The output
What exists in year eight, with no further money from the company
The question a reader is entitled to have answered before anything else. Everything below this section is the method; this is the output.
The one-sentence version. In year eight, with no further money from the company, about 26,000 people in this corridor are still earning — 4,500 of them from craft and design — a community-owned company is trading at about a hundred and ten crore a year on contracts signed in its own name, and 1.36 lakh households sit on a register the community itself owns. That is the output. Everything else is how it is reached.
The case
The park will buy far more locally than it employs
The state's own estimate for a park of this class is about 500 direct jobs, for highly skilled professionals, against about 3,000 indirect jobs. AWS puts the wider figure higher still: more than 81,300 full-time jobs a year in the Maharashtra data centre supply chain by 2030.
Almost all the local opportunity is therefore in services, supplies and trades. A programme designed around that produces household income. One designed around the hope of direct data centre jobs does not, and loses credibility in its first year.
The design decision that follows: route a share of what the park, its contractors, Palava township and the Ring 3 corridor already buy through a community enterprise, at a scale their procurement teams can actually contract with. Investment becomes supplier payment, which continues after a grant ends.
Why this is a business question
The World Economic Forum reported in March 2026 that data centre growth risks stalling without a social licence to operate, citing a 125 per cent year-on-year rise in opposing sentiment and about 98 billion dollars of projects blocked or delayed — more than all previous quarters since 2023 combined. Indian sites in Andhra Pradesh and Telangana have seen resistance over land pooling, water, power subsidies and unmet job promises.
Palava is a favourable site: land already held, inside a planned township. Lower risk, not no risk.
What already exists, and the gap
Lodha Unnati runs from the Palava Skill Development Centre, Dombivli, inside this catchment: 900+ women registered, 400+ counselled, 200+ referred, 60 placed, against a goal of 3,000 into jobs in three years. AWS runs STEM centres, a data centre skilling programme with a Navi Mumbai engineering college, school nutrition and groundwater recharge in the region.
Capable programmes. The gap is not effort — it is that none of it is anchored to the park's own procurement, and none rests on a verified picture of these specific settlements. SETU would extend the existing centre, not compete with it.
Catchment
Three rings, 1.36 lakh households, one register
Select a ring to see what it is for.
Villages of Kalyan tehsil in this corridor include Nilaje, Katai, Usarghar, Ghesar, Hedutane, Davadi, Pisavali, Golivali, Sonarpada, Mangaon, Kole and Nandivali. Indicative, from public sources. Confirming which settlements fall in each ring is the first task of the baseline, because gaothan boundaries, newer chawl clusters and construction labour camps are frequently absent from official lists.
Evidence
What we know, and what we have to find out
Commerce
What the park, its contractors, the township and the corridor buy
Every line is a bill already paid to an outside vendor. Column three doubles as the equipment list for the apex and cluster facilities. Clusters specialise rather than all running the same three lines.
| What is bought | Who buys it | What the facility must have | Start |
|---|
Recommended opening lines in clusters one and two: construction-phase meals, housekeeping, and trade helper placements. All three have demand today, need modest capital, and produce visible income inside the first year. The network reaches eight to ten lines by year four. The full demand picture comes from the Phase 2 audit of about 180 buying organisations, not from this list.
The creative corridor
A creative corridor beside the digital one
Migration moves people, and it also moves skill. These are overwhelmingly migrant households, drawn from a dozen states over three decades. Tailoring, carpentry, metalwork, pottery, basketry, painting and the food crafts are already here — held by people working as helpers, loaders and guards because nobody has asked. The baseline’s tenth instrument finds out, household by household.
Programme
One team, one household register, four workstreams
Reporting
Five primary goals, seven secondary, no claim on seventeen
Select a goal for the targets and the indicators this programme would actually report. Shaded tiles are primary.
Every committed indicator has a baseline value measured in Phase 2 across 1.36 lakh households, appears in the same order every quarter, is supplied in a form the company's sustainability reporting can use, and is presented back to the community in Marathi.
Delivery
Seven phases, twenty-eight quarters, seven decision gates
Select a phase. Gates fall at the end of Q3, Q6, Q9, Q14, Q20, Q24 and Q28 — each one a test the programme can fail while there is still time to correct it.
Team
Five cluster teams, one apex — and the fall budgeted as carefully as the rise
- Months 1–3, eight. Director, two deputies, baseline research lead, finance, administration.
- Months 4–9, 34 core plus 140 field staff. The nine-instrument baseline across 1.36 lakh households.
- Months 10–18, 70 staff and 90 cadre. Fifty spokes formed, first twenty learning centres, apex and first two cluster facilities.
- Year 3, 185 staff and 450 cadre. All five clusters live; scale-up begins.
- Years 4–5, 240 staff and 626 cadre — the peak. Five cluster teams of about 24, an enterprise and apex team of about 60, 20 thematic leads, 22 shared services and an 18-person research cell. Roughly 720 of the 860 are from the catchment itself.
- Year 6, 180 falling to 120, with 520 cadre. Production, quality and supervisory staff move onto the enterprise payroll — about 400 people by the end of the year.
- Year 7, 55. The enterprise carries its own workforce; the agency mentors.
Underfunding the handover years is the commonest reason programmes of this kind collapse at the point of exit. Running the programme is itself one of the larger local employers here, and that is reported rather than quietly counted twice as an outcome.
Budget
Indicative seven-year cost, and why it is this size
Planning estimates at about ±25 per cent, at 2026 prices, for about sixty settlements and 1.36 lakh households in Rings 1 and 2, organised in five clusters. To be rebuilt after the baseline.
A programme costing two or three crore a year in a catchment of 5.6 lakh people is a pilot with a proposal attached. It produces the result the existing skilling centre here has produced: good outreach, negligible placement. Sized to the catchment and the build-out, the honest number is an order of magnitude higher — and the cost per outcome falls from about ₹1.30 lakh to about ₹47,000, because the fixed costs are the same either way.
Seven-year total by head
| Head | ₹ cr | % |
|---|
What is not in the budget
- Rent for the apex centre, the five cluster facilities or the fifty worksheds — space is assumed in kind. If rent must be paid, add roughly ₹1.8–2.2 crore a year.
- Power and water at the facilities, assumed provided with the space
- Equipment for work lines not yet chosen, which follows the baseline
- Land, civil construction or building works beyond fit-out
- Any provision for carbon credit revenue, which is not a near-term income line
If it has to be staged
- Stage by cluster, not by component. Cutting components removes the parts that make the rest work.
- Two clusters, about ₹84 cr. Full baseline across all of Rings 1 and 2 — not reduced, because the register is the platform — then the apex centre and two cluster facilities covering ~55,000 households, about 10,000 in better work.
- Four clusters, about ₹151 cr. Roughly 1.05 lakh households, about 20,000 in better work.
- Full programme, ₹190 cr. All of Rings 1 and 2, 1.36 lakh households, about 26,000 in better work, plus the materials and research work and the full evaluation cycle.
- Do not stage the baseline, the crèche network or the payment system. Cutting them removes the evidence, the women, and the members respectively.
Engagement
How engagement is actually executed
The first ninety days
- Days 1–15. Agreement signed including space and the data sharing protocol. Field base taken in the catchment. Courtesy calls before any field activity.
- Days 16–45. Settlement listing and walk-throughs. A transparency meeting in every settlement, including the honest statement that direct jobs are few.
- Days 46–90. Baseline field work with consent, 140 field staff deployed. Grievance mechanism live. Welfare board and health scheme camps, so a visible benefit arrives long before any facility exists.
Then a fixed calendar
- Weekly. Field team review; camp-gate contact with migrant workers, whose population turns over.
- Monthly. Community committee in every spoke with minutes displayed; frontline worker and contractor meetings.
- Quarterly. Steering committee; government departments; procurement pipeline review.
- Annually. Public review of results and accounts, in Marathi, in every settlement.
Engagement is measured, not asserted. Households contacted this quarter; attendance at community committees with women's attendance and speaking participation recorded separately; grievances closed within the stated time; and whether households can correctly state what the programme does and does not offer, tested on a sample each year.
The ask
What is requested
- Space in kind for one apex centre and five cluster facilities, within or adjacent to the park and across the corridor, written into the memorandum of understanding.
- Introductions to the anchor operators, the EPC contractors and the facility management companies, and consideration for empanelment on fair commercial terms once qualified.
- A procurement intent — a share of relevant spend across the park, the township and the corridor to be sourced from the community enterprise once quality and compliance standards are met.
- A batch security verification protocol, agreed in Phase 1, since clearance at volume sits on the critical path and is the commonest cause of a late first contract.
- A commitment of at least five years. A three-year commitment funds the evidence and the institutions and then stops at exactly the point where household income begins, which is worse than not starting.
- Agreement to the data sharing protocol — aggregate results shared, individual household data never transferred, the register passing to a community-held trust at exit.
- A senior representative on the steering committee, and an introduction to the anchor operators' community teams.
Sources and caveats
Park, policy, operator and existing-programme figures are drawn from public announcements and reporting and have not been confirmed with the companies. Acreage and capacity differ between the MoU announcement and later developer positioning, so the design does not depend on either being exact. Census figures are from 2011; health indicators are NFHS-5 (2019–21) for Thane district as a whole, which includes rural and tribal blocks. Catchment population and household counts are planning estimates built from those sources and adjusted for growth; confirming them is the first instrument of the baseline. Participation shares, attrition rates, targets and budget figures are planning estimates at about ±25 per cent, to be replaced by baseline findings.
- Business Standard — Maharashtra and Lodha MoU, September 2025; data centre land monetisation, July 2026
- Free Press Journal — Maharashtra cabinet approval of green integrated data centre parks, October 2024
- About Amazon India — AWS investment in Maharashtra; community programmes near its Mumbai data centres
- Lodha Group — Palava data centre park material; Lodha Unnati programme pages
- World Economic Forum, March 2026 — social licence and data centre growth; Rest of World, 2026 — local resistance in India
- Census of India 2011, Kalyan-Dombivali; NITI Aayog district profile for Thane drawing on NFHS-5
[Organisation name] · [Name, designation] · [phone, email] · Draft for discussion, not for circulation.