Two businesses.
One balance sheet.
A ₹11,500 crore bet.

The question. Deepak Nitrite earned high returns for a decade on capital that barely appeared as capital expenditure. It is now spending roughly twice its FY2026 equity on greenfield plant. The investment question is whether the returns survive the change in method.

The earnings picture. Consolidated EBITDA has fallen every year since FY2023 — ₹1,337, ₹1,199, ₹1,176, ₹1,041 crore — profit after tax from ₹852 to ₹551 crore, and return on capital employed from above 30% to 14–15%. Over the same years the announced programme grew from ₹1,500 crore to ₹11,500 crore and the balance sheet went from debt-free to a stated peak project debt near ₹6,800 crore. Then Q1 FY2027 was the best quarter the company has reported: revenue ₹2,592 crore, EBITDA ₹554 crore at 21%, profit after tax ₹345 crore.

Whether that quarter is the old business returning or the new one starting remains unresolved. What follows is the interpretation of the evidence, and the last section says what it does not.

A valuation exists on the interactive DCF model and gives ₹684 a share against a market price of ₹1,611.95 on 22 September 2026. No investment view is formed, for reasons stated under What is not known.

What actually sets the earnings

Two segments with genuinely different economics. Four years of audited segment data, and the annual record contradicts what eighteen earnings calls suggest.

The calls make Phenolics sound like the volatile commodity chain and Advanced Intermediates like the steady contract business. At the annual level it is the other way round. And the capital went to the falling segment: of ₹1,029 crore of FY2026 segment capital expenditure, ₹1,004 crore went into Advanced Intermediates, whose assets grew from ₹1,995 to ₹3,468 crore across four years while its EBIT fell from ₹555 to ₹107 crore.

Plant operations complicate the phenol-cycle explanation

The evidence does not support a simple explanation in which rising volume offsets a falling spread to hold margins constant. A margin is a ratio, so volume largely cancels out. The available observations favour a focus on plant availability and conversion economics, while the small sample limits any causal conclusion.

The available phenol spread series now runs FY2020 to June 2025 — 380, 807, 662, 629, ~426, ~464 dollars a tonne, alongside separately reported five-year statistics of an average near 500 and peak near 700. The latter peak cannot be the maximum of the displayed series, which includes 807; the periods or measurement bases have not been reconciled. Set it against the margin in the five quarters where both are observed:

The reported correlation is −0.04. Q3 FY2025 had a higher spread than either adjacent quarter, yet a lower margin than Q4 FY2025. The highest spread in this five-quarter table is 528 in Q1 FY2023; the lowest margin is 8% in Q1 FY2026. On the three consecutive quarters sharing one basis and one publisher it is −0.60: the spread stays inside a 12% band while the margin runs 9, 16, 8.

Management explains Q3 FY2025 without mentioning the spread — an annual plant shutdown, "there is an inventory cost, there is a shutdown cost", and the imports it invited, buyers stocking "to compensate for a perceived extended shutdown." Shutdowns recur roughly every eighteen months. And in August 2024 management states the priority outright: "more than spreads, it is important to see how we are converting things."

The working interpretation is that domestic import-parity pricing and plant operations help explain the stable annual margin. Its unit economics vary about half as much as the spread. Other influences on margins include plant availability, the import response to an outage, and conversion into acetone, IPA and now MIBK/MIBC. Ambit named a fall below $650/t as a risk to estimates in January 2022; the spread has been below that in every period since Q1 FY2023 and reached 350–360 in Q4 FY2024, while annual margins remained within the 12.0–13.5% band. That history does not establish that margins cannot fall.

The unresolved part: Q4 FY2026 at 20% and Q1 FY2027 at 24% are the two highest quarters in the eighteen-quarter record, they are consecutive, nothing before them exceeded 16%, and the available spread series stops before both.

Advanced Intermediates is where the earnings actually went

This segment lost ₹448 crore of EBIT in four years — more than the entire decline in consolidated profit after tax. Revenue fell 16% and flattened; EBIT fell 81%. This is consistent with a major price-and-cost compression; revenue alone cannot establish that volumes or the order book were intact — anyone modelling a recovery here is modelling margin. Management's five stated causes, in the order they appeared:

1. Customer destocking. In August 2023: "Despite inventory destocking by our customers, our margins in AI remained steady." Within two quarters they were not, and the explanation kept being extended.
2. Chinese overcapacity, then dumping. "significant overcapacity that has been built in China"; "persistently underpriced product availability from China has prevented a broader recovery"; and specifically, "Deepak manufactures a little less than 1 lakh tonnes of sodium nitrite, but China has 5x that capacity." Named products: sodium nitrite and DASDA. Potentially structural; the permanence of the capacity reset is unresolved.
3. Raw-material cost the contract book would not pass through, stated as a mechanism: "my OBA prices do not move in line with toluene prices." The contract structure that steadied this segment on the way up delayed the pass-through on the way down.
4. A freight disadvantage into its main export market. "the customer in Europe has to essentially pay less for freight if the material is coming from China as compared to India." Roughly half this segment is exported.
5. Self-inflicted, volunteered in February 2026. "the pricing pressure and the margin pressure, to be completely honest, were a little bit self-inflicted... because we decided tactically to be aggressive in placing our product in the market." On para-nitroaniline: "it was anyways a soft market. I think by increasing our production and supply to the market, we may have softened it a little bit more." That is management saying it depressed its own realisation to hold share — and the same instinct is building ₹11,500 crore of capacity.

And a sixth that is not deterioration. From FY2026 the segment carries pre-operative expenses of Deepak Chem Tech projects that are not yet running, "around INR 15 crore for the quarter", confirmed for Q1 and Q2 FY2026. At least ₹30 crore of pre-operative expense depressed FY2026 segment EBIT of ₹107 crore — an expense equivalent to 28% of the reported EBIT. Add it back and the FY2026 margin is about 5.4% rather than 4.2%. Whether commissioning removes the drag cleanly is a forecast to test, rather than an established outcome.

Management's target is unchanged: "around 17%, 18% is the target... on a standalone basis", said in February 2025 when the quarter had printed 3%.

The method that made the returns, and what replaces it

Phenol capacity was commissioned at 200,000 tonnes in 2018 and reached about 250,000 by August 2022, 300,000 by FY2024, 350,000 by Q3 FY2025, with 400,000 targeted. Utilisation against the original nameplate ran 118% in Q4 FY2022 and 150% by Q4 FY2024. Management called the capital cost of that expansion "not a lot at all" and named it as the moat: expanding an existing phenol plant "costs a fraction" of a greenfield one.

What "a fraction" meant, once: 2020–21 notes price a doubling of the 30,000-tonne IPA line at ₹50 crore from internal accruals, against about ₹300 crore of revenue expected from a line that size — roughly six rupees of revenue per rupee of capital, against management's own 2:1 project criterion. Both figures are third-hand and the ₹300 crore was guidance at pandemic-inflated IPA prices, so treat six as an upper bound. Even discounted hard, it is the only quantification in the research of the mechanism that made the company's returns.

The polycarbonate programme reverses that method. It is greenfield-scale, on licensed or acquired technology, and must earn a return on capital that is actually deployed. The evidence reviewed does not establish that the company has done that before at this scale.

The capital programme, and how it was announced

These rows are not the same object — a board approval, a state memorandum, an asset purchase, a total outlay — so no multiple across them means anything. What holds is the like-for-like: the polycarbonate programme alone went from ₹5,000–7,000 crore in August 2022 to ₹9,000 crore by November 2025, and every restatement moved up. Two executives gave materially different totals for the same thing in the same August 2022 call, the earliest instance of a recurring pattern — a further-capex figure given as ₹1,000–1,500 crore and then ₹1,500–1,600 crore by the same speaker in one Q1 FY2027 call, against a stated ₹3,200 crore total that reconciles with neither.

Cumulative spend into Deepak Chem Tech was ₹599 crore at Q2 FY2024, ₹656 crore at Q3, ₹709 crore at Q4 FY2024. The commitment grew far faster than the spend, which is why the funding question arrived late and all at once.

The only outside estimate of the programme's product economics is Avendus Spark's mid-cycle assumption: polycarbonate at $975 a tonne against phenol at $500 and bisphenol A at $310. One house's number, not a disclosure, and the only such estimate in the research reviewed.

Funding

Debt-free on a consolidated basis through FY2024 — Deepak Phenolics prepaid ₹161 crore across FY2023 and cleared the rest by Q3 FY2024 — with consolidated debt about ₹1,000 crore by March 2025.

In August 2025 management gave peak debt to equity of 1.5× and peak debt of ₹7,000–7,500 crore. In August 2026 it gave peak debt to equity as below 1×, with project debt about ₹6,800 crore. The change was not explained. Whether it reflects a larger equity base, a smaller draw or a different definition is unresolved and is an important unresolved financing question.

Against that: stated cash generation near ₹1,000 crore a year, cash of about ₹900 crore, net worth ₹6,214 crore at Q1 FY2027. The ₹1,000 crore is operating profit before working-capital movements. Audited net cash from operations after working capital and tax was ₹539 crore in FY2026, against capital expenditure of ₹1,185 crore. That gap is the funding problem in audited numbers. The Praxair HyCO agreement takes the industrial-gas step off the balance sheet and genuinely reduces the requirement; it should not be netted silently into the headline.

Management's record

The numbers they gave. In November 2022 Sanjay Upadhyay set a normalised Phenolics EBITDA band of 16–22%, with above 25–27% "may not be sustainable" and below 16% "may not be the right margin"; MIBK/MIBC at 20–22%, phenol as a base product at 16–18%. February 2024 reaffirmed a blended consolidated band of 16–20%. What happened: consolidated EBITDA margin was 15%, 14% and 13% across FY2024–FY2026, below the band in all three years, before 21% in Q1 FY2027. The band was not wrong so much as untested at the bottom — set in year two of the window, then three years spent below it.

The forecasting record.

Management admitted in FY2025 that project delays cost return on capital, and volunteered in Q2 FY2026 that Deepak Chem Tech projects had slipped. The honest reading is not that this management is unreliable: its project timelines have consistently moved right and its capital numbers consistently up, and both should be discounted rather than taken at the stated date.

Where they have been candid, because it cuts the other way. Maulik Mehta gave a self-critical account of phenol capacity planning, saying consultants' demand assumptions led the company to size the original plant wrongly. He declined an analyst's bullish ROCE scenario in Q4 FY2022. He conceded FY2023 Advanced Intermediates growth was all price and no volume when asked directly. He admitted progress toward the 20–22% AI margin target was slower than guided. He volunteered the self-inflicted pricing pressure quoted above. And management has consistently refused to give figures it did not have — spreads, project IRRs, polycarbonate margins, phenol volumes — rather than inventing them.

Valuation, and what the price requires

The model gives ₹684 against ₹1,611.95. Read backwards, holding everything else at default, the price implies:

The two implied operating margins exceed the four-year audited history. The cost of equity is a valuation assumption, not a company operating outcome. Asset turnover has a historical comparison: the IPA debottleneck ran at roughly 6×. But that was debottlenecking an existing plant with its land and utilities paid for, which is structurally not what a greenfield polycarbonate chain does — management's own criterion for backward integration, ~2:1, is a third of the debottlenecking rate. Taken one variable at a time, the price implies unusually strong operating assumptions or a lower discount rate. Combinations of smaller changes can also reconcile the model with the price; none of these single-variable thresholds is individually necessary.

Terminal value is 111.6% of enterprise value. The explicit period is cash-negative while the ₹11,500 crore is spent, so the present value of the explicit years is negative ₹1,225 crore. ₹684 is the output of the terminal assumptions, not an independent estimate.

The street's record, both directions. Systematix ₹675 (Jul 2020, shares at 529); InCred Reduce ₹1,745 (Sep 2021, shares at 2,416); Edelweiss Buy ₹3,030 (4 Jan 2022) and Yes Securities Reduce ₹2,100 (5 Jan 2022) — two houses initiating 930 rupees apart in the same week; Avendus Spark Buy ₹2,500 (Mar 2024, shares at 2,110). At ₹1,611.95 on 22 September 2026, with an unchanged 136.4 million share count, the shares are below the listed 2021–2024 targets, above the ₹675 target from 2020, and 37% below the January 2022 price. An unnamed house also put FY2022 EPS at 45.6 against an audited ₹1,067 crore, about ₹78 a share: low by roughly 40%. The street under-forecast the cycle in 2020 and over-forecast the company in 2022 and 2024. The market has been marking this business down for four years while earnings fell, and the model still says it is dear.

What would have to be true

Single-variable routes to the price include: Phenolics margin re-rates structurally above anything in four audited years and holds; the polycarbonate chain earns 3.36× asset turnover against management’s roughly 2× criterion; Advanced Intermediates recovers past its best year; or the cost of equity is approximately three percentage points below the model’s default. These are separate sensitivities, not a claim that one must occur: a combination of smaller changes can also support the price. For the model to be too pessimistic, the cheapest routes are that the Q4 FY2026 and Q1 FY2027 Phenolics margins prove structural rather than cyclical, and that the ₹30 crore-plus pre-operative expense drag falls on commissioning without being offset by new depreciation or operating costs while the portion of the Advanced Intermediates decline attributable to customer destocking unwinds.

What is not known

• What the phenol spread did after June 2025. The available series stops there, and Q4 FY2026 and Q1 FY2027 carry the two highest Phenolics margins on record. Whether that surge is cyclical or operational is the single most valuable unresolved fact about this company.
• Whether Chinese sodium-nitrite and DASDA capacity is a permanent reset of the Advanced Intermediates margin. If it is, the pre-2024 margin is not the right anchor for any recovery. Indian anti-dumping investigations were initiated in January 2025 on some of these products; their outcome is public and has not been assessed in this research.
• Whether the pre-operative expense reverses cleanly on commissioning. The test is the FY2027 and FY2028 segment notes: the notes should distinguish removal of pre-operative charges from replacement depreciation, operating costs and other earnings movements. EBIT alone cannot settle the explanation.
• Whether this business is competitive without anti-dumping duties. The second-hand 2020–21 notes contain both views without independent evidence for either — one arguing the company already exports to China below international prices, the other that the promoter "knows how to lobby and will get ADD on all his products". Eighteen calls never address it directly. Whichever is right decides whether the moat is a cost position or a regulatory one. Export realisations against international prices would largely settle it.
• Why peak debt to equity moved from 1.5× to below 1× with no explanation.
• The standalone-versus-Advanced-Intermediates perimeter, left unreconciled twice and deferred offline once.

How to interpret the valuation. A view has in fact been assembled: the default assumptions produce ₹684 against the dated ₹1,611.95 market price; different operating assumptions, discount rates or combinations could reconcile them. The conclusion remains provisional for a specific reason. Fifteen of twenty-eight model parameters are analyst assumptions. Two — the Phenolics margin and the Advanced Intermediates recovery — have been reassessed as the evidence developed. They remain material sources of uncertainty. Human review of these assumptions and external calibration remain outstanding. Confirming or changing either margin assumption could materially change the conclusion; ₹684 is an illustrative output under the stated defaults, not an endorsed target price. This limitation is material.

Evidence and source quality

The research draws on four layers. Audited statements take precedence where sources disagree.

1. Twenty-two statutory and annual documents, including the Integrated Annual Reports for FY2024–FY2026, provide four years of audited consolidated statements and a full two-segment split. The model’s statement inputs come from the annual reports. Seventeen results filings have unreliable extracted text, including six image-only documents, and were not used for numerical model inputs.
2. Eighteen consecutive earnings calls, Q4 FY2022 through Q1 FY2027, provide management’s explanations and guidance.
3. Twenty-five sell-side documents include B&K’s chemicals channel checks, Avendus Spark’s initiation and dailies, and work by InCred, Edelweiss, Yes Securities, Ambit, Kotak and Systematix. The phenol-spread observations and broker targets come from these. Commodity tracking and target-price forecasting are different products; using the former does not endorse the latter.
4. Historical research notes, May 2020–October 2021, relay two calls whose original transcripts were not available. These are second-hand accounts, not independently verified evidence.

The source collection spans 131 documents. Specific distinctions matter: Q1 FY2027 Phenolics revenue was ₹1,775 crore and EBIT was ₹418 crore; the regional peer-utilisation comparison was approximately 60–70%; and Oman investment of ₹11 crore in Q4 FY2024 was part of ₹27 crore cumulative investment.

The segment interpretation rests on the audited statements; the debottlenecking comparison rests on second-hand historical notes; and the spread series rests on broker research. The analysis was prepared on 22 September 2026 with AI assistance. Human review of the assumptions and external valuation calibration remain outstanding.

Primary references: FY2026 Integrated Annual Report, FY2024 Integrated Annual Report, and company financial results and earnings calls.