realistic options. The commercial and utility solar sectors are anticipated to be as much as 45 GW by 2027. First Solar (“FSLR”), as one of the largest domestic solar module manufacturers, is expected to increase domestic production capacity from 3 GW to approximately 11 GW by 2027. This is incredible and should be celebrated. However, many of us have witnessed a handful of large purchasers, through a series of headline grabbing announcements, have more or less reserved all capacity through 2026. This leaves most of the industry, well, still without a realistic option.

So, how might supply and demand-side stakeholders work to fill the void in domestic c-Si PV modules? Well, it’s going to require cooperation, commitment, and policy that the industry can reliably plan around. It seems as if we now have the policy. Demand-side stakeholders need to accept higher priced PV modules (without economic penalty) and supply-side stakeholders need economic conditions that compel domestic CapEx investments. Said another way: they both need carrots.

IRA Manufacturing Provisions

The Supply-Side Carrot: IRA Manufacturing Provisions

The IRA provides for $30 billion in Manufacturing tax credits (MTC) and an additional $10 billion in manufacturer Investment tax credits, both of which are applicable to a broad range of technologies and sub-sectors. Manufacturers can only utilize one or the other, meaning that a manufacturer who applies for MTC incentives, would not be eligible for ITC benefits. For our purposes, we’ll focus on MTCs, specific to those applicable to c-Si solar PV module manufacturing[3].

Table 1: US Solar Manufacturing Capacity

The IRA provides different incentive levels for different manufacturing processes involved in module production, shown in Table 1. For purposes of evaluation, let’s take a typical 72-cell 545W bifacial Mono-PERC module composed of M10 wafers (182mm) as a baseline. As mentioned previously, existing domestic c-Si module manufacturing is confined to assembly only. So, we know that at a minimum, the $0.07/W incentive is most attainable. This is shown as “Option 1” in Figure 3. Subsequent scenarios show the build-up of different incentives on a per-Wdc basis with increased vertical integration at each step. Even if in the near-term manufacturers leverage only the assembly credit, that alone is significant enough to close the economic gap between domestic and imported modules.

Figure 3: IRA MTC Incentives

Option 4 shows the maximum level of incentives available from polysilicon production to module assembly. Although $0.17/Wdc of incentives is available, it seems unlikely that Option 4 is meaningfully adopted as the incentive for poly production is comparably small, and current US poly production is negligible. The incentives for wafers and cells production are more valuable and at the very least, will be strongly considered.

Figure 4: Announced US Manufacturing

Already, early results are nothing short of astonishing. In just a few months since passage of the IRA, a multitude of manufacturers have announced plans to establish new US manufacturing facilities, not only increasing total module production capacity, but more importantly, reshoring wafer and cell manufacturing. One might reasonably expect this trend to continue, with additional manufacturers investing in new domestic facilities, adding both topline capacity and strengthened supply chain control. While acknowledging that a lot will change between now and 2025 (the timeline in which any new facilities may be realized), it’s getting easier to imagine a promising domestic manufacturing market. We’re off to a good start.

The Demand-Side Carrot: IRA Domestic Content Provisions

The Domestic Content tax credit bonus is the corresponding incentive provided for demand-side stakeholders, designed to substantially offset the increased cost of domestically produced components. There is much to be learned and revealed in pending final guidance issued by Treasury/IRS, but for now, we are aware of the primary qualifiers. Domestic Content eligibility is achieved so long as the following conditions are met:

  1. Requirement that 100% of steel and iron should be produced in the United States and done so consistent with the regulations under the “Buy America Act.”
  2. Meet or exceed a threshold percentage of the total costs of manufactured products (including components) in a facility are mined, produced, or manufactured in the United States. Thresholds start at 40% for projects beginning construction before 2025, rising in 5% increments to 55% after 2026.

This framework falls short of levels required for action-oriented decision-making, of course. One has to assume the (very splashy, headline grabbing) announcements of huge purchase orders are riddled with caveats and qualifiers that look to final rules/regulations/guidance. The current comment period hopefully results in clear and objective guidelines. In the meantime, we’ll have to look to precedent and use a bit of deduction. The relevant regulations under the Buy America Act provide that all steel and iron manufacturing processes must take place in the United States, except metallurgical processes involving refinement of steel additives. The regulations further provide that steel and iron requirements apply to all construction materials made primarily of steel or iron, but do not apply to steel or iron used as components or subcomponents of other manufactured products. In Section 26 USC §45(b)(9)(B)(i) of the IRA, reference is made to "Buy America" 49 CFR 661, related to Federal Transit Authority (FTA) procurement for bus depots and train stations. One could potentially draw analogies from those regulations to ascertain when such a manufactured product will be treated as produced in the United States. Under those regulations, for a manufactured product to be considered produced in the United States:

  1. All of the manufacturing processes for the product must take place in the United States
  2. All of the components of the product must be of U.S. origin. A component is considered of U.S. origin if it is manufactured in the United States, regardless of the origin of its subcomponents.

In applying the regulations and provisions to solar components, the “end product” here is the PV system, and “components” at a very minimum include modules, inverters, batteries, and piers. It is unknown how racking will be treated, where the example of single-axis trackers has multiple potential outcomes. One can argue that assembly of trackers in the field constitutes manufacturing of a “component” and if found agreeable, then the source of those subcomponents is made irrelevant, meaning tracker costs could be fully recognized in meeting Domestic Content requirements. However, if such argument is not found agreeable, then the individual sub-components (torque tubes, rails, controllers, etc) will determine the cost eligibility of trackers, most likely around 50% assuming they are manufactured domestically.

Project Scenarios

Building upon previously discussed US market conditions, the state of US domestic manufacturing, and preliminary Domestic Content requirements, we will need to establish assumptions to fill the void of unknown IRA conditions and the project level “control/baseline” for which Domestic Content analysis will be applied.

Qualitative assumptions:

  1. Racking costs will be eligible (including piers), minimum of 50% if not 100%
  2. Electrical BOS (E-BOS) components will be eligible
    1. Utility projects: 67% of DC and 25% of AC
    2. Commercial: 50% of DC and 25% of AC
  3. String inverters will be manufactured domestically, central inverters are imported
    1. Large utility projects (see Scenario 1 below) only use central inverters
    2. Small utility and commercial projects utilize string inverters
  4. All projects will have qualified for 30% ITC as a baseline, and subsequently meets labor-related wage requirements. Qualifying for the Domestic Content provision enables a 40% ITC.

To substantiate the assumptions here, I assume that utility projects will utilize “Big Lead Assembly” (BLA) type solutions as manufactured domestically by Shoals, and that these components will qualify for Domestic Content qualification. While commercial projects are less prone to utilizing these types of solutions, I assume that these projects will at least use combiner boxes and other domestically sourced E-BOS components.

Quantitative assumptions

Table 2: Project Model Cost Assumptions

To begin estimating how many domestically produced modules future projects will need to qualify for the bonus ITC adder, I have used the cost assumptions in Table 2[4], approximated using a combination of recent verbal quotes and extrapolating broader component costs trends. We can debate the merit or “correctness” of these cost assumptions if that’s the sort of thing you enjoy doing, but bear in mind that for the purpose of this analysis, the relative domestic and imported costs are all that really matter.

Determining Eligible Cost Basis and Relative Added Cost

We analyzed four project scenarios, as shown in Table 3, based upon what is currently known about the IRA provisions and using placeholder assumptions to fill unknown conditions. While we will go into greater detail of only Scenario 1, we summarize the results across all Scenarios and construction years (ranging from 2024 to 2026 and beyond) using the same methodology.

Table 3: Project Scenarios

Using IRA guidance and the assumptions already outlined, a 2024 Large Utility project would have a cost stack of $0.646/Wdc (excluding civil, labor, and other items out of the denominator for domestic content purposes), from which we can identify the eligible cost basis needed, seen in Figure 5. To qualify for the domestic content bonus ITC, 40% of the incurred/relevant project costs need to qualify as domestic content – in this case, that is $0.258/Wdc.

Figure 5: Cost Basis Estimation

After backing out racking and E-BOS costs (per above, we assume those are domestically manufactured), we are left with $0.076/W of component cost that must be covered by domestic content to reach the 40% threshold (Figure 6). The domestically manufactured PV modules “plug” needed is 18 MW (13.8%) which adds $0.021/W compared to a project with only imported modules. If only 50% of racking costs are eligible, then the project will need to procure approximately 36 MW (27.7%) of domestic modules, a cost adder of approximately $0.048/W.

Figure 6: Module Cost Target

Projecting this out through 2027, assuming full tracker cost eligibility, a Large Utility project will have an increased dependence on domestic modules to meet the increasing cost threshold requirements, growing to nearly 37 MW by 2027. However, despite the need to procure increasing amounts of domestic modules to qualify for the bonus ITC, neither the gross nor relative cost adder grows significantly. This, of course, is largely attributable to expected component cost reduction, which one must believe will occur via increased production capacity, competitive forces, and the (mostly undefeated) learning rate effect.

Figure 7: Module Target By Year

The matrix below shows the full spectrum of scenarios. The key difference between Scenario 1 – explored in depth above – and the scenarios in the table below, is the inclusion of domestically procured inverter costs for the cost stack, which of course reduces dependence on domestic modules to achieve the minimum cost threshold. As mentioned previously, we are assuming that string inverters will be manufactured domestically, as opposed to central inverters which for Scenario 1, are assumed imported. The reasoning for this key difference is based simply on the notion that the barriers are lower due to fewer components, higher volumes, and prior market precedents.