1. Company overview & snapshot

Mahavir Die Casters is a private aluminium high-pressure die-casting manufacturer with a long operating history in Faridabad and Manesar. Public sources consistently describe it as one of the larger independent HPDC players in North India, with around 18,000 tons of annual casting capacity and multiple plants spanning Faridabad, IMT Manesar and now Binola on the Delhi-Jaipur corridor.

There has been meaningful change within the last 12 months. Karan Gupta was appointed Managing Director in June 2024, and Haryana consent records show activity at the Binola facility, including approval for consent to establish in February 2025 and consent to operate in March 2025. That makes the company especially interesting now: leadership transition plus footprint expansion often creates a window where management becomes more open to operational benchmarking and cost-control tools.

Unlike listed peers, Mahavir’s public narrative is light on sustainability rhetoric and heavy on manufacturing capability. That is useful. The right commercial framing is margin and plant control, not ESG. A company navigating new capacity and a generational or leadership handoff may respond well to a pilot that proves value quickly and stays outside the core production controls.

2. Energy profile

DISCOM / supply (name early): Faridabad, Manesar and Binola are likely DHBVN HT industrial connections; verify the service address and bill account per unit.

2.1 Bill band, tariff & demand

The working group band is ₹40–90 lakh/month [~] for a multi-plant HPDC operation, inferred from its operating profile rather than a disclosed invoice. Confirm two recent DHBVN bills from the pilot unit, including sanctioned demand, recorded MD, kWh, PF incentive / penalty, tariff category and ToD treatment. Establish whether each unit is separately billed and whether melting / holding is electrically or fuel driven.

No public ISO 50001 evidence was found. The company does show automotive quality credentials and has sufficient process discipline to win major automotive business, but energy management appears to be an operating sub-function rather than a public strategic narrative. That can actually help Stamped: the conversation can focus on bill reduction and plant-to-plant variance without first having to out-theory a formal corporate sustainability team.

Likely pain points include furnace holding losses, nonstandard startup/shutdown behavior, compressor and cooling utility drift, plant-wise MD differences, and hidden variation in kWh or rupees per kg casting across units. The expansion into Binola adds another layer: new or ramping sites often distort what “normal” looks like in the rest of the network.

3. Operations, equipment & digital stack

Mahavir’s process chain appears to include HPDC, CNC machining, liquid painting, sub-assembly, robotic deburring and in-house tool capabilities. That means it is not just a casting shop; it is a multi-stage industrial system with process, utility and finishing loads interacting across several plants.

Operations are likely multi-shift and throughput-driven. In this environment, the big rupee leaks are rarely mystery line items on the invoice; they come from ordinary behaviors that remain unpriced inside the plant, such as unnecessary furnace hold time, overlapping startup windows, inefficient compressed-air usage or different operating habits across sites.

Digital maturity is difficult to confirm publicly. There is no clear public evidence of site-wide SCADA or EMS deployment, but a supplier of this scale should have at least machine-level and maintenance data. The likely problem is not zero data; it is fragmented data that does not roll cleanly into one monthly cost narrative a managing director can act on.

4. Stamped Energy fit analysis

This is a strong Band A fit because the business combines energy-intensive operations, multiple North India sites and a current management/expansion transition. The best Stamped wedge is cross-site comparison anchored in the electricity bill: “Which plant or utility pattern is leaking the most power, and how do we prove the correction fast?”

The read-only architecture should also matter here. A private manufacturer adding capacity may resist anything that sounds like a disruptive IT or automation project. Stamped’s value is that it can sit on top of existing meters and plant data, avoid PLC writes, and still produce assigned rupee actions with a next-bill test.

Alternatives include internal engineering, occasional consultant studies and intuition-driven management reviews. Stamped becomes compelling if it helps a new MD tighten operational control without launching a large capex or digitization program.

4.1 ICP scorecard

Process intensity, Haryana geography and multi-site complexity pass. The ₹30 lakh monthly bill gate is highly likely but unverified at the pilot unit; data maturity and the current utility owner are unknown. This remains a 9/10 discovery priority, not a guaranteed commercial qualification.

4.2 Fit score rationale

HPDC, CNC, painting and several sites offer a credible opportunity to find avoidable MD and utility variance. Recent leadership and Binola activity provide timing but also justify a deliberately small, non-disruptive first ask.

4.3 Wedge (parser-critical)

The strongest wedge is: establish a one-plant DHBVN baseline, identify synchronized casting, compressor, cooling or paint-support loads that create avoidable MD, then verify the production-safe correction on the next invoice before comparing plants.

4.4 Objections & competitors

Expected alternatives are internal electrical teams, OEM dashboards and consultant audits. Ask whether they assign an action owner, quantify the rupee effect and reconcile it to MD / PF on the bill. Stamped is an action-and-verification layer, not a replacement for plant controls.

4.5 Pilot design

Choose one stable Faridabad or Manesar unit; obtain two DHBVN bills, production / shift context and read-only access to available interval or feeder data. Success is a ranked action, named owner, approved operating test and visible movement in the relevant bill metric. Pause if bill, data or intervention ownership is unavailable.

5. Before you reach out

• Verify the current decision-maker first; public records support Karan Gupta as MD, but reception confirmation still matters before sending personalized outreach.
• Confirm whether the initial wedge should be Faridabad only or a network view across Faridabad, Manesar and Binola.
• Ask whether new-capacity ramp at Binola is creating comparison noise in energy cost benchmarks across older plants.
• Lead with owner/MD language: “Which plant is leaking the most power in rupees?” rather than technical jargon.
• Do not assume a formal EMS exists; instead ask how monthly bills are currently reviewed and by whom.
• Probe for furnace-holding, compressor and paint-shop utility issues before talking about AI or analytics.
• Keep the first pilot narrow and non-threatening: one plant, one bill, 90 days, read-only.
• Be ready for a maintenance champion to emerge only after top-level interest; this is probably a top-down sale first, bottom-up validation second.

6. Risks, flags & sources

6.1 Integrity / controversy / regulatory (search explicitly)

No substantiated fraud, lawsuit outcome, tax raid, labour controversy, NGT / PCB enforcement action or promoter-integrity allegation was found in the reviewed company pages, registry-derived profiles, CRISIL item and consent references as of 2026-07-12. Searches included “Mahavir Die Casters controversy”, “Mahavir Die Casters pollution”, “Mahavir Die Casters NGT”, “Mahavir Die Casters lawsuit”, “Mahavir Die Casters fraud” and “Mahavir Die Casters labour”. Consent activity is permitting evidence, not evidence of a violation. Run MCA, HSPCB and litigation checks before contracting.

6.2 Data quality flags

• Revenue, employee and monthly electricity estimates are based on private-company databases and public website material rather than audited annual-report disclosure.
• Public information confirms multi-site operations but does not clearly quantify plant-wise capacity or electricity spend by unit.
• Public sources confirm MD appointment and Binola consents, but detailed plant-equipment disclosure is limited.

6.3 Sources consulted

• https://mahavircasters.com/about/
• https://mahavircasters.com/contact-us/
• https://mahavircasters.com/plant-5/
• https://in.linkedin.com/company/mahavir-die-casters-pvt-limited
• https://www.thecompanycheck.com/people/mahavir-die-casters/U29299HR1991PTC031331
• https://www.tofler.in/mahavir-die-casters-pvt-ltd/company/U29299HR1991PTC031331
• https://ocmms.nic.in/OCMMS_NEW/industryLogin.action
• https://www.crisilratings.com/mnt/winshare/Ratings/RatingList/RatingDocs/MahavirDieCastersPrivateLimited_September%2003_%202024_RR_350743.html

6.4 Expanded account diligence

Corporate, commercial and site map

Mahavir is best treated as a manufacturing network, not as one Faridabad shop. The public description of four operating facilities, the Faridabad registered presence, the established Manesar manufacturing footprint and the 2025 Binola consent record all point to a distributed operating model. That distinction matters for Stamped. A group-level electricity number would be interesting, but it is not an actionable starting point: tariff, sanctioned demand, production mix, meter hierarchy and the electrical/fuel split may differ at each site. The first discovery call should therefore ask for the legal entity on the relevant DHBVN bill, the exact plant address, which unit has a stable production mix, and whether the bill is paid centrally or by plant.

The company is private, so revenue, employee, capacity and charge data used in this file are third-party or company-asserted signals rather than audited financial disclosure. The ₹321 Cr FY2024 figure and roughly 18,000-ton annual capacity are useful for prioritisation, not for a commercial claim. Likewise, open-charge data is only a signal of bank-financed assets; it should not be framed as financial distress. On a call, use operational facts that the prospect can confirm—four sites, HPDC, machining, finishing and expansion—not a guessed P&L.

The HSPCB/OCMMS result for Binola is a genuine timing signal. It records a February 2025 CTE approval and a March 2025 CTO approval after an earlier CTE application was rejected in October 2024. It does not say why the prior application was rejected, whether conditions were imposed, or whether other Mahavir sites have equivalent current records. The appropriate account action is a quiet compliance check before contracting: request current consent certificates for the selected site and confirm the process scope. Do not use the earlier rejected application in prospect-facing outreach.

Process and energy model

An HPDC part typically passes through alloy receipt and melting/holding, die-casting, trimming and deburring, CNC machining, washing or surface preparation, painting and inspection/assembly. Not every stage necessarily sits in every Mahavir plant, and the public sources do not disclose the meter hierarchy. Still, this chain gives a practical discovery map. Casting cells can create large coincident electrical demand through hydraulic power packs, die-temperature control, cooling, vacuum, extraction and ancillary motors. Compressors and chillers often serve multiple cells, so a machine-level view without utility context can misattribute the cost. CNC and robotic deburring are more variable with throughput; paint and air-handling loads can remain elevated between production runs.

For the pilot, first separate electrically driven loads from thermal fuel. Electric melting or holding may be a major bill driver; where furnaces are PNG/LPG-fired, Stamped should not imply that electricity software will reduce fuel directly. It can still target demand overlap, cooling, compressors, pumps, extraction, curing support and off-shift auxiliary load. Ask the electrical lead to mark the major incomers, each feeder’s served loads, and any shared utility feeders before trying to calculate ₹ per casting. A production-normalised metric is useful only after confirming that alloy mix, casting weight, scrap/rework, shift pattern and finishing route are comparable.

The ₹40–90 lakh/month [~] working group band is deliberately broad and not an invoice claim. It is a prioritisation hypothesis for a multi-site HPDC business, and it can be wrong in either direction. The validation packet is two consecutive DHBVN HT bills for one plant, interval/MD where available, sanctioned demand, PF incentive or penalty lines, tariff schedule, and a simple shift/output calendar. The team should also ask whether the company has rooftop solar, open-access power, DG operation, or multiple connection numbers. These affect the counterfactual: Stamped should reconcile savings to the relevant grid component of the invoice, not claim credit for generation or production changes.

Digital readiness and implementation route

There is no verified public evidence of a named EMS, historian or SCADA vendor. That is a data-gap, not evidence of absence. A mature die caster may have machine OEM HMIs, PLC tags, energy meters, maintenance logs and utility readings that are simply not public. The opening technical question is therefore “what can be read safely today?” rather than “do you have SCADA?” Path A is read-only access to existing main and feeder meters plus shift/production context. Path B, if feeder data is not usable, starts with bill and main-meter patterns to identify the highest-value boundary and only then agrees a non-disruptive metering/data plan. No PLC writes, control changes or production interlocks should be proposed in the first meeting.

A credible 90-day design has four gates. In week one, nominate a plant sponsor and electrical owner, approve the one-plant boundary and collect bills. In weeks two to three, align meter intervals with production, startup and maintenance events; create a baseline that excludes obvious shutdowns. In weeks four to eight, issue a small prescription queue—such as staggered casting-cell ramps, compressor pressure/off-shift control, paint-support scheduling, or a PF investigation—with the owner, expected ₹ mechanism and safety/quality guardrail recorded. In weeks nine to twelve, compare against the relevant DHBVN invoice and production-normalised evidence, documenting what changed and what did not. A result is valid only if production quality, throughput and operating safety were maintained.

Buyer map, objections and contact discipline

Karan Gupta is the current outreach primary because registry-derived records indicate appointment as MD in June 2024. That should be confirmed by reception before use. The better day-to-day champion remains an electrical, utilities, maintenance or plant-manufacturing leader who owns the bill and can test a scheduling change. The MD’s job is to sponsor a bounded proof run and resolve cross-plant priority conflicts; an electrical leader’s job is to validate meters, tariffs and operating constraints. A finance contact may be valuable at close because invoice reconciliation must be accepted as the proof mechanism.

Anticipate three objections. “We already have machine dashboards” invites the question of whether they turn a demand spike into an assigned action and reconcile it to the bill. “We cannot disturb production” is answered by read-only data access and production-approved, reversible sequencing tests. “We have solar / an electrical team” is not a reason to dismiss the account: solar changes the grid draw; it does not automatically stop idle utility loads or peak overlap. In each case, avoid saying Stamped replaces the EMS, maintenance team, APFC panel or OEM controls. It is the action and bill-verification layer above them.

Expanded discovery and risk controls

Before sending technical material, ask: Which unit is currently most stable? Does each plant have a separate DHBVN bill? What are the sanctioned and recorded demand values? Which furnaces are electric versus fuel-fired? Which feeder covers compressors, cooling, paint and casting support? Are there recurring PF penalties, maximum-demand excursions or tariff timing concerns? Who can approve a safe test, and how is output/quality protected? These questions qualify the opportunity without requesting sensitive production recipes.

6.5 Research conclusion

Mahavir remains the strongest of these six targets on physical energy opportunity: multi-site HPDC, process depth and a recent site/leadership transition support a cross-plant control narrative. The commercial risk is not lack of possible savings; it is getting a clean first boundary in a complex private group. Lead with one Haryana plant, one DHBVN bill and one owner-backed operational question. Only expand to benchmarking after a bill-linked action has been verified.

6.6 Call-planning worksheet

The first twenty minutes should produce decisions, not a generic product demo. Start by asking the plant name and legal billing entity, then whether its current main bill is DHBVN HT and who can share two redacted invoices. Ask the electrical lead to describe a recent demand or PF surprise in their own language. Then ask which supporting systems—compressors, cooling, extraction, paint or furnace auxiliaries—remain live when casting output falls. This identifies a testable operating boundary without requesting proprietary die, customer or part-cost data.

If the plant says it already has an EMS, ask whether it can answer four operational questions: which specific load created the maximum-demand interval; who owned the corrective action; what production constraint was checked; and whether the action changed the following invoice. A “yes” does not disqualify Stamped—it may mean the integration route is easier. A “no” establishes the gap without criticising the team. If data is unavailable, offer a bill-and-meter-map review rather than promising cross-site analytics.

The expected pilot cadence is deliberately conservative. No action should change die temperature, casting parameters, safety procedures, emissions equipment or quality checks. The plant chooses the action and stop conditions. Stamped’s role is to rank evidence, attach a rupee mechanism, record the owner and reconcile outcomes. This makes the proof run compatible with an automotive supplier’s quality discipline and avoids the false expectation of a black-box optimisation project.