AIA Engineering Limited: The Grinding Media Duopoly

I. Cold Open & Roadmap

Somewhere in the Atacama Desert, a steel drum the size of a small apartment block is turning on its axis at a copper mine. Inside it, several hundred tonnes of hard steel balls are tumbling over each other in the dark, smashing chunks of ore into a slurry fine enough that the copper can be chemically separated out. The drum consumes an enormous amount of electricity. The balls inside it wear away — slowly, invisibly, atom by atom — and every few weeks a truck arrives to top them up. Nobody outside the plant thinks about this for a moment. It is the least glamorous step in the entire supply chain of the energy transition.

That tumbling, grinding, wearing-away is a business. A large one. And a company headquartered in Ahmedabad, listed on the NSE with a market value of roughly ₹39,000 crore as of late August 2026, has spent four and a half decades turning it into one of the more unusual industrial franchises in India.¹

AIA Engineering makes what the industry calls high-chrome mill internals: grinding media (the balls), mill liners (the armour plating inside the drum), and diaphragms (the internal partitions). Its products are consumables. They are destroyed in the act of being used. That single fact — that AIA sells something the customer must buy again next month, forever — is the load-bearing wall of the entire investment case, and it was chosen deliberately by the founder almost fifty years ago.

If the physical picture is still fuzzy, use this one. A grinding mill is a rock tumbler scaled up to industrial absurdity. You load ore and steel balls into a rotating cylinder; the cylinder lifts them and drops them; the falling balls crush what is underneath. The balls wear down and are topped up continuously. The liners bolted to the inside wall of the drum take the same beating and get replaced perhaps once a year. Everything AIA sells is one of the things being destroyed inside that cylinder. The company's entire commercial argument is about how fast those things are destroyed, and what happens to the plant's electricity bill and metal recovery while they are being destroyed.

That is why this is a metallurgy business and not a steel business. A tonne of steel is a tonne of steel. A tonne of grinding media that lasts 30% longer, breaks less often, and lets the mill run 5% harder is a different product sold on a different basis — and, as we will see, it takes years to convince a mine of that.

The company reported FY26 revenue from operations of ₹4,420 crore and consolidated net profit of ₹1,270 crore, its best year ever.² It shipped 258,002 tonnes of product, of which 159,813 tonnes went to mining customers and 98,189 tonnes to cement, thermal power, and general industry.² It is, by common industry description, the world's second-largest producer of high-chrome castings for grinding applications, behind the Belgian-founded, Chilean-owned Magotteaux.³

Here is the tension the rest of this piece is built around.

AIA's modern investment case is not "we sell a lot of grinding balls." It is a conversion thesis. The great majority of the world's mining grinding media is still cheap forged or cast steel — plain carbon steel, essentially, forged into balls and sold on price. AIA's product costs more per kilogram, wears more slowly, breaks less, and — the company argues — lowers the customer's total cost per tonne of ore ground.

If the mining industry gradually converts from forged steel to high-chrome, AIA's addressable market multiplies without the company having to win a single share point from Magotteaux. Management has repeatedly framed the opportunity as somewhere between one and 1.5 million tonnes a year across gold, copper, and iron ore, against a total wear-parts opportunity it puts at 2 to 2.5 million tonnes.⁴ Against FY26 shipments of roughly a quarter of a million tonnes, that is the whole story.

The problem is that stories like this get tested. AIA's got tested in FY25, when revenue fell to ₹4,287 crore from ₹4,854 crore and sales volumes dropped about 14% in a single year.[^5] That is not a rounding error and it was not a cosmetic accounting artefact. It happened in the very segment the bull case depends on, and it happened for reasons — customer inventory cycles, trade policy, delayed conversions — that a "structural, multi-decade" framing is supposed to be insulated from.

So: does the thesis survive?

The short answer this piece will build toward is that it survives in a narrower form than the promotional version. What follows is the full case — how a failed foundry became a near-monopoly in Indian cement, how that near-monopoly funded a bet against cheap steel, what the moat actually consists of when you press on it, and what management is and isn't telling investors right now as it holds ₹4,500 crore of idle cash and declines to give a volume number for the year ahead.

A note on posture before we begin. AIA is a company with an unusually loyal shareholder base and an unusually articulate management, and the two reinforce each other: the executives give long, thoughtful, first-principles answers on earnings calls that are genuinely more informative than the industry norm. That is a real asset. It is also a reason to read them carefully rather than credulously, because an articulate explanation of why a business is special is still an explanation, not evidence. Where this piece can test a claim against the company's own record, it does. Where it cannot, it says so.

It starts, as these stories usually do, with someone losing all their money.

↑ Back to Top

II. Origins: A Foundry That Almost Failed (1979–2005)

In 1975, a young metallurgical engineer came home to Ahmedabad from IIT Kanpur with a B.Tech and not much else. Bhadresh Shah had graduated in metallurgy in 1974 — the institute would later give him a Distinguished Alumnus Award for entrepreneurship — and rather than take a job, he decided to build a foundry.⁵ With a loan from Central Bank of India and money from his father, he set up a small operation in 1976 making pressure die castings, supplying compressor makers and textile units around Gujarat.⁶

It failed almost immediately. The castings had quality problems, customers complained, and Shah lost essentially the whole of his roughly ₹1 lakh of capital inside the first year.⁷

What he did next is the single most consequential decision in the company's history, and it had nothing to do with metallurgy. In 1977 he stopped selling components to original equipment manufacturers and started selling replacement components to power plants.⁶ The logic is worth sitting with, because it explains everything the company has done since.

Selling to an OEM means you sell once, into a purchase order that gets re-tendered on price. Selling replacement parts means you sell into a wear cycle: the part is consumed, the plant cannot run without it, and the buyer is the plant manager who cares more about downtime than about the last five percent of price. It is a lower-glamour business with better repeat economics and a buyer whose incentives are aligned with quality rather than procurement savings.

In 1979 Shah co-founded Ahmedabad Induction Alloy with a friend, making steel alloys and iron castings for power plants — the entity that would eventually become AIA Engineering.⁶ The company was incorporated in 1978 under that name and only later took its present form.⁴

The technical breakthrough came from Europe. In 1988, Shah partnered with the Belgian consultancy Slegten SA for design expertise, a moment he later described as a turning point.⁶ Three years later, in 1991, that relationship deepened into a joint venture with Magotteaux International — the Belgian company that was then, and remains today, the world's largest producer of high-chrome grinding balls.⁶ Magotteaux took 51%; Shah held 49%. Through that JV, AIA learned to make high-chromium castings properly.

It is worth pausing on what that means physically, because the product is more interesting than "a metal ball."

A high-chrome grinding ball is not steel with chrome painted or plated on it. It is a casting of white iron alloyed with a large fraction of chromium — the chromium and carbon in the melt combine into extremely hard chromium-carbide crystals suspended in a tougher matrix, like gravel set in concrete. The carbides resist abrasion; the matrix stops the whole thing shattering when it slams into rock at speed.

Getting that balance right is the whole trick. Too hard and the ball is brittle and fractures, filling the mill with useless fragments. Too tough and it wears away too fast, and the customer is back to buying more of them. Tuning it to a specific ore hardness, a specific mill diameter, and a specific throughput is a metallurgy and heat-treatment problem that takes years of iteration to master. AIA produces something on the order of a hundred different alloys for this reason.⁷ Chemistry is the easy part. Chemistry that works in this mill, on this ore body, at this throughput, is the hard part — and it is the only part a competitor cannot buy.

The partnership did not last. Shah spent 1996–97 at Magotteaux's Belgian headquarters and came away convinced the Belgians were not paying enough attention to the Indian venture; he returned to India in 1997, and the agreement was wound down around 2000.⁸⁴ He bought out Magotteaux's stake and renamed the company AIA Engineering.⁶

That divorce is the founding fact of the modern company. AIA did not enter the high-chrome market as an insurgent with a new idea — it entered as the former junior partner of the incumbent, carrying the incumbent's process knowledge, with a cost base in Gujarat rather than Wallonia. Its opening move in export markets was, in the plainest terms, the same product at a materially lower price.⁴ Its competitor for the next twenty-five years would be the company that taught it the trade — and, as we will see in the trade-remedy section, would later petition governments to keep it out.

In 2003, AIA acquired Vega Industries in the UK, the beginning of an overseas sales-and-service network that would eventually stretch from Dubai to Santiago.⁴ And in 2005 the company went public, pricing its IPO at ₹315 a share and raising roughly ₹148 crore.⁶ Shah had reportedly considered private equity — the decision reportedly crystallised during an airline conversation — and chose the listing instead.⁶ At the time of the offering, revenue was around ₹406 crore and net profit around ₹52 crore; by FY14 those had become roughly ₹2,000 crore and ₹325 crore.⁶

The same year, AIA acquired Welcast Steels, a Bengaluru-based high-chrome grinding media maker founded in 1972, taking a 74.85% stake.¹³ Welcast has remained a small, separately listed appendage ever since — under ₹100 crore of revenue and occasionally loss-making — and its fate two decades later becomes a small but instructive test of AIA's capital discipline, discussed later.

What the record of these three decades actually establishes, stated carefully: a founder who chose a repeat-purchase business model after a first-attempt failure, licensed world-class metallurgy through a joint venture rather than inventing it, took full ownership when the partnership stopped serving him, built a services arm before he needed one, and funded the whole thing conservatively. What it does not establish is any claim of technological uniqueness. AIA's opening competitive position in export markets was cost, not capability — the same product, materially cheaper.⁴ The capability advantage, to the extent it exists today, was accumulated afterwards.

Two things about the founding matter for a reader in 2026, and one thing does not. What matters: the replacement-parts business model was a deliberate choice made in year two, not a happy accident, and the metallurgical capability was acquired from the global leader rather than invented. What does not matter much: there is no "previous management" chapter here. Shah has run this company continuously since before the IPO, and the board reappointed him as Managing Director for a further five years from October 1, 2026.⁹ Almost every judgement in the rest of this piece — good and bad — is a judgement about one operator's decisions over five decades.

The first place those decisions compounded was not mining. It was cement.

↑ Back to Top

III. The Cement Fortress: Building a Near-Monopoly at Home (2000s–2010s)

A cement plant is, from AIA's point of view, an almost perfect first customer. It runs mills continuously. It grinds a material — limestone and clinker — that is abrasive but predictable, unlike ore bodies that change composition as a mine deepens. Its plant managers are measured on throughput and specific power consumption, both of which are directly affected by what is tumbling inside the mill. And in India in the 2000s, there were a great many of them being built.

AIA went after that market with a combination that its competitors, mostly foreign, could not easily match: European-derived metallurgy, an Indian cost base, and — crucially — a service model. The company did not simply quote a price per tonne of grinding media. It audited the mill. It measured the grinding efficiency, proposed a different ball-size distribution or a different liner profile, ran the change, and measured again.

This is what management still means, three decades later, when it insists AIA is not a parts vendor. On the August 2026 earnings call, Executive Director Kunal Shah put it bluntly: "when I go to a cement plant and I tell them, you have a 50-ton per hour mill, and I'm going to improve that by 20%, that requires me to take ownership of the whole grinding process itself."¹⁰

The result was dominance. AIA is generally described as holding a 90–95% share of India's cement grinding-media market, and roughly 35–40% of the global cement segment outside China.¹¹ Those are the numbers most often cited by brokerages covering the stock, and they are consistent with the company's own positioning as India's only significant domestic manufacturer of high-chrome mill internals.⁴ An independent investor should treat "90–95%" as an industry estimate rather than an audited statistic — AIA does not publish a market-share table — but the direction is not seriously in dispute.

The mechanism for exporting this playbook was Vega. Originally the international sales, service, and technical support arm, Vega Industries grew into a network of wholly-owned entities across the UK, the Middle East, the United States, South Africa, and China, each functioning as a local technical presence rather than a distributor.⁴ The distinction matters. A distributor takes an order. A Vega engineer walks the plant, pulls samples, and comes back with a proposal that involves changing what the customer buys. The physical product ships from Gujarat; the relationship is local.

The customer list this produced reads like a roll-call of global heavy industry. In cement: Lafarge, Holcim, Heidelberg, Cemex, Italcementi, Votorantim. In thermal power: NTPC, BHEL, the Indian state electricity boards, and equipment makers including Alstom and Hitachi.⁴ These are not accounts you win with a low quote. They are accounts you win by surviving a technical audit, and then keep by not causing an unplanned mill stoppage for a decade.

There is an underappreciated financial consequence to running this model from a single manufacturing base. Because AIA ships physical product across oceans and holds stock in multiple geographies for customers who cannot afford to wait, the business is working-capital heavy in a way that pure-service businesses are not.

Inventory days have run north of 200 for years, receivables have historically sat in the 65–95 day range, and the cash conversion cycle has repeatedly exceeded eight months.¹³ Exports have accounted for roughly 70–75% of sales, which also makes reported margins sensitive to the rupee.¹³ A reader should carry this forward: AIA's returns are good, but they are earned on a balance sheet that ties up a lot of capital per rupee of sales — one reason the company has always needed to be conservatively financed.

What did the cement business actually buy AIA, in strategic terms? Three things, in descending order of durability.

First, a cash engine. Cement customers reorder on a predictable cycle and, at a 90%-plus domestic share, AIA was effectively selling into an annuity. That funded everything that followed without recourse to debt — a point we will return to when we look at a balance sheet that has been essentially borrowing-free for a decade.¹

Second, a proving ground. The service-plus-product model was refined against hundreds of cement mills before it was ever pointed at a copper mine. When management says today that its "legacy is not a part manufacturer, legacy is a solution provider," the claim has a genuine twenty-year track record behind it in cement.¹⁰ That is the strongest form of evidence available for a soft capability: not an assertion, but a business that was actually built that way.

Third — and this is where an investor should be careful — a set of switching costs that are real but bounded. In cement, AIA's share is so high that the relevant question is not whether it can win customers but whether it can lose them, and here the honest answer is that the fortress is a mature, slow-growing one. India's cement industry has been consolidating, and cement grinding is a well-understood problem; the incremental value AIA can add to a mill it optimised five years ago is smaller than the value it added the first time.

There is also a slow structural drag hiding inside this segment that rarely gets discussed. The "others" bucket includes thermal power, where high-chrome internals are used to grind coal. Coal-fired generation is not a growth market in most of the world, and Shah has said plainly for years that AIA's growth would come from mining rather than from declining coal-based power.⁷ So within a segment that looks flat in aggregate, cement is holding and one sub-segment is structurally shrinking. That is not a crisis at 38% of volume, but it does mean the base business has to run to stand still.

The segment framing tells the story. What the company calls "others" — cement, utilities, and general industry — contributed 98,189 tonnes of FY26's 258,002 tonnes, or about 38%.² In the June 2026 quarter it was 25,416 tonnes of 64,644, again roughly 39%, growing about 7% year on year.¹² That is not a business in decline. But it is not the growth story either, and it has not been since the early 2010s. Back in FY14, mining and non-mining were nearly evenly split; the mix has been drifting toward mining for a decade.⁴

So the cement fortress is best understood as what it is: a high-share, high-margin, low-growth base that pays for the option on something bigger. The interesting question — and the one that determines whether AIA is a compounder or a well-run niche manufacturer — is whether the same playbook works when the mill is sixty times larger and the customer is BHP.

↑ Back to Top

IV. The Mining Pivot: Betting Against Cheap Steel Balls (2010s–2020s)

Around 2008, AIA started chasing iron ore mines in Brazil, and by 2011, having taken the lion's share of the Indian cement space, the company formally turned its attention to supplying high-chrome mill internals into mining.⁶⁴ The bet it was making can be stated in a single sentence: that mine operators would eventually stop buying grinding media on price per kilogram and start buying it on cost per tonne of ore ground.

To see why that is a hard sell, consider the customer's arithmetic.

Grinding media is roughly a tenth of the operating cost of a mid-sized mine.⁴ A high-chrome ball can cost meaningfully more per kilogram than a forged steel ball — brokerage estimates have put the capital cost premium for chrome internals at around 20%, with forged media selling 20–40% below high-chrome on a like-for-like basis.¹³⁴ To justify that, the chrome ball has to wear slowly enough, break rarely enough, and improve mill throughput or metal recovery enough that the mine comes out ahead. It usually does, on a full lifecycle basis.

But here is the organisational problem hiding inside the economic one. The saving shows up over quarters, in throughput and power and recovery — line items the mine's procurement department does not own. The higher price shows up immediately, in a line item it does. AIA is asking a buyer to accept a certain, immediate, measurable cost increase in exchange for an uncertain, delayed, harder-to-attribute benefit that will be credited to someone else's performance review. Every difficulty AIA has ever had converting a mine traces back to that asymmetry, and no amount of metallurgy solves it.

That is why the conversion has been slow. And that slowness is precisely the opportunity. Estimates of high-chrome penetration in mining grinding media have hovered under 20% for years — HDFC Securities put it at less than 20% in 2022, with the rest served by conventional forged and cast products; Keynote Capitals had estimated 15–20% two years earlier.¹³⁴ Note that these figures have not moved much across the decade of published research, which is itself a data point: this conversion has been "about to accelerate" for a long time.

Structurally, though, the pivot worked. Mining is now the larger half of AIA by a clear margin — 159,813 tonnes of FY26's 258,002, or about 62%.² In the June 2026 quarter, mining was 39,228 tonnes against 25,416 for everything else.¹² A company that began as a cement supplier with a sideline in power plants now sells the majority of its output into copper, gold, and iron ore.

Getting there required capacity, and AIA's capacity history is a useful proxy for how confident management has been at each point. Installed capacity was around 260,000 tonnes in the mid-2010s and roughly 340,000 tonnes for several years after.⁴ By FY22 it was 390,000 tonnes, expanding to 440,000 in FY23, with a further 80,000 tonnes planned at about ₹200 crore of capex to reach 520,000 tonnes by the end of FY24.¹³ That last expansion did not happen as announced. It was cut to 36,000 tonnes and then, in the December 2024 quarter, halted altogether, with the company deciding to stay put at around 460,000 tonnes in India and pause ongoing work.[^5] Today management describes the base as 430,000–440,000 tonnes.¹⁰

That sequence — announce 520,000, cut, then pause — is worth naming plainly, because it is the same behaviour pattern that shows up later around Ghana, China, and volume guidance. AIA has a consistent habit of announcing capacity or growth ambitions on the assumption that conversions will arrive, and then quietly deferring when they do not. It is not reckless; nothing was built and stranded, and that is genuinely to management's credit. But it does mean announced plans at this company carry a lower information content than they would at a firm with a tighter announcement-to-delivery record.

Alongside capacity, AIA also bought in outside know-how rather than developing everything internally — as it had with Slegten and Magotteaux decades earlier. It entered a technical collaboration with EE Mill Solutions LLP of the USA, which specialises in optimising energy efficiency and output of AG, SAG, and ball mills through redesign of mill internals — head and shell liners, grate liners, pulp lifters — and began developing the mining liner business on that basis.¹³ Mill liners matter strategically because they are the second consumable in the same mill: winning the liner alongside the grinding media roughly doubles the wallet share per customer and, more importantly, gives AIA design control over the mill rather than just supply of its ammunition.

The international build-out that supported this has been notably unambitious in the best sense. AIA has not bought scale; it has bought control of relationships it already had.

In August 2023, Vega Industries (Middle East) FZC agreed to take 30% of Mining Products and Service Pty, an Australian mining-liner and process-engineering business, with an option to acquire a further 40% over three years.¹⁴ In November 2024 it exercised part of that, taking its stake from 43% to 56% for about AUD 2.13 million; the target had done AUD 34.5 million of turnover and AUD 4.8 million of profit the prior year.¹⁵ In December 2025 it took the final 14% to reach 70%, for AUD 5.6 million.¹⁶ In the same period it moved to full ownership of a Peruvian entity, Vega Industries Peru, for a consideration of a few thousand Peruvian soles — a formation, not an acquisition.¹⁷

The pattern is unmistakable: small, incremental, control-increasing purchases of businesses AIA already partly owned and fully understood. There is a real analytical conclusion here, and it is not simply "management is disciplined." It is that AIA has never bought its way into a market it could not sell its way into. That is a genuine constraint on the downside — you cannot write off an acquisition you did not make — but it is also a constraint on the upside, because it means every tonne of growth has had to be won one mine at a time, through the trial process described in the next section.

The company's own history of prior deployments is bounded but clean on the evidence available: this piece found no record of a large impairment, shuttered venture, or written-down acquisition in AIA's recent corporate record, though that is a statement about what public filings and brokerage coverage disclose over roughly the past decade, not a guarantee that nothing exists. Read at the right confidence, the record supports "no large value-destroying deal has surfaced," not "management is a proven capital allocator." Those are different claims, and only the first one is evidenced — because a company that has never deployed a large sum has also never demonstrated it can.

Now the competitive map, because "duopoly" is a word that needs testing.

At the high-chrome tier, AIA's rival is Magotteaux, wholly owned since 2011 by Santiago-listed Sigdo Koppers. Magotteaux describes itself as the world's number one high-chrome grinding ball producer, with 390,000 tonnes a year of high-chrome ball capacity, a further 275,000 tonnes of forged steel and low-chrome balls, and 69,000 tonnes of castings, across 15 production plants selling into 150 countries. It sold 544,000 tonnes physically in 2024 and claims a 15% share of high-chromium grinding balls in mining.³ AIA, by comparison, has roughly 430,000–440,000 tonnes of installed capacity concentrated in India and shipped 258,002 tonnes in FY26.¹⁰² Brokerage estimates of the split of the high-chrome market have put Magotteaux around 35% and AIA around 25%.⁴

But — and this is the framing error most bull cases make — the high-chrome market is the small market. The dominant players in grinding media overall are the forged-steel producers: Moly-Cop, the largest forged-ball manufacturer globally, and ME Elecmetal, alongside a long tail of regional producers.¹⁸ Between them, the two high-chrome specialists hold a modest slice of total grinding media tonnage precisely because most of the tonnage is still forged.

Which reframes the competitive question. AIA and Magotteaux are not fighting each other for a fixed pie. They are, in effect, joint evangelists for a product category, competing hardest at the margin over which of them gets the account once a mine decides to convert. The real contest is against forged steel — against the incumbent, cheaper, good-enough substitute that has held roughly four-fifths of the market for as long as anyone has been measuring it.

Magotteaux, however, competes in another arena too. In Brazil, anti-dumping and countervailing duties were imposed on AIA following a petition by Magotteaux; the anti-dumping duty was terminated and the countervailing duty cut from 6.5% to 2.9% in July 2025.[^5] That is worth naming plainly: AIA's largest competitor has used trade remedies as a commercial weapon against it, successfully, for years. The duopoly is not a comfortable cartel. It is two companies with different cost bases, one of which has a European manufacturing footprint and the legal standing that comes with it.

Which brings us to the harder question. What, exactly, stops a customer from switching?

↑ Back to Top

V. How the Core Business Actually Wins (or Doesn't)

The clearest evidence AIA has ever produced for its own thesis arrived in October 2025, when a Vega subsidiary announced a $32.9 million order — roughly ₹291 crore — from a large international copper miner in Chile for high-chrome grinding media, to be executed over eighteen months starting November 2025.¹⁹[^5] Chile is the world's largest copper market. It is also, by AIA's own account, a market where the top two mines run Chinese forged media.¹⁰ Winning a high-chrome conversion there, in that market, against that incumbent product, is the single best real-world validation of the cost-per-tonne argument the company has been making for fifteen years.

By the August 2026 call, that order was contributing roughly 3,000 to 3,500 tonnes a quarter and management expected it to continue.¹⁰ Do the arithmetic and it is about 5% of quarterly volume — meaningful, not transformative. That ratio is the honest scale of what a marquee conversion win looks like for this business.

The economics of conversion. AIA's pitch is not a price pitch. It is a cost-per-tonne-milled pitch, and it can only be made by a supplier willing to take responsibility for outcomes it does not fully control. This creates a specific commercial rhythm: long qualification, sticky retention. A mine does not swap grinding media the way it swaps stationery suppliers, because a bad ball can cost more in lost throughput in a week than the entire annual media budget. That asymmetry is genuinely protective once you are in.

It is equally punishing on the way in. Kunal Shah described the trial process on the Q1 FY27 call with unusual candour: the journey "of doing a trial, the iteration is uncertain in a way that it could take 3 months, it could take 2 years because every mine site is different."¹⁰ Management started with smaller mills, is now working through medium and large ones, and told investors to expect updates that "may continue to be bland in terms of just saying status quo" for three or four quarters.¹⁰

That is the moat and the vulnerability in the same sentence. The qualification cycle that makes the relationship durable also makes revenue timing unforecastable and puts the pace of growth substantially in the customer's hands.

Switching costs, tested. The strongest disconfirming evidence against AIA's stickiness claim comes from the company itself.

In FY25, management attributed the volume decline partly to a customer lost to competition amid duty-related uncertainty, alongside inventory corrections by key customers and delayed conversions.²⁰²¹ A single lost account in a year is a low base rate and management explicitly framed it as rare. But it establishes the principle: these accounts are not contractually locked. AIA has confirmed that its supply contracts fix volumes, not prices, with pricing reset on a quarterly or half-yearly basis.¹³ Stickiness here is behavioural and technical, not legal.

The buyer-side dynamic reinforces the caution. Because the purchase decision on grinding media sits with mine managers responsible for productivity rather than with central procurement, product placement is inherently mine-specific and time-consuming — AIA has described its own transition as moving "from being a pure cost leader to cost plus value-added supplier."¹³ That phrasing is more revealing than it looks. It concedes that the starting position was price, and it implies the value-added position is something the company has been building toward rather than something it has always had.

There is a second, subtler test. If AIA's switching costs were as strong as the narrative implies, one would expect pricing power to show up cleanly in margins. Instead, the company describes a pass-through model: realisations rise when freight and ferrochrome rise, and other expenses rise alongside them. Non-executive director Sanjay Majmudar confirmed exactly this in August 2026 — realisations went up "as the freight goes up, as the raw material prices go up, as we do the pass-through," and "my other expenses would have also gone up."¹⁰ Gross margin has sat in a consistent 60–61% band.¹⁰ That is the profile of a business with good, stable, defensible economics — not one with the ability to price above cost inflation at will. Investors should hold the switching-cost claim at that calibration.

The "solution versus product" argument, examined. Management's sharpest articulation of its own moat came in response to an analyst question about how AIA's mill liners compare with competitors who specialise in composite or rubber liners. Kunal Shah's answer was that this is the wrong comparison entirely. Material choice, he argued, is dictated by operating conditions — metal liners in SAG mills, rubber in smaller mills, composites in between — and a liner specialist "can only bring intervention and innovation on the material and squeeze in a little more then it's a cost question." His framing: "The material can only go so far. The magic comes from the design conversation... The design intervention only happens if you understand the process. And none of the incumbents across the spectrum deal with process."¹⁰

He then drew the line even harder: a forged-ball maker "has an insignificant influence at the customer's operating conditions. He will buy from another vendor or a third vendor or one of 30 forged players in China making the product." Whereas AIA's proposition is to walk into a copper producer grinding 2,000 tonnes an hour and offer to improve throughput 10%, cut power 10%, and lift recovery by two to five percentage points.¹⁰

This is the most compelling version of the bull case anyone will articulate, and it is directionally credible — the cement track record is real evidence that AIA can do process work, not just parts work. But an investor should notice two things about it.

First, it is a claim about AIA's approach, and approaches are copyable in a way that patents are not; nothing in it explains why Magotteaux, with 15 plants and a Chilean parent embedded in South American mining, cannot make the same argument. Second, if the solution proposition were as differentiated as described, one would expect it to command a premium margin. Management says explicitly that it does not: the goal is stickiness, and the margin profile "does not necessarily translate into a very different margin profile."¹⁰ A moat that produces retention without pricing premium is a real moat. It is just a narrower one than the rhetoric implies.

Porter's five forces, briefly. Supplier power is moderate: ferrochrome and steel scrap are the key inputs, both traded and both volatile, and management noted in August 2026 that ferrochrome was "on the higher side currently" while insisting it was passing costs through.¹⁰ Buyer power is moderate-to-high — these are BHP, Rio Tinto, Anglo American, Vale, Newmont, and their peers, buyers with in-house metallurgists and a structural preference for multiple vendors.⁴ The threat of substitutes is the entire game: forged steel is the substitute, it is cheaper, and it still holds roughly four-fifths of the market. New entry is concentrated in China. Rivalry at the high-chrome tier is duopoly-like today, but is the least interesting of the five forces here.

The Chinese question, placed where it belongs. Asked directly on the August 2026 call whether Chinese players were undercutting AIA and threatening conversions in unprotected markets, Kunal Shah's answer was "Not really" — and then he described something that deserves more attention than the denial. The Chinese, he said, "have made a dramatic entry into the forged space where, you know, as much as the top two mines, I think, in Chile today use Chinese forged media. So, you know, China has become the largest producer of forged and a very important and there are like 20 forging companies, you know, now in the fray supplying to all sorts of mines across the world."¹⁰

His argument for why this does not threaten AIA is structural rather than empirical: forged is "one product, one grade, one material, and then it's a distribution business," whereas high-chrome "is a custom business" requiring front-end engineering. On that basis, "I don't think Chinese presence in high chrome is there today."¹⁰ He then added the caveat that most managements omit: "Now that may change in the future. All of these are questions of conjecture."¹⁰

Treat this as an untested claim, not a proven one. What is established: Chinese producers scaled a wear-parts export business from nothing to global leadership in forged media, in the same customer accounts, within roughly a decade. What is not established: that the engineering barrier to high-chrome is durable against a competitor that has already demonstrated it can localise, scale, and win those exact buyers. The relevant precedent is not "have they entered yet" — it is "how fast did they enter the adjacent segment." That is the single most important forward risk to the moat claim, and the thing to watch is not aggregate Chinese exports but Chinese-brand participation in high-chrome tenders specifically.

Myth versus reality. Four consensus statements about this company deserve correction before the rest of the analysis proceeds.

Myth: AIA is a duopoly with pricing power. Reality: it is one of two specialists in a product category that holds roughly a fifth of its own end market, competing mainly against a cheaper substitute rather than against each other, and it explicitly runs a cost pass-through pricing model rather than a premium-pricing one.¹⁰¹³

Myth: the mining conversion story is structural and therefore smooth. Reality: the company's own FY25 says otherwise, and management has stopped forecasting the pace.²¹¹⁰

Myth: anti-dumping duties protect AIA. Reality: on the balance of the evidence, trade remedies have cost AIA more volume than they have protected — duties in Brazil obtained on a competitor's petition, a dumping duty in the United States its customers absorb, and tens of thousands of tonnes lost in Canada and South Africa.[^5]²¹¹³

Myth: the cash pile is dry powder for expansion. Reality: management has said the near-term uses are trials, a corporate headquarters, and land — with distribution explicitly deferred and buybacks explicitly not contemplated.¹⁰

A Hamilton Helmer reading. Of the seven powers, the best fit is process power — accumulated metallurgical and grinding-process know-how built over decades, hard to replicate quickly because it lives in iteration rather than in documents. Layered on that are modest switching costs from the technical-service bundle and the qualification cycle. What AIA does not have: scale economies (Magotteaux is larger and it does not appear decisive), network effects (none), branding power in any consumer sense, or a cornered resource — there is no patent-like exclusivity that prevents a competent competitor from matching a specification over time, and AIA's own history proves it, since it entered this market by matching its former partner's product at a lower price.

Process power is a real power. It is also the slowest-acting and most quietly erodible of the seven. It does not prevent competition; it only makes competition expensive and slow. Which is exactly what AIA's own conversion history looks like from the other side of the table.

↑ Back to Top

VI. Current Management: Ownership, Incentives, Capital Discipline

The most striking number on AIA's shareholding page is not a financial metric. It is 58.5% — the promoter holding, unchanged quarter after quarter for years, sitting alongside 16.9% held by foreign institutions and 22.1% by domestic institutions as of June 2026.¹

For a company of this size, that is an unusually concentrated founder position, and it has a specific analytical consequence. A promoter who owns 58.5% captures 58.5 paise of every rupee of value created and bears 58.5 paise of every rupee destroyed. Empire-building is expensive for someone in that position in a way it is not for a professional manager with options. Bhadresh Shah, 74, has held the Managing Director's chair for essentially the company's entire life and was reappointed for another five years from October 1, 2026.²²⁹

The compensation data supports the alignment reading. Shah's FY25 total compensation was about ₹1.2 crore, entirely cash, with no ESOPs — roughly 23 times median employee pay at the company.²² Set that against FY26 net profit of ₹1,270 crore and the extraction rate is close to a rounding error.² For an Indian promoter-run industrial with a controlling stake, that is a genuinely low-extraction structure, and it is the kind of fact an activist investor would find nothing to attack in.

It is worth being precise about who actually speaks for the company, though, because it is not the founder. AIA's earnings calls are conducted by Kunal Shah, Executive Director for Corporate Affairs, and Sanjay S. Majmudar, a non-executive, non-independent director.¹⁰ Bhadresh Shah does not take investor questions. That is not unusual for a promoter of his generation and it is not a governance defect — but it does mean public assessment of this management is an assessment of two people articulating a strategy set by a third, and it makes the succession question a genuinely open one at a company where a 74-year-old founder was just handed another five-year term.²²⁹ Succession planning has not been disclosed in any detail, and for a business whose competitive claim rests on accumulated judgement rather than on documented process, that gap matters more than it would elsewhere.

The capital allocation record. AIA has run with essentially no borrowings for a decade — ₹10 crore of debt against ₹8,007 crore of reserves as of March 2026 — while paying steady dividends.¹ The board recommended ₹16 per share for FY26, with a record date of September 5, 2026.² Dividend payout has been modest at around 12% of profit.¹

In August 2024 the company completed its first-ever buyback: 10 lakh shares at ₹5,000 apiece, ₹500 crore in total, about 1.06% of equity, with a record date of August 20, 2024.²³ The stock rose over 5% on the announcement to an intraday high of ₹4,697.²³ Two readings of that event are available, and both are true. It was shareholder-friendly, and it was late — the cash pile had been conspicuous for years before the board acted.

There is one prior capital-allocation episode worth putting next to the "disciplined operator" characterisation, because it complicates it. In December 2023, AIA offered to buy out the remaining 25.15% of its listed subsidiary Welcast Steels for about ₹25 crore via reverse book building.²⁴ The delisting failed.

Public shareholders validly tendered just 62,099 shares — far short of the 90% threshold required under SEBI's delisting regulations — and the offer lapsed with AIA acquiring nothing.²⁵ It is a small transaction and no capital was destroyed. But it is a data point about pricing discipline that cuts in two directions at once: AIA declined to raise its price to close a deal it wanted, which is disciplined; and it misjudged what minority holders would accept, which is a failure of execution on a trivially small transaction. Investors evaluating the "clean capital allocation" claim should hold both.

Return ratios, and what they are actually saying. AIA's returns have been strong but are trending in a direction that deserves explanation rather than applause. Brokerage estimates put return on net worth at 18.6% in FY23, 17.1% in FY24, and 15.3% in FY25, with return on capital employed sliding from 22.2% to 18.3% over the same span; the FY26 recovery brought reported ROE back to around 17% and ROCE to about 21%.[^5]¹

The decline was not primarily an operating problem. It was a denominator problem. Cash accumulated on the balance sheet faster than it was deployed, and cash earns treasury yields, not business returns — AIA booked ₹85.35 crore of treasury income in the June 2026 quarter alone.¹⁰ The company is now sitting on ₹4,500–5,000 crore of cash.¹⁰ Whether that is prudence or a shortage of attractive reinvestment opportunities is the central capital-allocation question, and management's own answer is closer to the first: Majmudar told investors the board wants to hold "a little higher cash, I would say at least for a few more quarters," while the trials play out, before considering distribution options, and explicitly said a buyback is "not contemplating it in near future."¹⁰

One further capital-allocation exposure deserves a mention, because it is invisible in the headline numbers. AIA's profits have at times depended materially on export incentives that are set by policy rather than by the business. Changes to India's RoDTEP scheme — the removal of the benefit on grinding media and a reduced rate elsewhere — cost the company an estimated ₹60 crore in FY22 alone.¹³ That is roughly a tenth of that year's net profit, gone because of a line in a government notification. It is a small recurring reminder that a large exporter operating from a single country carries policy risk on both the revenue side and the incentive side.

Credibility, assessed through behaviour. For most of its history, AIA gave the market specific tonnage guidance and explained shortfalls in granular terms. The FY25 miss was attributed openly and in detail — inventory corrections, logistics disruption, delayed conversions, a duty-related customer loss, slower mining-liner ramp-up.²⁰²¹ That is a non-evasive explanation, and it matters: managements that blame "the macro" and stop there are telling you something different from managements that name five specific mechanisms.

But there has been a real, traceable shift in disclosure posture, and it is older than most commentary suggests. After FY25, management declined to give FY26 volume guidance, citing global volatility, while indicating an ambition of 25,000–30,000 tonnes of annual additions.²¹ Around October 2025 the framing was that AIA expected to return to adding 30,000–40,000 tonnes a year from new conversions "in next 2-3 quarters."[^5] FY26 delivered an increase of 2,559 tonnes.² That gap between stated ambition and delivered volume is the most concrete test of management's forecasting reliability available, and AIA failed it — not through dishonesty, but by consistently underestimating how long conversions take.

By August 2026, guidance had been withdrawn a second year running. Asked directly whether FY27 would reach 280,000–290,000 tonnes, Majmudar's answer was "I have, frankly, no specific answer to give on this," with a request to "wait for one more quarter."¹⁰ Asked to update an eighteen-month-old realisation guidance of ₹160–165 per kg when actual realisation had exceeded ₹180, Kunal Shah declined, responding with a list of the variables he could not forecast — the dollar, ferrochrome, scrap, competition, shipping.¹⁰

The charitable reading is that this is intellectual honesty about a genuinely unforecastable trial pipeline, and there is real merit to it — a management that refuses to guide is at least not setting up the next miss. The less charitable reading is that a company which repeatedly missed its own volume ambitions has responded by ceasing to state them. Both readings are consistent with the evidence. What is not ambiguous is the direction: investors are being asked to accept less disclosure than they used to get, at the same time as they are being asked to fund more spending.

↑ Back to Top

VII. The FY25 Stress Test: When the Structural Story Met a Real Slowdown

Every structural growth story eventually meets a year that does not cooperate. AIA's was the twelve months ended March 2025, and it is the most important stretch of history in this entire piece, because it tests the exact claim the bull case rests on.

The numbers first. Revenue fell to ₹4,287 crore from ₹4,854 crore, a decline of about 11.7%.[^5] Sales volume fell roughly 14%, to 255,443 tonnes.²¹ Fourth-quarter volume was 68,741 tonnes, down 3.8% year on year, and net profit for that quarter came in at ₹285 crore.²¹²⁶ EBITDA margin held up reasonably — around 26.8% for the year on brokerage calculations — because AIA passes costs through and because a smaller volume of higher-value product can flatter realisations.[^5] But there is no honest way to describe a 14% volume decline as anything other than a demand event.

Management's attribution was specific and, importantly, mostly about customer behaviour rather than competitive loss: inventory correction by key customers, logistics disruption, delayed new conversions, a slower ramp-up in the mining liner sub-segment, and one customer lost to a competitor amid anti-dumping duty uncertainty.²⁰²¹

Now weigh it properly, because this is where most analysis stops one step early.

Is this evidence recent? Yes — it concluded seventeen months before this writing. Does it involve the same business and the same growth engine as the claim it tests? Yes — mining volumes, mining liners, and mining conversions are precisely the mechanism the thesis depends on. Is it large? Yes — a double-digit revenue decline and a 14% volume decline is a serious result, not noise. Was it remediated? Partly.

FY26 recovered to record revenue, record profit, record EBITDA of ₹1,744 crore, and a 39.5% EBITDA margin, with a record quarterly EBITDA of ₹502 crore in the March 2026 quarter.²²⁷ But volumes recovered only to 258,002 tonnes — barely above FY25 and still well below the FY24 level implied by a 14% decline.² The profit recovery was substantially a price, mix, and currency recovery, not a volume recovery.

That distinction is the whole finding. AIA's FY26 record year was earned on roughly the same tonnage as its disaster year.

And here is the older evidence that bounds the claim further. FY25 was not the first time trade policy cost AIA volume. In FY22 the company lost an estimated 35,000–40,000 tonnes from Canada and South Africa because of changes in customs duty structure and supply-chain disruption — offset at the time by roughly 30,000 tonnes gained from new customers.¹³ Add the Brazilian duties won by Magotteaux's petition, and a 9.6% dumping duty in the United States which AIA's customers have been absorbing, and a pattern emerges.[^5]²¹ A meaningful part of AIA's geographic footprint is shaped by trade policy — sometimes protecting it, more often obstructing it. That is not a moat. It is an exposure, and it runs in both directions.

Kunal Shah acknowledged the geography of the problem himself in August 2026, describing Brazil and Canada as markets "where already there are, you know, duty actions and an incumbent who's, you know, using the global tariff situation to their advantage."¹⁰

So what is the verdict on the thesis?

The FY25 evidence does not reject the conversion thesis. There is no sign that mines that converted to high-chrome converted back, no evidence of systematic share loss to Magotteaux or to Chinese entrants in high-chrome, and the Chilean win in the following year is affirmative evidence that conversions still happen in the hardest markets. The falsification test the claim needed to survive — did customers abandon the product — it survived.

But it decisively narrows the claim. The version of the story that says "structural conversion growth insulated from the mining cycle" is dead, killed by the company's own FY25. The version that survives is narrower and less exciting: high-chrome conversion is a real, multi-decade, still-early structural trend, and AIA participates in it with a genuine technical position — but the timing of that participation is controlled by customer capex cycles, customer inventory behaviour, trade policy, and trial durations that management has repeatedly underestimated. Growth is real and lumpy, not real and smooth.

The KPI that resolves the revised claim is simple and specific: mining segment volume, measured annually rather than quarterly, over the next two to three years. If mining tonnage compounds at a visible rate through FY27 and FY28, the structural reading holds and FY25 was an air pocket. If AIA delivers another year where total volume moves by low single-digit thousands of tonnes while profit is carried by realisations and currency, then the conversion engine is running slower than the story requires, and the company should be valued as a high-quality cyclical rather than a structural compounder.

A separate claim, separately tested: the balance sheet. The bull case leans heavily on financial strength, and that claim deserves its own falsification test rather than being taken as self-evident. The right test is not "is there debt today" but "what has this company done in its worst moments."

The record is unusually clean on this one. Share count has been essentially static — roughly 9.43 crore shares for years, reduced slightly by the 2024 buyback rather than increased by issuance.¹³¹ There has been no rights issue, no qualified placement, no convertible, and no episode of refinancing distress in the listed era. Borrowings have been trivial to nil throughout, including through FY21 and FY22 when profits fell and freight and input costs spiked.¹³ And critically, AIA financed its capacity expansions from internal accruals rather than from debt or equity, which is precisely why the announced-then-paused 520,000-tonne expansion cost shareholders nothing.⁴[^5]

So this claim survives its test intact — a rarer outcome than it sounds, and the strongest single element of the bull case. It is worth stating plainly because so much else in this analysis has to be qualified: AIA's financial position is not a narrative, it is a fact, and it means the company can be wrong about the pace of conversion for several more years without any solvency or dilution consequence for existing shareholders. What it cannot do is make the conversion happen faster.

One more piece of context sharpens the verdict rather than softening it. FY25 was not a mining-industry catastrophe. Copper and gold were not in freefall; the majors were not shutting mills. This was an ordinary period of customer destocking and deferred decisions — and it still took 14% off AIA's volume. A business whose growth engine can be stalled by an ordinary inventory cycle in its customer base is, by definition, more cycle-exposed than a business whose growth comes from its own product roadmap. That is not a criticism of AIA's product. It is a statement about where in the value chain the company sits: downstream of someone else's capital cycle, selling something the customer can defer for a quarter or two without immediate consequence.

The activist question that follows naturally is whether the FY25 experience should have changed anything about how the company is run — whether a business with this demonstrated cyclicality should be holding a quarter of its market value in cash while declining to distribute it, or whether that cash is precisely the right response to volume that arrives in lumps. Management has chosen the second answer.

That question will be settled substantially by where the company chooses to put its money next.

↑ Back to Top

VIII. Capital Deployment: Buybacks, Bolt-Ons, and the Ghana/China Bet

There is a counterfactual worth taking seriously. Sigdo Koppers built Magotteaux's global footprint substantially through acquisition — it took full ownership of the Belgian company in 2011 and now runs fifteen plants across the world.³ AIA, over the same period, built one manufacturing base in India and a network of sales-and-service entities, and topped up minority stakes in businesses it already partly owned.

That is a strategic choice, not an accident, and management defended it explicitly in August 2026. Asked why AIA would not simply build a plant in Chile or Peru to serve the market it says is most important, Kunal Shah's answer was structural: "manufacturing of our type of products requires a local ecosystem. I can't be one foundry of 10 in the whole country, right, and everything needs to be imported from India, it does not work." India, he argued, "remains a very, very attractive location for our type of products, which means reasonable cost, adequate manpower with the technical skills."¹⁰

There is something else in that answer that a careful listener should catch. Shah also said the bottleneck is not logistics at all: "I don't think we've come till to that point now, so I set up a factory, and the customer is saying I don't want to use your product because I think it'll fail, what do I do then?"¹⁰ In other words, management's own diagnosis is that the constraint on growth is customer acceptance, not proximity. That is a materially different problem from the one the overseas plants were designed to solve.

Which brings us to Ghana and China. The original strategic logic was clean. AIA had lived through freight costs going up fivefold and back down and up again, which made the total-cost argument to a mine impossible to hold steady. Lead times from India to South America were long. So the company announced two modular plants of 50,000 tonnes each — one in China to halve shipping time to South America, one in Ghana to serve West African gold — at a combined estimated cost of about $50 million, roughly ₹435 crore, built in stages to stay capital-efficient. As of late 2025, the China facility's first phase was expected to be commissioned within a year and Ghana approvals were targeted over the same period.[^5]

Here is where those plans actually stand, in management's own words on August 12, 2026. In Ghana: "we've identified the location, we are now in dialogue with the government to make sure that all what we understand and what we need is available." In China: "we have set up a small lab kind of a facility, we are now exploring the possibilities of what kind of, you know, infrastructure or investment we should make in a phase." And the summary judgement: "We have not shelved those plans. It is currently in a slow mode."¹⁰

Kunal Shah went further: "while the intent continues, I don't have meaningful progress to report on it, you know, for different reasons," adding that AIA remains "at the national stage" of considering manufacturing outside India and "for now, we continue to be, you know, based out of India for all practical considerations going forward."¹⁰

Apply the correct standard here. An announced plant is not a plant. A signed government dialogue is not a permit. A lab facility is not a foundry. Roughly a year after commissioning was expected to begin, one project has a site and a conversation and the other has a laboratory. This is real optionality — the strategic case for proximity to South American and West African mining has not been refuted — but it has produced no revenue, no capacity, and no confirmed timeline, and the original guidance has already slipped by a year with the goalposts now removed rather than moved. It should be valued as an unpriced option, not as capacity in the pipeline.

The other technical bet. Running in parallel is what AIA calls the New Generation Discharge System, or NGDS. Management describes it as an intervention in how material exits a mill, aimed at the three problems that matter most to a mine — throughput, power consumption, and falling metal recovery as ore grades decline. Kunal Shah called it "a sum total of everything that the company has done over these many years" and, in the Q4 FY26 results commentary, the company framed a South American customer's decision to order a second mill conversion as validation.¹⁰²⁷

Two things must be said about NGDS in the same breath. First, the strategic logic is genuinely strong: if AIA can bundle grinding media, liners, and discharge system into one solution, it converts three separate vendor relationships into one integrated one, which is exactly how switching costs get built. Management is explicit that this is the point — the goal is "stickiness," not a higher margin, and Shah said directly that the pricing metric and margin profile would not change.¹⁰ Second, it is in trials.

There is no standalone addressable market for it; Shah was emphatic that "there is no standalone TAM. I don't want to sell NGDS."¹⁰ Trials at small mills have gone well; medium and large are in progress; management asked for three or four quarters of "bland" updates and declined to disclose the outcome of the second South American trial at all, saying "to carve out one and say it has worked well, again, does not feed into it."¹⁰

Judge it against the company's own conversion record: AIA has a long history of technical capability and a demonstrated tendency to take longer than expected to turn that capability into tonnage. NGDS should be held at the same discount.

And now the capex signal. In August 2026, AIA raised FY27 capital expenditure guidance from ₹130 crore to ₹350–400 crore.¹² The composition is revealing. About ₹170–200 crore is for a dedicated plot of land and a new corporate house. Another ₹50–100 crore is for additional land for possible future brownfield or greenfield expansion. The balance covers maintenance, debottlenecking, and a hybrid solar-wind renewable power project on which about ₹30 crore was spent in the June quarter.¹⁰

Read that allocation carefully. Roughly half to two-thirds of a near-tripled capex budget is going into land and a head office — not into capacity, not into the overseas plants, not into the conversion engine. Meanwhile the existing Indian asset base runs at around 60% utilisation, and Majmudar told investors current capacity could "easily" support 300,000–350,000 tonnes without further building.¹²¹⁰

The one item in that budget with a clean, measurable payback is the smallest: a hybrid solar-wind project that became operational shortly before the August 2026 call, on which management expects the benefit to show up in coming quarters.¹⁰ Power is a major input for an energy-intensive casting and heat-treatment operation, and AIA has been chipping away at this for years — it previously installed nine wind turbine generators covering roughly 17% of its energy requirement.¹³ This is unglamorous, high-certainty capex of exactly the kind a mature manufacturer should be doing continuously, and it is the part of the FY27 budget least likely to disappoint.

There is a defensible argument for each of the larger items too. A company holding ₹4,500 crore of cash can afford a corporate house, and buying optionality on land ahead of demand is what AIA has always done — Majmudar's rationale was that "capacity creation itself takes time" and "I cannot run and say that I will set up a plant and then you come."¹⁰ But the pairing is what matters. In the same quarter, on the same call, this management raised spending sharply, directed the largest single component of it at a headquarters building, declined to give a volume number, declined to update a realisation number that reality had already overtaken by ₹15 a kilogram, and declined to say whether the flagship trial had worked.

Spend more, say less. That is not evidence of a governance problem — there is no related-party issue here, no leverage, no complexity, and a promoter taking ₹1.2 crore a year out of a ₹1,270 crore profit pool. But it is a change in the information investors receive, and it arrives at exactly the moment when the thesis most needs verification. It is the clearest thing to watch over the next two to three quarters, and it deserves to be treated as a live question rather than settled either way.

↑ Back to Top

IX. Bull vs. Bear, Risk Radar, and What to Watch

The bull case, stated at its strongest. AIA owns a dominant, cash-generative position in Indian cement grinding media that requires little capital to defend and funds everything else.¹¹

It is one of only two credible global suppliers of the high-chrome product category, with a genuine process-power advantage built over five decades and validated by its FY26 results — record revenue, record profit, a 39.5% EBITDA margin, and the highest per-kilogram realisation in company history.²¹² The addressable conversion opportunity in mining, on management's own framing, is roughly one to 1.5 million tonnes a year against current shipments of about a quarter of that.¹⁰ The balance sheet carries essentially no debt against ₹8,007 crore of reserves and ₹4,500–5,000 crore of cash — meaning AIA can fund any conversion opportunity that materialises without touching capital markets.¹¹⁰

The founder owns 58.5% and takes almost nothing out.¹²² The qualification cycle that makes new business slow to win makes existing business genuinely hard to displace, evidenced by a customer-loss base rate that management could count on one hand across a bad year.²¹ And there is a real, if unquantifiable, secular tailwind underneath all of it: ore grades are falling worldwide, which means miners must grind more rock to produce the same metal, which means more grinding media consumed per ounce and a stronger economic case for anything that improves recovery. Management has framed its whole current product effort around exactly that problem.¹⁰

The bear case, stated at its strongest. FY25 already demonstrated that "structural growth insulated from cycles" was an overstatement — a 14% volume decline in one year, driven by customer inventory behaviour and capex timing, is what cyclicality looks like.²¹ FY26's record profits were earned on essentially flat volume, meaning the recovery was priced, not sold.² The company has now declined to give volume guidance for two consecutive years while its own stated ambition of 30,000–40,000 tonnes of annual conversion additions produced 2,559 tonnes in FY26.[^5]² Trade policy has repeatedly bitten: 35,000–40,000 tonnes lost from Canada and South Africa in FY22, a 9.6% dumping duty in the US, and Brazilian duties won by Magotteaux's own petition.¹³²¹[^5]

Chinese producers have already proven they can go from nowhere to the largest global forged-media supply base and into Chile's two biggest mines, and the reassurance that they have not entered high-chrome is a statement about the present, which management itself flagged as conjecture about the future.¹⁰ The Ghana and China plants are a year behind with no revised timeline.¹⁰

Working capital is deteriorating — receivables at 90 days of sales in the June 2026 quarter against 73 a year earlier, with inventory days also up.¹² A near-tripled capex budget weighted toward land and a head office, announced alongside withheld guidance, is at minimum an odd combination.¹⁰ And the asset base runs at roughly 60% utilisation, which means the reported returns are being earned on a plant that is a third idle and a balance sheet a quarter of which is cash — a structure that flatters nothing and forgives nothing if volumes stay flat.¹²¹⁰

Risk radar — material mechanisms only.

Mining capex cyclicality. The mechanism is direct and was demonstrated in FY25: when copper and gold prices soften or mine budgets tighten, operators defer discretionary media upgrades and run down inventory, even when lifecycle economics favour conversion. AIA's growth is a discretionary line item in someone else's capital cycle.

Chinese competitive entry. The mechanism is price-led share capture, first consolidating forged media — already achieved — then potentially migrating up-market. The barrier is engineering and process knowledge, which is real but is the kind of barrier that erodes over years rather than holding permanently.

Trade policy and anti-dumping. The mechanism runs both ways. Duties imposed on AIA close markets and cost volume, as in Brazil and Canada. Duties protecting AIA, where they exist, can be reversed, immediately reopening low-cost import competition. Neither direction is under the company's control.

Input-cost volatility and freight. Ferrochrome, chrome ore, and steel scrap are the raw inputs; ocean freight sat at $8,000–9,000 a container in mid-2026 with transshipment ports congested and the Red Sea situation unresolved.¹⁰ AIA passes these through with a lag, which means margin compression in the quarter costs rise and margin expansion when they fall — a real driver of reported profitability that has nothing to do with the underlying business.

Execution risk in Ghana, China, and NGDS. Capital and management bandwidth committed to projects that have slipped, with no revenue and no confirmed timeline.

Geographic and shipping concentration. Because essentially all manufacturing sits in Gujarat and roughly three-quarters of sales are exports, every tonne AIA sells crosses an ocean. Container availability, transshipment congestion, and Red Sea routing were all cited as live constraints in mid-2026.¹⁰ Competitors that manufacture inside their end markets do not carry this exposure, which is a structural cost and reliability disadvantage rather than a cyclical one — and it is the exact problem the paused overseas plants were meant to address.

Note what is not on this list. There is no leverage risk, no refinancing wall, no meaningful cybersecurity or data-privacy exposure in a business that sells castings, and no obvious technological-disruption vector — the physics of comminution has not changed and no software is going to grind rock. The risks here are commercial, cyclical, and political, which is a narrower and more analysable set than most industrials carry.

Where a skeptical investor would push hardest. Not at governance — there is genuinely little to find. The promoter takes almost nothing, the balance sheet is unlevered, the corporate structure is simple, there is no portfolio of unrelated businesses to break up, and the auditor's reports carry no qualifications that have surfaced in public coverage.

An activist's angle here would be entirely about capital and disclosure: roughly ₹4,500–5,000 crore of cash sitting on the balance sheet earning treasury returns, mechanically dragging return on capital down; a single ₹500 crore buyback in the company's entire listed history; a dividend payout of about 12%; and a management that has now told investors twice in a row to wait a quarter for guidance while raising capex three-fold, largely for real estate.¹⁰¹ The response to each of those points individually is reasonable. The response to all of them together is harder to make.

The peer comparison is instructive here too. Sigdo Koppers deployed capital into Magotteaux and built a fifteen-plant global footprint that now sells 544,000 tonnes a year — more than double AIA's shipments — with local manufacturing embedded in the South American market AIA describes as its single largest opportunity.³ AIA's capital-light, India-only approach has produced a cleaner balance sheet and, by most measures, better returns on capital. It has not produced comparable tonnage growth, and it leaves the company shipping across oceans into a market where its main competitor already manufactures. Both strategies are defensible. Only one of them has been tested at scale in South America.

The three KPIs that actually matter. Not a dashboard — three things.

First, mining segment volume growth, measured year over year rather than quarter to quarter. This is the single best read on whether the conversion thesis is compounding or stalling, and management has said explicitly that it does not track a quarterly run rate — "it's annual" — so quarterly moves like the 5,400-tonne sequential dip in mining volume in June 2026 should be read as timing, not signal.¹⁰ Two consecutive years of visible mining tonnage growth would confirm the narrowed thesis; another flat year would not.

Second, per-kilogram realisation together with EBITDA margin. These have to be read as a pair, because realisation rising on freight and ferrochrome pass-through is not the same event as realisation rising on mix and pricing power. Realisation exceeded ₹180/kg in the June 2026 quarter while EBITDA margin compressed 411 basis points sequentially to 36.35% — a textbook illustration of why the two numbers must be watched together.¹²

Third, whether announced capital actually converts into capacity or revenue. Specifically: does the ₹350–400 crore of FY27 capex produce anything other than a corporate house and land, does the China facility move beyond a laboratory, and does the Ghana site progress from government dialogue to a permit? Announcement-to-commissioning is the honest measure of this management's execution on greenfield work, and the record so far in this cycle is one slipped year with no replacement date.

↑ Back to Top

X. Epilogue: The Lesson of a Quiet Compounder

There is a particular kind of company that financial media almost never writes about and that patient investors occasionally find: a business with a boring product, one founder, one factory complex, no debt, no drama, and a market position built so slowly that no single year explains it.

AIA Engineering is that. Nearly fifty years after Bhadresh Shah lost his entire capital on bad castings and pivoted to replacement parts for power plants, the company he built has become the world's second-largest producer of a product almost nobody outside heavy industry can name, selling into more than a hundred countries from a base in Gujarat, with a promoter who still owns most of it and takes a salary that would not raise eyebrows at a mid-sized private company.

It got there without a single transformative acquisition. There is no marquee deal in this story, no bet-the-company merger, no roll-up. The Australian and Peruvian transactions of the last three years are measured in single-digit millions of dollars, and one of them cost less than a used car.¹⁶¹⁷ What growth AIA has produced came from a slower and less narratable activity: sending engineers into mills, running trials that took months or years, and persuading one plant manager at a time that a more expensive ball is a cheaper ball.

There is a symmetry worth noticing in how the story began and where it now sits. Shah's first business failed because the castings were not good enough, and he recovered by choosing a market where the customer buys again every month. Half a century later, the company's constraint is the mirror image: the castings are demonstrably good enough — the Chilean copper conversion proved that in the hardest available market — and the binding constraint is persuading a new customer to try them for the first time. AIA solved the quality problem in the 1980s and the distribution problem in the 2000s. The problem it has not solved, and the one that determines the next decade, is the adoption problem.

That is also why the comparison with Magotteaux is more interesting than a market-share table suggests. Two companies with broadly similar products and broadly similar technical claims have chosen opposite answers to the adoption problem: one manufactures next to its customers on three continents and accepts the capital intensity that comes with it; the other manufactures in one country, keeps the balance sheet pristine, and ships. Neither has broken the forged-steel incumbency. Whichever approach eventually does will tell the industry something about whether high-chrome conversion is fundamentally a logistics problem or fundamentally a persuasion problem — and AIA, by its own management's account in August 2026, has now concluded it is the latter.¹⁰

The investing lesson, though, is not that patience always pays. It is the more uncomfortable one that FY25 taught, and that FY26 only partly repaired: a structural, multi-decade conversion story can have a large cyclical component sitting inside it, invisible until the cycle turns. AIA's mining pivot did not fail its stress test — customers did not revert to forged steel, the technical position held, and a Chilean copper major placed a $32.9 million order in the year that followed. But it emerged narrower than it went in. What survives is a real conversion opportunity, participated in by a company with a genuine technical edge, on a timetable that belongs to its customers rather than to itself.

Everything that matters from here is therefore measurable rather than rhetorical. Does mining tonnage compound, or does another flat year arrive dressed up in favourable realisations? Do the Chinese forged specialists stay in forged? Does a company holding ₹4,500 crore of cash and declining to forecast its own volumes deploy that capital into the conversion engine, or into land and a head office?

Management has asked investors to wait a quarter, and then another. That is not unreasonable for a business whose sales cycle is measured in years. But the request should be understood for what it is: an invitation to trust a narrative at precisely the moment its author has stopped putting numbers to it. The next few years of mining volume data — not the framing around them — will settle whether the duopoly with Magotteaux is a durable structure or simply a temporary arrangement between two specialists in a market that four-fifths of the world still buys on price.

AIA Engineering Limited

Table of Contents

AIA Engineering Limited: The Grinding Media Duopoly

I. Cold Open & Roadmap

II. Origins: A Foundry That Almost Failed (1979–2005)

III. The Cement Fortress: Building a Near-Monopoly at Home (2000s–2010s)

IV. The Mining Pivot: Betting Against Cheap Steel Balls (2010s–2020s)

V. How the Core Business Actually Wins (or Doesn't)

VI. Current Management: Ownership, Incentives, Capital Discipline

VII. The FY25 Stress Test: When the Structural Story Met a Real Slowdown

VIII. Capital Deployment: Buybacks, Bolt-Ons, and the Ghana/China Bet

IX. Bull vs. Bear, Risk Radar, and What to Watch

X. Epilogue: The Lesson of a Quiet Compounder

References