Cyient DLM Is Paying Today's Margin to Buy a $1 Billion Option The Tape Has Stopped Loving
A record order book, a deliberate 135bp EBITDA squeeze, and a design-led pipeline that only shows up in revenue after FY2029 — right as the market de-rates AI infrastructure.
CYIENTDLM.NS · Earnings Call · 2026-10-15
A growth quarter with a margin problem management won't call one
Cyient DLM's Q2 FY27 report (2026-10-15) reads like two stories wearing one set of numbers. The good story: revenue rose 26.5% year-on-year to INR 3,931 million, and H1 revenue grew 30.2%, putting the company broadly on the 30–35% trajectory it keeps guiding toward. The awkward story: EBITDA margin fell to 8.7% from 10.0% a year ago — a 135 basis point squeeze — and the headline profit flattered itself. Reported PAT of INR 340 million includes a one-time INR 182 million earn-out reversal; strip that out and normalized PAT was INR 158 million, up 25.8%, for a normalized margin of 4%.Management chose to narrate the margin as a purchase, not a mistake, and did it with unusual specificity. CFO R.M. Subramanian sized the drag directly: “it is marginal investments in terms of what we have done, which is in the range of 100-150 basis points. As a policy, today, we are not capitalizing any of the cost what we are doing.” — R.M. Subramanian, Chief Financial Officer · 2026-10-15 The conservatism is the point — product development for the build-to-spec programs runs straight through the P&L rather than being parked on the balance sheet, so there are no future amortization surprises.
Chairman Krishna Bodanapu then gave the sharpest, most falsifiable commitment on the call. Where the CFO framed marketing costs as forward-looking investment, the chairman made them a one-and-done:
That is the hinge of the whole thesis. If the sales and marketing engine is genuinely fully built and simply under-absorbed, then the next leg of revenue growth drops to the EBITDA line almost mechanically. If it isn't — if more selling investment keeps appearing — then the margin recovery is a story that keeps getting deferred a quarter at a time.We will not add any more sales and marketing costs till we at least hit about or at least net new, we will not add any more sales and marketing costs till we hit at least double our revenue on a quarterly basis. That's just the issue. It is a point-in-time issue that there is fully loaded.
The order book, and the option hiding behind it
The tangible evidence of demand is strong and repeated: a record order book of INR 28.1 billion, up roughly INR 2.1 billion sequentially, a book-to-bill of 1.6x, and six consecutive quarters of order momentum. Execution runs 18–24 months: “if you look at on the order book what we have, it is usually from 18-24 months. That is what the range.” — Santhosh Seshadri, Analyst · 2026-10-15 On a ~40% win ratio from a pipeline north of $1 billion, management flagged $400–500 million of order intake as the realistic take.The bigger, more interesting number is the one that isn't in the order book at all. Cyient's build-to-spec business — recently renamed electronic products — carries a roughly $1 billion opportunity across four programs that management deliberately excludes from backlog because the contracts aren't yet production-ready: “There are four major programs which we are working right now... All the four programs are really coming now currently close to $1 billion, which are not there currently in our order book.” — Rajendra Velagapudi, Managing Director and CEO · 2026-10-15 These are seven- to nine-year engagements with co-owned IP, prototypes in customer testing, and production starting around FY2029 with ramp into FY2030.
That future option is what the entire margin framework is staged against: roughly 9–11% EBITDA today in the "strengthen" phase, 11–13% in the FY2027–29 "expand" phase as build to spec scales, and 13–18% from FY2030 in the "transform" phase. The nearby proof point is box build, already 27% of revenue and growing 85% year-on-year — evidence the mix is genuinely migrating up the value chain rather than just being promised toward it.
Pivoting into a cooling consensus
Here is the tension a reader should sit with. Cyient is adding three new engines to its core four verticals: AI and data center infrastructure, semiconductor capital equipment, and robotics. Semiconductor is already contributing revenue; the other two are nascent. Asked directly how real they are, the CEO was candid rather than promotional: “on semiconductor capital equipment, we are already there... On the other two areas where I said earlier, I think we just started exploring it right now, and it may take some time for us to get to a number there.” — Rajendra Velagapudi, Managing Director and CEO · 2026-10-15So the company is pivoting toward a theme the tape has, for now, stopped rewarding. Across the last 90 and 30 days, the AI data centers cluster is one of the market's biggest decliners — dozens of tickers red, with adjacent names like high bandwidth memory and co package optics down sharply on the shorter windows. Yet over the trailing 360 days those same clusters were among the biggest winners. That is the signature of a theme that has run and is now cooling — exactly the environment in which a small-cap announcing its arrival feels late rather than early.
Nor is the positioning unique. In the same reporting window, TSMC discussed its AI demand outlook, Nanya talked AI infrastructure, and ASML spoke to high-volume manufacturing. Crowded consensus, not company-specific insight. The genuinely Cyient-unique angle is design-led manufacturing with co-owned IP and multi-year manufacturing responsibility — a model that is hard to copy and hard to accelerate.
There is a subtle cross-current, too: the same memory tightness that the global tape is pricing is showing up as a constraint on Cyient's own conversion — the chairman cited continued "supply chain shortages, longer lead times, and geopolitical uncertainties, especially driven by what is happening in the memory market" as the reason backlog is not translating to revenue faster.