What project owners need to know.
The crystalline silicon solar value chain has five stages: polysilicon, ingot, wafer, cell and module. They are often discussed as one industry, but each is a different business with different capital intensity, different skills and different competitive dynamics. India's capacity is distributed very unevenly across them. According to ICRA analysis reported in September 2026, module capacity stood at 217 GW as of August 2026, cell capacity at 35 GW, and ingot and wafer capacity at about 2 GW from a single completed facility. A NITI Aayog report puts import dependence at close to 100% for polysilicon, above 90% for wafers, above 60% for cells and above 40% for modules. The country has built the chain from its downstream end, which is the easiest to enter and the least defensible, and the ALMM framework is now working backwards up the chain to correct that.
Five stages, five different businesses
Polysilicon purifies metallurgical silicon into feedstock of very high purity. Ingot growth turns that feedstock into a single crystal, usually by the Czochralski method. Wafering slices the ingot into wafers. Cell processing turns a wafer into a device that generates current. Module assembly turns cells into a laminated, framed, warrantable product.
They sit in a line, which makes them look like one industry. They are not. Capital intensity, process difficulty, energy consumption and the nature of the competitive moat change at every step. Polysilicon is a chemical plant. Ingot growth is a slow, energy-intensive crystallography problem where yield is decided by furnace conditions. Wafering is precision machining at enormous volume. Cell processing is semiconductor-adjacent, with the equipment intensity to match. Module assembly is a lamination and quality-control operation.
This matters commercially because the barriers are not evenly spaced. Moving downstream from wafer to module is comparatively straightforward. Moving upstream from module to wafer, or wafer to polysilicon, means entering a genuinely different business each time — which is exactly what India is now attempting.
- Treat each stage as a separate investment case, not a segment of one
- Note where energy, purity and yield replace assembly throughput as the constraint
- Expect upstream moves to require new skills, not just new machines
- Check which stage a supplier or partner is actually credible in
India built the chain from the wrong end
The growth has been real and fast at the downstream end. NITI Aayog records module capacity rising from roughly 2.3 GW in 2014 to 100 GW by August 2025, and cell capacity from under 1.2 GW in 2014 to 25 GW by March 2025. ICRA analysis reported by pv magazine India in September 2026 puts module capacity at 217 GW as of August 2026 and cell capacity at 35 GW, up from 13 GW in July 2025, with cells projected to reach 100 GW by December 2027.
Upstream, the picture is different. The same analysis identifies about 2 GW of completed ingot and wafer capacity from a single facility. That leaves module capacity at roughly six times cell capacity and roughly a hundred times ingot and wafer capacity. Polysilicon has no meaningful domestic capacity at all, and reporting in 2026 describes MNRE weighing a dedicated support scheme to back more than 10 GW equivalent of polysilicon manufacturing.
An inverted chain of that shape has a specific consequence: nameplate module capacity is not the same as domestic value addition. A module assembled in India from an imported cell, cut from an imported wafer, grown from imported polysilicon, is an Indian module by assembly and an import by value.
- Read module nameplate capacity as assembly capacity, not value addition
- Track the cell and wafer ratios, not the headline module number
- Expect the cell ramp to pull hard on wafer demand well before 2028
- Watch polysilicon policy — it is the last unaddressed stage
Import dependence, stage by stage
The NITI Aayog report quantifies the dependence directly: close to 100% for polysilicon, above 90% for wafers, above 60% for cells, and above 40% for modules measured on available supply after exports. It also notes that nearly 59% of India's photovoltaic cell and module imports in 2024 came from China.
The gradient is the story. Dependence falls steadily as you move downstream, which is the mirror image of where capacity has been built. Every gigawatt of new domestic cell capacity increases wafer demand that domestic wafer capacity cannot currently meet, so upstream import dependence does not fall automatically as downstream capacity grows — in the short term it can rise.
For a project developer or manufacturer reading this as a market map, the practical implication is that the scarcity is upstream and the competition is downstream. That is unusual, and it is temporary. It is also why the policy instruments have started moving in the other direction.
- Map your exposure stage by stage, not as a single import share
- Expect downstream growth to raise upstream import volumes near term
- Treat wafer supply as the binding constraint for new cell capacity
- Model supply risk on the stage you actually depend on
The global backdrop that sets the price
None of this happens in isolation. The 17th edition of the ITRPV, drawing on data from 38 companies and research institutions across the value chain, reports global module shipments stabilising at around 706 GW in 2025, and global crystalline silicon manufacturing capacity at the end of 2025 exceeding 1,230 GW for polysilicon, ingot and wafer production, 1,260 GW for cells and 1,460 GW for modules.
Capacity at that scale, well ahead of annual shipments, is what keeps global pricing pressure on every stage. It is also why domestic manufacturers in a newly building market face a structural cost comparison rather than a temporary one. ICRA's assessment, as reported, is blunt on this point: domestically manufactured components are unlikely to be competitive against Chinese imports on cost alone.
That does not make domestic upstream capacity a bad investment. It means the investment case has to rest on something other than beating a global cost curve — policy-protected demand, supply security, proximity to customers, or integration across stages. Anyone building a plant should be able to say which of those their case rests on.
- Do not build an upstream case on beating global cost alone
- Name the actual basis of the case: policy, security, proximity or integration
- Track global capacity against shipments as the pricing backdrop
- Stress-test the plan against continued oversupply, not against recovery
What the policy clock actually requires
ALMM is the instrument working backwards up the chain. List-I covers modules. List-II for cells came into force on 1 June 2026, and enlisted cell capacity had passed 30.3 GW by the seventh revision as of 30 April 2026, with the MNRE ALMM page recording a ninth revision on 21 August 2026. List-III extends the framework to wafers from 1 June 2028; reporting of the proposed requirement describes at least three wafer manufacturing units with aggregate capacity of 15 GW per annum.
Read as a sequence, the intent is clear enough: make each stage progressively harder to satisfy by import, starting downstream and moving up. For anyone planning capacity, the useful question is not whether the deadlines will hold exactly, but which stage their business will be standing in when the requirement reaches it — and whether the qualification work, which takes far longer than construction, has started.
Current requirements should always be confirmed directly with MNRE rather than taken from summaries, including this one, as revisions have been frequent.
- Work back from the date the requirement reaches your stage
- Start supplier and product qualification well ahead of construction
- Assume revisions — check the live MNRE position before committing
- Plan for the stage above yours to be scarce when your capacity lands
How JRST supports this requirement
JRST works on the upstream half of this chain — ingot and wafer manufacturing and the equipment, materials and consumables around them — helping projects test which stage they should be entering, what that entry demands technically rather than commercially, and how to sequence supplier and product qualification against the policy timetable. Figures on this page are drawn from public reports, industry roadmaps and policy reporting and should be validated for the specific project; capacity figures move with each revision and compliance requirements must be confirmed with the relevant authorities.
Discuss your requirement
Mohammed Saif Zaveri connects JRST's industrial content to execution conversations.
As Co-Founder and designated partner of JRST Technology LLP, Mohammed Saif Zaveri works across industrial growth, strategic partnerships, client conversations, equipment strategy, and project pathways for silicon, solar, semiconductor, and advanced-manufacturing opportunities.
This knowledge page is part of JRST's public industrial knowledge base, designed to help buyers move from search terms and early research toward a structured technical-commercial discussion.
View Mohammed Saif Zaveri's profileConnect on LinkedInFrequently asked questions
What are the stages of the solar value chain?
Polysilicon, ingot, wafer, cell and module. Polysilicon purifies silicon feedstock; ingot growth produces a single crystal; wafering slices it; cell processing turns a wafer into a working device; module assembly laminates cells into a finished product.
How much solar manufacturing capacity does India have?
ICRA analysis reported in September 2026 puts module capacity at 217 GW as of August 2026, cell capacity at 35 GW, and ingot and wafer capacity at about 2 GW from one completed facility. Polysilicon has no meaningful domestic capacity.
Why is India's solar manufacturing described as inverted?
Because capacity is concentrated at the easiest, most downstream stage. Module capacity is roughly six times cell capacity and roughly a hundred times ingot and wafer capacity, so most upstream value is still imported.
How dependent is India on solar imports?
A NITI Aayog report gives close to 100% for polysilicon, above 90% for wafers, above 60% for cells and above 40% for modules on available supply after exports, and notes nearly 59% of 2024 cell and module imports came from China.
How large is the global solar manufacturing base?
The ITRPV 17th edition reports global module shipments of about 706 GW in 2025, against end-2025 capacity exceeding 1,230 GW for polysilicon, ingot and wafer, 1,260 GW for cells and 1,460 GW for modules.
When does ALMM apply to wafers?
List-III is set to extend ALMM to wafers from 1 June 2028. Reporting of the proposed requirement describes at least three wafer manufacturing units with aggregate capacity of 15 GW per annum. Confirm current terms with MNRE, as revisions have been frequent.
How can JRST help across the value chain?
JRST works upstream — ingot, wafer and the materials and equipment around them — helping projects scope which stage to enter, what the entry actually requires technically, and how to sequence qualification against the policy clock. Assessments are project-specific and must be validated.
Primary sources and further reading
Last reviewed 2026-09-23. Technical scope, policy eligibility, availability, and commercial terms should be independently confirmed for each project.