Contracted versus speculative expansion

Two companies announce the same plant on the same morning — the same product, the same capacity, the same capital expenditure, commissioning in the same year. On the page they are indistinguishable. Yet one of them has a signed buyer for every unit the plant will make before a single brick is laid, and the other has a hope that the market will be there when the plant opens. That one fact — invisible in the capex number, absent from the headline — decides whether the expansion compounds value for a decade or manufactures a glut that destroys it.

This is the distinction between contracted and speculative expansion. Contracted capacity is built against demand that already exists in writing: an offtake agreement, a power-purchase contract, a firm order backlog, an anchor customer, a long-term supply agreement. Speculative capacity is built against demand that is merely expected — "we see the market growing," "demand will materialise." Both are legitimate strategies. But they carry opposite risk, and the reader's job in this module is to tell them apart from the disclosures, because the market, dazzled by the size of the spend, routinely does not.

Module 086 established that a capex announcement means nothing on its own — you read it against the return the last cycle earned and whether the whole industry is building at once. Module 101 established that capacity is a physical fact, counted in tonnes or megawatts or rooms, not in rupees. This module adds the third question, and it is the one that most often separates a good expansion from a value-destroying one: is the output sold, or is it a bet?

The reason this matters so much is the lag. A plant takes years to build and more years to ramp, so the demand that justifies it must be forecast across that whole horizon. Contracting the offtake in advance collapses that forecast risk: the buyer, not the builder, has committed to take the volume, so utilisation is secured on day one and the J-curve — the loss-making ramp when full depreciation sits on low output — is short and shallow because the plant fills fast. Speculative capacity keeps the entire forecast risk on the builder's own balance sheet. If the demand appears, the plant earns; if it does not, the same asset sits idle, ramps for years, or sells into a price the glut has crushed.

The rupee capex figure the announcement leads with is the noise; the offtake behind it — signed, hoped-for, or absent — is the signal, and the crowd cheers the first while the answer lives in the second.

Contracted and speculative expansion leave different fingerprints in the disclosures. You are looking, in the annual report, the investor presentation and the concall, for whether a buyer is committed before the capacity exists:

• An offtake agreement — a buyer contractually commits, before the plant is built, to purchase a specified quantity of its future output. The strongest form is a take-or-pay clause, where the buyer pays for a minimum volume whether or not it actually lifts it, so the revenue is secured regardless of the buyer's own demand.
• A power-purchase agreement or concession — in power and infrastructure, a long-term contract fixes the buyer and often the tariff for fifteen to twenty-five years, making the project's revenue bankable before financial close.
• An anchor customer or anchor tenant — a large committed customer (or, in real estate, a lease-signed tenant) whose long-term contract underpins the economics of the new capacity and effectively de-risks the build.
• Order backlog covering the new capacity — a signed order book large enough that the expansion is being built to execute work already won, not work hoped for.
• A dedicated plant or supply agreement — in specialty chemicals, a block built for one molecule under a long-term contract with a named customer; the demand is locked to the asset.

Against these, the speculative tell is the absence of any of them, dressed as confidence: "we expect demand to materialise," "the market is growing at X%," "we are building ahead of the curve." Merchant capacity — capacity built to sell into the open market at prevailing prices — is speculative by construction, however strong today's price looks.

The discipline is to hunt for the counterparty. A contracted expansion can always name its buyer and point to the contract; a speculative one names a market and points to a forecast. When the disclosure describes the demand as a trend rather than a signed commitment, you are reading a bet — which may still be a good one, but must be judged as a bet, not banked as secured growth.

Two composite expansions, announced the same quarter, each guiding to roughly ₹1,000 crore of capex. [illustrative]

The first is a specialty-chemicals maker, Meridian Fine Chemicals illustrative, building a dedicated block for a single advanced intermediate. The presentation names the structure without naming the customer: a five-year supply agreement, a take-or-pay floor covering about 70% of the block's output, and a customer-funded portion of the plant cost. The offtake is signed before the build, so Meridian's utilisation is largely secured on commissioning and its ramp is short. The residual risk is real but narrow — customer concentration, and the 30% of capacity still sold merchant — not the existential "will anyone buy this" risk. This is contracted expansion, and the disclosure lets you see it.

The second is a metals producer, Vanshi Metals illustrative, doubling smelter capacity into a strong current price, output to be sold merchant. [illustrative] There is no offtake, no take-or-pay, no anchor — the case rests on a demand forecast and today's price. Open the CWIP and the industry context (086): three rivals announced similar expansions the same year, all reading the same high price. Vanshi's plant will commission in three years into the supply the whole industry is building now, and it carries the full demand risk itself. This is speculative expansion, and its identical ₹1,000 crore headline hides an opposite risk shape. An investor reading only the capex sees "two companies investing for growth"; one reading the offtake sees one plant pre-sold and one plant betting the market will still be there when the glut arrives.

The norm inverts by sector: in some industries capacity is contracted by default and in others it is speculative by default, so the same announcement — "we are doubling capacity" — should reassure you in one sector and worry you in another. The inversion is sharpest between regulated power, where a new plant without a signed buyer would barely be financed, and merchant commodity metals, where the identical addition is a bet on the cycle. Read the announcement against the sector's default, then check whether this particular expansion follows it or breaks it — a rare uncontracted power plant, or a rare contracted metals expansion, is the exception that changes the reading.

Knowing an expansion is contracted tells you the demand risk has been transferred; it does not tell you the contract will hold. An offtake agreement is only as good as the counterparty behind it — a take-or-pay from a fragile buyer, or a PPA with a discom that cannot pay on time, secures revenue on paper that may not arrive in cash. Contracted de-risks demand, not the credit of the buyer, and a chain of contracts is only as strong as its weakest party.

It does not tell you the price in the contract is a good one. Capacity can be fully contracted at a thin or loss-making margin — an order book won by underbidding, a PPA signed at a low tariff to secure financing. Contracted at a bad price is secured value destruction, not security; the volume is locked but the economics may not be worth locking. This is the same warning the order-book carries: a growing backlog won at loss-making margins is a warning, not a comfort.

And it does not make speculative expansion automatically wrong. A low-cost producer building merchant capacity counter-cyclically, when rivals are retrenching, can create enormous value precisely by carrying the demand risk others will not. The distinction sorts the risk shape; the judgement of whether the bet is worth taking still needs the cost-curve position, the balance-sheet strength to survive a wait, and the sector cycle. Contracted is safer, not always better.

The first trap is banking speculative capacity as secured growth. An expansion is announced, the model adds the full new capacity to future revenue at today's price, and the entire question of whether anyone has agreed to buy it is skipped. Building the forecast as though the plant is pre-sold, when it is a bet on the market, is how analysts and managements alike fund gluts with a straight face.

The second is reading a strong current price as proof the bet will pay. Merchant capacity earns the price that prevails when it commissions, years later, into the supply the strong price is tempting everyone to build now. The price that justifies the expansion is the very thing the expansion will erode — extrapolating today's price across the ramp is the commodity investor's recurring error.

The third is mistaking a hope dressed in confident language for a commitment. "We expect demand to materialise," delivered with conviction and a market-size chart, is not an offtake. The tell is that the demand is described as a trend and never as a counterparty — a signed buyer can always be named and pointed to; a market can only be forecast. When you cannot find the buyer in the disclosure, the buyer is a hope.

The fourth is assuming the sector norm without checking this expansion. Not every power plant is contracted and not every chemicals block is dedicated. Treating a merchant power project as if it had a PPA, or a standard multi-customer chemical as if it were a locked dedicated block, applies the sector's comfort to an expansion that has quietly broken from it — which is exactly where the norm-based reader is most exposed.