The most instructive thing about this week’s perovskite announcement is not the number in the headline. It is the distance between the headline and the factory floor — and how much of that distance the announcement itself leaves undescribed.

On October 23, Insolation Energy Limited (INA Solar), through its subsidiary Insolation Green Energy Private Limited, announced a five-year strategic partnership with US-based Caelux Corporation for the commercialisation of 3 GW of advanced solar modules. The modules would combine Caelux’s proprietary “Energy Producing Glass” — a perovskite-based power-generating glass layer — with INA Solar’s N-Type TOPCon platform, under the product name Hybrid-Tandem™. Commercial production is targeted at module efficiency of up to 29% by 2028. The companies said they will work together on technology validation, manufacturing readiness, and regulatory approvals, including India’s Approved List of Models and Manufacturers (ALMM) requirements.

The announcement arrived as a paid press release carried by ANI via PNN — a detail that matters, because it sets the evidentiary status of everything in it. The figures below are the companies’ claims, reported accurately and then examined. None of them has been independently verified by a certification body, a customer order book, or a shipped product.

The numbers, handled honestly

Start with the three numbers that will travel furthest, and what each one actually is.

The 3 GW is a partnership scope, not shipped capacity and not a firm order book. It is the volume the two companies say they intend to commercialise over a five-year agreement. No production start date, no annual ramp schedule, and no first-shipment milestone are given in the announcement. A reader who treats “3 GW agreement” as “3 GW of perovskite modules now in production” would be misreading it by the full length of a manufacturing ramp.

The 29% is a target for commercial production by 2028, not a measured module efficiency. No certified test result, no cell or module area, and no measurement protocol accompanies it. It is also worth placing against the right baseline: conventional TOPCon modules ship today at efficiencies in the low-to-mid twenties, and the Navitas Solar agreement discussed below targets up to 28% against a conventional-TOPCon figure of around 25%, per pv magazine India’s reporting. A 29% commercial module in 2028 would be a genuine step up from today’s TOPCon product — but it is a promise about a future production line, not a laboratory record and not a product specification.

The 25 GW is a claim by Caelux chief executive Scott Graybeal, quoted in the release: “With this agreement, Caelux’s global contracted volume with module manufacturers exceeds 25 GW.” That is a company statement about its own contracted pipeline, not a disclosed list of contracts. The release names no other counterparties, and no independent source has published a contract-by-contract breakdown that would let a reader check the total. It should be read as a CEO’s characterization of demand, not as an audited figure — and it should be read alongside the fact that “contracted volume” in this industry routinely means multi-year partnership scopes like the INA deal itself, not modules sold.

None of this makes the announcement unimportant. It makes it an announcement, and announcements in perovskite manufacturing are best read as statements of intent whose real test is everything that happens next.

What the technology actually is

The technical idea is straightforward, and it is the reason Caelux’s approach is worth more attention than a conventional perovskite startup pitch.

In a standard silicon module, the front glass does nothing electrical. It protects the cells, admits light, and survives weather. Caelux coats that cover glass with a thin perovskite film stack, turning the glass itself into a first power-generating layer. The perovskite layer absorbs the higher-energy part of the spectrum; the silicon TOPCon cells underneath absorb the rest. The result is a tandem device — two absorbers stacked in one module — built by modifying the glass rather than redesigning the cell line.

That distinction is the whole business model. Caelux, founded in 2014 as a spin-out from the California Institute of Technology, does not make modules. Under its model, module manufacturers consign glass to Caelux’s Baldwin Park, California facility, where the perovskite film stack is applied; the coated glass is shipped back and integrated into existing module assembly lines with, the company says, minimal process changes. The pitch to a manufacturer is therefore not “replace your factory” but “add a layer to your existing factory.” For a company like INA Solar, which already operates module lines, that is a much easier adoption story than a greenfield perovskite line — and it is why Caelux’s partnerships are structured as multi-gigawatt scopes rather than pilot orders.

There is also a physical reason the glass approach is attractive, and a physical reason it is hard. Perovskites are famously sensitive to moisture: the material degrades when water reaches it, which is the central stability problem for the whole field. Embedding the perovskite layer within glass is encapsulation by architecture — the layer is protected by the very component that already has to survive twenty-plus years outdoors. That does not solve stability on its own; the layer still has to survive thermal cycling, ultraviolet exposure, and lamination, and the interfaces between the perovskite stack, the glass, and the encapsulant are their own engineering problem. But it does narrow the stability question: it is about one coated component inside a conventional module, not an entirely new device stack.

Why India, and why this is the second deal in the pattern

The INA agreement is not Caelux’s first move in India. In July 2026, Caelux announced commercialisation agreements with two Indian manufacturers, Rayzon Solar and Navitas Solar, each targeting approximately 5 GW of modules over five years using perovskite-coated glass with silicon technology. The Navitas agreement — 5 GW of hybrid tandem modules targeting efficiencies of up to 28%, with commercial production expected in 2028 — resurfaced in coverage on October 22, when Navitas announced up to $20 million in growth capital from Swiss impact investor responsAbility to build a solar cell facility. So within a single week, the Indian market has seen a fresh 3 GW Caelux agreement and a renewed public airing of a 5 GW one.

That clustering is the real story. Caelux is not testing the Indian market with one partner; it is assembling a portfolio of Indian manufacturing partners, all aimed at the same 2028 commercial-production window. India is a logical place to do it: a large domestic module-manufacturing base, an ALMM regime governing which models can be deployed in government-backed projects, and manufacturers competing on efficiency as the market matures past its first capacity wave. A technology promising a multi-point efficiency gain without a new cell factory is aimed precisely at that moment.

The ALMM point deserves a plain explanation, because the announcement names it as a next step without saying what it entails. The Approved List of Models and Manufacturers, maintained by India’s Ministry of New and Renewable Energy, is the list of module models and manufacturers eligible for government-supported projects. Getting a new hybrid-tandem model onto that list is not a formality: it requires the product to exist, to be tested, and to pass the listing process. The announcement’s own sequencing — technology validation, manufacturing readiness, then regulatory approvals — is honest about the order of operations. Validation comes first, and it has not yet been reported as complete.

The gap between announcement and module

That sequencing is where the scrutiny belongs. Between a signed partnership and a shipping product, a perovskite-on-glass program has to clear at least four gates, and the announcement is silent on the status of all four.

First, technology validation: the coated glass must demonstrate, on INA Solar’s actual TOPCon line and at module scale, the efficiency gain and the stability profile the partnership assumes. Laboratory demonstrations on small areas do not answer this question. Perovskite films are process-sensitive — grain structure, defect density, and uniformity depend on exact thermal and coating conditions — and a coating process that works on a pilot line does not automatically transfer to meter-wide glass at gigawatt throughput.

Second, manufacturing readiness: Caelux must be able to coat glass at the volume and yield the 3 GW scope implies, and INA Solar must be able to integrate that glass into its lines without degrading its existing TOPCon yield. The announcement gives no coating capacity figure, no yield target, and no timeline for line integration. INA Solar’s existing footprint — three module facilities in Jaipur with a combined 5.5 GW annual capacity, plus a 4.5 GW TOPCon cell facility under development in Narmadapuram, per the release — is real manufacturing scale, but it is TOPCon scale, not hybrid-tandem scale. Converting it is the work.

Third, certification and listing: the hybrid-tandem modules must pass the testing and certification regime that precedes ALMM listing, and must do so as a new product category — a perovskite-silicon tandem — rather than as a routine TOPCon variant. Regulators and test labs will want stability data, and stability data takes calendar time. Accelerated testing can compress some of it, but outdoor exposure data cannot be manufactured faster than the weather allows.

Fourth, economics: the release says the technology is “designed to boost energy generation, increase energy density and reduce the levelized cost of electricity (LCOE).” That is a design goal, not a measured cost. The coating adds a process step, a supplier dependency, and a new failure mode to a tightly optimized module cost structure. Whether the efficiency gain pays for the added cost depends on coating price, yield, and field lifetime — none of which are in the announcement.

This is the same gap that defines the entire perovskite field right now, and it is why this desk treats partnership announcements as the beginning of a story rather than its conclusion. The field’s laboratory records have been silicon-competitive for years. Its commercial question has always been manufacturing at scale, and manufacturing at scale is measured in yield, uptime, and shipped megawatts — not in gigawatts of announced scope.

What to watch

Four things will tell readers whether this agreement becomes a product.

Technology validation results: any disclosed pilot-line data from INA Solar’s facilities — module efficiencies at stated areas, measured under a named protocol — would convert the 29% target from an aspiration into a trajectory. Until then, the target is just a target.

ALMM progress: a listing application, or a listed hybrid-tandem model, would be the first public regulatory fact in the story. The announcement names ALMM as a next step; the listing itself would be the evidence that the step was taken.

Capacity signals: coating-line investment by Caelux, or line-integration work at INA Solar’s Jaipur or Narmadapuram facilities, would show capital following the press release. Partnership announcements are cheap; coating equipment is not.

The 2028 window: both the INA and Navitas agreements point at commercial production in 2028. If that year arrives with listed products and shipped volumes, Caelux’s portfolio strategy will have been vindicated. If it arrives with renewed targets and no products, the pattern will be legible in the other direction.

One further watch item is the competitive context the announcement does not mention. Caelux is not the only company pursuing perovskite-silicon tandems, and not the only one pursuing them in India. The race to commercial perovskite tandems is genuinely global, and India’s role in it is as a manufacturing base, not merely as a market.

The bottom line

INA Solar’s agreement with Caelux is a serious-looking industrial bet: a real manufacturer with 5.5 GW of existing module capacity, a technology approach that modifies glass rather than replacing factories, and a 3 GW, five-year scope aimed at a 29% commercial module by 2028. It is also, at this moment, entirely a bet. No validated product, no certified efficiency, no ALMM listing, and no shipment has been disclosed. The 25 GW of global contracted volume is a CEO’s claim about a pipeline of scopes like this one.

The perovskite field’s commercial question has always been manufacturing scale, not lab records, and this announcement is best understood as a statement about how one American technology company and one Indian manufacturer intend to answer it: by coating the glass that already goes on every module. Whether that answer works will be decided by validation data, coating yields, and a listing — not by the size of the announced number. The number to watch is not 3 GW, 29%, or 25 GW. It is the first certified hybrid-tandem module that ships.

No investment recommendation is made or implied anywhere in this article.

Sources and further reading

  • ANI via LatestLY (October 23, 2026, 05:36 PM IST), “INA Solar Partners with US-Based Caelux in 3 GW Agreement; Targets 29% Module Efficiency by 2028” — the primary source for the partnership terms, next steps, executive quotes, and INA Solar’s capacity figures. Carried as a PNN press release; ANI disclaims responsibility for its content.
  • Energetica India (October 24, 2026) and Outlook Business (October 2026) — trade re-reporting of the same announcement, confirming the terms.
  • pv magazine India (October 22, 2026) — the Navitas–Caelux agreement (5 GW, up to 28% target vs. around 25% conventional TOPCon, production expected 2028) and Navitas’s cell-manufacturing expansion.
  • Precedence Research / GlobeNewswire (October 2026) — Caelux’s July 2026 agreements with Rayzon Solar and Navitas Solar (approximately 5 GW each over five years), and the caution that partnership figures are planned volumes, not operating capacity or realized sales.
  • Dealroom company profile — Caelux founded in 2014 as a Caltech spin-out; the consignment-glass model (manufacturers’ glass coated at Baldwin Park, California, and returned for integration into existing lines).
  • Photonics Spectra (August 2023) — Caelux’s $12 million Series A3 (Series A total $24 million) and its Baldwin Park ramp toward 100 MW of perovskite-coated glass.
  • Taiyang News — Caelux’s Caelux One perovskite-coated glass product and its UNSW/ARENA research partnership on tandem-cell characterization.

Background on perovskite moisture sensitivity, film-process sensitivity at scale, and the tandem architecture is drawn from the standard perovskite processing literature; no laboratory record or certified figure is quoted in this article beyond the companies’ own stated targets, which are identified as targets throughout.