Zetwerk IPO: Sales are rising fast, but the company needs cash to keep up with demand

1. IPO Overview

Zetwerk Manufacturing Businesses is raising Rs 2,830 crore through new shares at Rs 143-151 a share, alongside an offer for sale of up to 144,525,036 shares by promoters and investors. Most of the fresh money will repay borrowings of the company and nine subsidiaries, with the balance left for acquisitions that are not yet identified and general corporate needs.

The point to notice is the mix. The sale by existing holders is large relative to the fresh raise, and the company remains loss-making on a consolidated basis with cash still being absorbed by working capital. Promoters held 15.86% before the offer on a fully diluted basis and are selling up to 28,461,344 shares, so public investors will own most of a business that has grown partly by buying other factories.

2. What the company does

Zetwerk sells a way to get things made without owning every factory that makes them. A customer gives drawings and a delivery date, Zetwerk gets the goods built across outside workshops and its own plants, and delivers finished products. It also bulk-buys metals and resells them to manufacturers.

Runs a universal factory stitched from outside workshops and own plants

The core idea is a universal factory. Instead of one large owned plant, Zetwerk directs spare capacity at thousands of independent suppliers as if they were one factory. Its software, Zetwerk OS, breaks AutoCAD drawings into parts, estimates cost from past orders, picks suppliers, tracks production and quality, and manages logistics.

Most volume is made by outside specialists in sheet-metal work, fabrication, moulding, machining, casting and assembly. Selectively, Zetwerk owns finishing and assembly plants for complex, high-value or certification-heavy work, or where being near a large customer helps. As of June 2026 it ran 26 own facilities across four countries, including India.

The company contrasts this with two older models. A vertically integrated maker owns dedicated plants and grows by building more plant. A contract maker builds to a brand's design in its own factory under bilateral contracts. Zetwerk calls itself a manufacturing platform, where assets stay with suppliers but a central system decides how they are used.

Earns a margin between customer price and supplier cost, plus a trading spread

In manufacturing, Zetwerk contracts as seller to the customer and as buyer from suppliers, usually on fixed-price orders that stop the supplier raising price later. It takes title to the goods and keeps the difference between what the customer pays and what procurement, manufacturing, logistics, quality checks and overheads cost.

In commodities, branded Terra91, it aggregates demand, buys ferrous and non-ferrous metals and petrochemicals such as aluminium, zinc, steel, copper and lead in bulk, and supplies manufacturers on a principal basis. It keeps a trading spread plus handling charges. The brand began as a way to get cheaper, faster inputs for its own network and then became a standalone sales line.

An order typically moves through enquiry and fixed-price quote, matching to several suppliers at once, execution with quality and logistics control, and delivery. Splitting one order across about seven suppliers on average in FY2026 allows parallel work in different places to shorten delivery time. For long engineering and installation scope, revenue is recognised over time by costs incurred, and expected losses are provided immediately.

Makes energy gear, precision parts and capital goods, and sells raw materials

Zetwerk reports two continuing segments. Manufacturing covers custom goods made through others and itself, including design, supply and installation. Ecosystem, or Terra91, covers sale of industrial commodities. A third line, Civil Infrastructure Works such as water pipelines and railway tracks, was approved for sale in early 2026 and is now shown as a discontinued operation.

Within Manufacturing, sales are split into three verticals. Energy Products covers solar modules and cells, trackers, transformers, frames, wires and cables, wind components and switchgear. Precision Products covers recurring parts such as hearables, wearables, smart watches, displays, printed circuit boards, battery parts, EV chargers and machined or moulded parts. Capital Goods covers project equipment such as pressure vessels, columns, reactors, heat exchangers, steel structures, girders, bridges and refinery or boiler parts.

Energy is the largest vertical, at about seven-tenths of manufacturing sales in FY2026. Precision and Capital Goods each contributed in the mid-teens that year, though the mix moves with large project deliveries.

The three-month period to June 2026, a stub that cannot be compared with a full year, showed Ecosystem at Rs 2,215.99 crore and Manufacturing at Rs 1,901.88 crore. That quarter flipped the full-year mix back toward commodities after a factory-led FY2026.

Serves startups to large industrials, mostly in India with a US leg

Customers range from young startups to mature industrials across utilities, renewables, consumer electronics, artificial intelligence infrastructure, aerospace, space and defence, oil and gas and industrial automation. Named buyers include Siemens Gamesa, Acer India, boAt, NextPower, CG Power, NTPC Renewable Energy, L&T MHI Power Boiler, DRDO's electronics laboratory, National Aluminium, Indian Oil, Mortenson, Nordex Acciona, Indian Air Force and Numaligarh Refinery.

In FY2026 the platform had 1,802 customers, with 1,613 in India and 189 outside, and buyers located across 29 countries. Manufacturing served 906 active customers through suppliers in 26 countries. The base included large listed and Fortune 500 names, which helps credibility but does not lock in orders because most buying is through short purchase orders or forecasts that can be changed.

India dominates sales, at about 83% of continuing revenue in the June 2026 stub and about 83% in FY2026. The USA was about 13% in that stub and about 14% in FY2026, with the rest spread across other countries. Manufacturing alone is more international, at about 30% from outside India in FY2026, because power equipment and precision parts travel more easily than bulk metals.

Wins repeat work by expanding basket, but buyers stay price sensitive

Zetwerk tries to grow by selling more categories to the same buyer, moving from one product to a full bill of material, and by mixing one-time project products with recurring components that original equipment makers reorder. Recurring components were about 51% of manufacturing transaction value in FY2026 and about 71% in the June 2026 stub. Design-led work, where Zetwerk also designs the product and not just builds to print, was about 14% in FY2026 and carries higher switching costs because replacing it needs re-engineering.

Repeat buying is high. Revenue from customers who had bought in a prior period was about 80% of manufacturing sales in FY2026 and about 89% in the June 2026 stub. Net revenue retention, which tracks how much retained buyers spent versus the prior year, was about 120% in FY2026 after dipping below 100% in FY2025. Average order value rose from Rs 4.59 crore in FY2024 to Rs 9.17 crore in FY2026, before easing to Rs 6.53 crore in the June 2026 stub.

Still, buyers can be tough. Some tenders go to the lowest bidder, customers periodically seek price cuts and stricter terms, and refusing can lose business. Top 10 customers were about 36% of continuing revenue in FY2026, so a cutback by one large buyer can swing sales and plant use.

Holds a large order book, while new plants ramp and old ones run hot

Manufacturing orders not yet delivered stood at about Rs 12,370.01 crore at March 2026 and about Rs 13,062.40 crore at June 2026, well above one year of factory sales. New manufacturing orders were about Rs 15,393.35 crore in FY2026, up from about Rs 11,327.11 crore in FY2025. That cover supports future sales if orders convert on time.

Own-plant use is uneven. Older plants run full, with the Kryfs Bhopal transformer-tank plant above 100% on extra shifts in FY2026. Newer or expanded plants built for future growth run light, with JZNEE torque tubes at about 4% and Winsharp televisions and monitors at about 6% in FY2026. The company frames low early use as planned ramp-up rather than weak demand.

Headcount has risen with scale, to 3,754 full-time staff at June 2026 from 2,544 at March 2024, with operations the largest function. Contract workers were about twice the permanent rolls, which fits a network that flexes with project flow.

3. Use of Funds

• Repayment or prepayment of borrowings of the company: Rs 1,300.00 crore.
• Repayment or prepayment through investments in nine subsidiaries, including Sharp Tanks, Terra91 International, Smile Electronics, Kryfs Power Components, Kryfs Transformers, Unimacts Global, Zetwerk Manufacturing SG, Zap91 Solar and Zet Town India: Rs 860.00 crore.
• Funding inorganic growth through unidentified and other strategic acquisitions, and general corporate purposes: balance of net proceeds, with general corporate and unidentified acquisitions capped at 35% of gross proceeds.
• The company will not receive any proceeds from the offer for sale of up to 144,525,036 shares; those proceeds go to the selling promoters, promoter group and investors.

4. Financials Overview

Sales dipped when low-margin trading was exited, then rebounded as factories grew. Factory profit on a cleaned-up basis rose each full year, but consolidated losses stayed large and cash from operations stayed negative. The June 2026 column is a three-month stub and is not comparable with full years.

Three-month stub.

Return on net worth, which scales owners' continuing loss against net worth, stayed negative in every period. Net asset value was Rs 31.26 a share at March 2026 and Rs 25.85 at June 2026, on a pre-issue Ind AS 33-adjusted share base. Manufacturing EBITDA margin held around 6%, while blended margins stayed thin because commodities earn about 1%.

5. What the financials tell us

The accounts tell a split story. Factories are executing better and orders are growing, but growth ties up cash, acquisitions cloud comparisons, and a business awaiting sale still owes more than it owns.

Sales dipped on a deliberate trading exit, then factories led the rebound

Total continuing sales fell in FY2025 and then rose by about two-fifths in FY2026. The fall came from Ecosystem, which shrank after the company quit apparel, textiles and other trading in FY2024, while Manufacturing still edged up by about 3%.

In FY2026 both parts grew, but factories grew far faster at about 50% and lifted Manufacturing to about 59% of sales. The June 2026 stub flipped again, with Ecosystem growing about 90% against Manufacturing at about 28%, so commodities were again the larger share at about 54%.

• The FY2025 dip was largely an exit, not a factory slump, because traded apparel sales fell from about Rs 1,172.20 crore to about Rs 59.20 crore.
• The FY2026 rebound was factory-led, with Energy Products doing most of the heavy lifting.
• The latest quarter leans again on low-margin commodities, so blended profitability needs watching.

For an investor, this means the top line is not one steady engine but two different ones. Factory growth drives value, while commodity swings move the total.

Headline losses hide improving factory profit, but a large non-cash charge dominates

Reported losses from continuing operations stayed large, at about Rs 852.50 crore in FY2024, about Rs 215.10 crore in FY2025 and about Rs 964.40 crore in FY2026 on a total-group basis. The FY2026 spike came mainly from exceptional items of about Rs 835.80 crore, dominated by a non-cash fair-value charge of about Rs 796.40 crore for raising some shareholders' diluted holdings.

Strip out exceptionals, share-based payments, joint-venture losses and other income, and the cleaner Adjusted EBITDA rose every full year, from about Rs 97 crore to about Rs 421.30 crore. Manufacturing EBITDA rose from about Rs 345.30 crore to about Rs 565.10 crore, with margin near 6% in FY2026. Gross margin improved from about 8% to about 12% over the same span and then held flat.

• Exceptionals sit above the bottom line, so they drive the headline loss more than factory operations do.
• Even on the cleaner view, owners remained loss-making in FY2026, so there is no positive earnings base to capitalise.
• Stubs are only three months and are not annualised, so the smaller June 2026 loss does not signal a full-year turn.

Judge factory execution on Adjusted and Manufacturing EBITDA, and judge the consolidated bottom line separately. The first is improving, the second is still deeply negative.

Growth consumes cash because customers pay later and suppliers are paid sooner

Cash used in operations was negative in all three full years and in both stubs, even as sales grew. In FY2026 the business earned about Rs 477.70 crore before working-capital changes but used about Rs 681.50 crore in operations. In the June 2026 quarter it earned only about Rs 3.80 crore before working capital and used about Rs 506.20 crore.

In FY2026 the drain came from higher other financial and current assets of about Rs 715 crore, higher customer dues of about Rs 336.10 crore and lower acceptances of about Rs 319.10 crore. In the June quarter it came from lower supplier dues of about Rs 318.90 crore, lower acceptances of about Rs 249.40 crore and higher stock of about Rs 167.60 crore, partly offset by lower receivables.

• Receivable days improved from about 63 to about 53 over FY2024-2026, but payable days fell faster, from about 89 to about 50, so the company funds more of the cycle itself.
• Net working capital rose to about Rs 1,374.20 crore at June 2026 from about Rs 710.60 crore at March 2026.
• Profits on paper therefore do not leave spare cash until billing, advances and supplier timing normalise.

This is the central funding tension. Sales growth first needs outside money, and the IPO debt repayment eases interest but does not fix the cash cycle.

New plants sit mostly empty while orders cover well over a year of sales

The company ran 26 own plants in FY2026 and at June 2026. Several newer sites built or bought for future growth operate at single-digit use, with JZNEE at about 4% and Winsharp at about 6% in FY2026, while Kryfs Bhopal ran above 100% on overtime. Low early use is presented as capacity built ahead of demand and phased ramp-up.

Demand cover looks strong. Confirmed but undelivered manufacturing orders more than doubled from FY2024 to about Rs 12,370 crore at March 2026, and reached about Rs 13,062.40 crore at June 2026, about 1.39 times FY2026 factory sales. Own-facility share of manufacturing sales rose from about 7% in FY2024 to about 33% in FY2026 and about 53% in the June stub, showing more work routed in-house.

• Money spent is not yet earning in new plants, so near-term sales depend on converting the book without overloading full plants.
• Utilisation is measured against capacity built for future growth, so low early rates are expected, but the pace of ramp is not yet proven.
• Execution risk sits in hiring, quality, shutdowns and logistics, not just order wins.

Debt and goodwill are large, and the business for sale owes more than it owns

Book borrowings on the capitalisation table were about Rs 1,942.20 crore at June 2026 against total equity of about Rs 4,094.10 crore. Total promises including non-fund facilities were far larger, at about Rs 5,208.10 crore at August 2026, so guarantees, letters of credit and bank limits still leave funding needs after the IPO. About two-thirds of those promises sit in the parent and about one-third in subsidiaries, with Unimacts Global the heaviest disclosed subsidiary borrower against thin, loss-making reserves.

Goodwill almost quadrupled after buying Kryfs, Smile, JZNEE, Angelina and taking control of Winsharp, to about Rs 1,067 crore at June 2026, about a quarter of equity. FY2026 therefore includes businesses absent in earlier years, so growth is part buying and writedowns would hit net worth. Contingent liabilities were about Rs 1,144.60 crore at June 2026, including corporate and bank guarantees.

The discontinued Civil Infrastructure line lost money every period and took a Rs 453 crore impairment for unbilled revenue and supplier advances in FY2026. Assets marked for sale of about Rs 810.10 crore at June 2026 sit below linked liabilities of about Rs 1,346.80 crore, leaving a net shortfall and possible further cash to exit.

• The planned Rs 1,300 crore plus Rs 860 crore IPO repayment would more than cover book borrowings but only about two-fifths of wider promises.
• Book debt and total promises are on different bases and dates and should not be mixed.
• Related loans and guarantees, including loans to associate Pinaka and borrowings from promoter-linked Creovate, add ties to watch, though no new material abuse is disclosed.

6. Valuation Analysis

At the top of the band, Rs 151 a share, the offer values the group at about 3.7 times post-issue book, which is the only workable lens since earnings are negative and peer multiples do not exist. That price pays for future conversion of a large order book and repair of the cash cycle rather than current profit or asset cover.

The multiple looks demanding rather than earned, given persistent operating cash absorption, thin blended margins and idle new capacity alongside goodwill and a for-sale drag. It would look fairer only if factory sales convert quickly into cash, with utilisation rising and working capital easing.

7. Peer Analysis

Basis: post-issue shares implied by Rs 2,830.00 crore fresh issue; P/E uses FY2026 owners' continuing loss; P/B and NAV use June 2026 equity plus fresh issue before expenses.

There is no listed peer, as the company states that no listed business in India or globally runs the same manufacturing-platform model.

That absence matters. Without a peer multiple, cheap or expensive cannot be judged relatively, and the negative P/E cannot show value. The economics split in two: factories earned about 6% EBITDA margin in FY2026 while commodities earned about 1%, yet commodities drove the latest quarter. Growth has also not produced cash, with operations absorbing cash every year while new plants at 4-6% use wait to be filled despite cover of well over a year of factory sales. Paying over three times book for continuing losses, goodwill at about a quarter of equity and a for-sale business with net liabilities therefore demands a cash-generative scale-up the accounts do not yet show.

8. Moat

Software plus thousands of suppliers lets it act like one large factory

What sets Zetwerk apart is orchestration, not ownership. Zetwerk OS unifies enquiry, design, supplier matching, production planning, quality and logistics across 6,979 suppliers in 26 countries and 26 own plants. That lets it split one order across several shops, run work in parallel and route around tariffs or capacity shocks.

Evidence of stickiness is real. About 89% of manufacturing sales in the June stub came from repeat buyers, and the order book covers well over a year of factory sales. No single small workshop can offer the same breadth, speed and multi-country backup. A pure marketplace without deep execution control cannot easily replicate quality assurance for complex, high-value work.

Big demand pools and policy shifts push work toward India

Outside forces help. The addressable market for manufacturing platforms is put at Rs 10.26-12.57 trillion in India and Rs 2.07-6.26 trillion in the USA in FY2026, with strong growth expected to FY2031. China-plus-one sourcing, Make in India, defence indigenisation and infrastructure spending shift orders toward Indian capacity. Long transformer lead times of 30-42 months let a faster 12-15 month promise win grid orders tied to renewables and data-centre power demand.

These tailwinds reach sales directly through new orders, larger baskets and multi-year cover, including take-or-pay terms where buyers commit minimum volumes. Growing design-led share adds switching costs because changing supplier needs re-engineering.

The edge is real but narrow and must keep compounding

Durability is moderate, not deep. Capacity is mostly rented, so scale must keep feeding data, quality records and supplier performance to sustain speed and cost. Vertically integrated makers and large contract manufacturers have deeper balance sheets and can add platform-like services, while customers can insource in downturns when price beats speed. If order growth stalls, data advantages fade and utilisation at new plants stays light, the moat wears quickly.

9. Risks

• Losses and cash burn: Continuing operations lost money every period and operations absorbed cash every year, forcing reliance on debt and equity. This is company-specific and has persisted, though pre-exceptional losses narrowed to FY2026.
• Supplier and input fragility: Top 10 suppliers were about 41% of continuing expenses in the June stub, mostly on short orders, with exposure to steel, aluminium, copper, freight, tariffs and forex. A delay or quality failure can hit delivery and margins.
• Customer concentration and cancellable orders: Top 10 buyers were about 29% of continuing sales in the June stub and about 36% in FY2026, largely on short orders that can be amended or cancelled. Losing one large programme would dent utilisation.
• Leverage, goodwill and guarantees: Total promises exceed equity, goodwill is about a quarter of equity after five deals, and guarantees to joint ventures and an associate are large. Writedowns or called guarantees would hit net worth and cash.
• Exit and legal overhang: The discontinued civil works line shows liabilities above assets held for sale and large impairments, alongside sizable contingents, a Bangladesh arbitration, pledged promoter shares and a charge sheet against a subsidiary. Crystallisation would need cash.
• Regulatory and plant ramp: Pending fire, building and labour approvals, a Greater Noida relocation and ramp of near-empty plants create execution risk. Industry-wide factors like US tariffs and commodity swings add volatility for all exporters.

10. Verdict

The call rests on factory-led growth with thin blended margins, cash still absorbed by working capital, idle new plants against strong order cover, and leverage plus goodwill alongside a for-sale net-liability drag. Those facts point to improving execution wrapped in weak cash conversion and acquisition-driven comparability.

At about 3.7 times post-issue book for a still loss-making, cash-absorbing group with no listed peer to anchor value, the price pays upfront for flawless conversion. The thesis works only if manufacturing orders turn quickly into collected cash with rising utilisation; it breaks if receivables and payables stay stretched, or goodwill and exit costs erode book value.