Tonbo Imaging IPO: sells battlefield night vision in via tenders

1. IPO Overview

Tonbo Imaging India is selling up to 12,174,046 shares at Rs 665 to Rs 700 a share through a 100% offer for sale, which means existing shareholders sell and the company receives nothing. At the top of the band, Rs 700 a share, the sale is worth about Rs 852.18 crore and values the company at about Rs 4,009.75 crore on its post-issue base.

The two things to notice are the structure and the timing. There is no fresh issue to cut debt or fund stock, even though cash is locked in inventory and short-term bank loans are high. A clutch of new investors bought about 59.10 lakh shares at Rs 700 in late September 2026 from early backers CEAQ India and CEAQ Singapore, so the cap equals the last private price.

2. What the company does

Designs heat cameras and aiming systems in Bengaluru

Tonbo makes electronics that let soldiers and machines see heat rather than light. Its cameras see at night and through smoke, dust and camouflage, and its software then tracks a target, shares the picture and helps aim or steer a weapon.

The range runs from sights a soldier carries to payloads that ride on vehicles, aircraft, ships and missiles. Named products include Spartan and Arjun thermal weapon sights, Bloodhound and Gazehound binoculars, Wolfpack panoramic imagers, Avenger gimbals that stabilise cameras on aircraft, ELPEOS targeting units for remote weapon stations, and TRAP seekers that guide missiles. A drone-killer based on high-power microwaves, called WaveStrike with a TRex sensor, is still in development and has booked no sales.

Tonbo calls itself an original equipment manufacturer, which means it owns the design and brand for the finished system. It keeps optics, infrared engines, video processing and tracking software in its Bengaluru research centre and outsources heavy assembly to outside electronics factories. It says it owns 100% of its technology since March 2024 and that its portfolio is free of ITAR, America's defence export controls, which lets it sell to more countries without US permission.

Governments and contractors pay after tenders and trials

Buyers are armies, navies, air forces, police and border forces, plus large defence contractors, called system integrators, that bolt Tonbo's module into a tank, drone, ship or missile they deliver.

There is no subscription or rental stream. A ministry or contractor issues a tender, Tonbo bids a fixed price built from detectors, lenses, displays and factory time, proves the product in trials, and then delivers a batch. Revenue is booked when control passes, usually on despatch or delivery. Contracts are short-term and non-recurring, so Tonbo keeps any saving if it builds cheaper but absorbs any overrun in material, labour, currency or rework.

Cash arrives much later. It takes six to 24 months from order to finished product, plus weeks for delivery and acceptance, plus years of warranty. Tonbo buys detectors upfront, funds the factory build, gives bank guarantees of typically 3% to 5% of contract value for one to five years, and then waits to be paid.

Soldier sights still bring most sales, but vehicle and missile payloads grow fast

The business has two live lines and one future bet. Tactical systems are finished products a soldier carries. Platform systems are modules someone else integrates, plus a move toward complete loitering munitions. Directed energy, the drone-killer, has no revenue yet.

Tactical was well over half of sales in every period, but its share is falling as platforms rise from almost nothing to about two-fifths of sales.

Home market now dominates after Europe order ends

India was about 93% of FY26 sales after Europe collapsed. A large European order executed across FY24 and FY25 had no follow-on in FY26, while West Asia disruption pushed other export shipments into FY27.

The customer mix swung just as hard. Foreign governments were 57% of sales in FY25 and 3% in FY26, while the Indian government rose to 71% and private integrators to 24%.

Concentration is stark. The largest customer was 44% of sales in FY24, 57% in FY25, 14% in FY26 and 55% in the June 2026 quarter. The top ten were over 77% in every full year and 98% in the June quarter. Customers served were 25 in FY24, 30 in FY25, 44 in FY26 and 12 in the June quarter.

Order book grows while bidding gets harder

Firm orders, backed by binding documents, stood at Rs 314.24 crore at June 2026, up from the FY25 low. Another Rs 118.15 crore arrived between July and September 2026. Platforms now make up almost two-thirds of the book, and domestic orders are about three-quarters.

Bidding shows pressure. Bids submitted doubled to 131 in FY26, but the win rate slipped each year.

Two outside factories build almost everything

Kaynes Technology and Avalon Technology build most output under non-exclusive terms. Tonbo proves a golden sample in Bengaluru, then the partner assembles to its drawings.

In FY26 Kaynes made about 45% of units and Avalon about 28%. In the June quarter Avalon made 87% of the small lot. A third unnamed partner built 30% of units in FY24 and then disappeared.

Research headcount triples while inputs stay imported

Permanent staff rose from 167 to 339 in two years, and the research team rose from 41 to 133 engineers. Research cost rose from Rs 4.24 crore to Rs 25.86 crore.

What decides growth is winning the next tender and then delivering it. Detectors, displays, optics and germanium come from a narrow global base, with international suppliers at 64% of purchases in FY26. Inventories and receivables therefore swing with order timing, and second-half weighting is heavy, with about 79% of FY26 sales in the second half.

3. Use of Funds

• Listing benefits only; the stated object is to list the shares and enable the sale by selling shareholders.
• No fresh capital is raised, so no amount is earmarked for debt repayment, working capital, capex or research.
• All proceeds, net of offer expenses, go to the selling shareholders, including promoters, CEAQ India, CEAQ Singapore and others, in proportion to shares sold.

4. Financials Overview

Three-month stub.

The trend is falling sales with resilient-looking margins and collapsing returns. Revenue fell about 23% in FY26 after growing in FY25, EBITDA held near 29% but PAT slipped, and return on equity fell from about 53% to under 10%. The June quarter, a seasonally light three months, made a loss on very low sales while staff and amortisation costs continued. FY24 is standalone with no subsidiaries, later periods are consolidated with two tiny foreign units.

5. What the financials tell us

Tonbo now owns its designs and builds through others, but FY26 exports slipped into FY27 so sales fell, reported profit was helped by releasing old warranty cover, and cash sits in stock funded by bank loans. The listing adds no cash, so recovery rests on shipping a concentrated order book.

Sales fell because shipments slipped, not because orders vanished

Revenue fell from about Rs 469.10 crore to about Rs 362.60 crore in FY26. Exports collapsed from about two-thirds of sales to about one-twentieth, with Europe down to almost nothing after a large order finished with no follow-on.

The company links the drop to West Asia disruption that delayed international execution and pushed expected orders into FY27. A Rs 51.44 crore order received in FY26 could not ship and sits in the June book for FY27. The order book therefore rose even as sales fell, which looks like timing rather than cancellation.

• For an investor, this bunches FY27 on a few delayed export shipments plus domestic tenders.
• The risk is that timing becomes loss if the customer defers again or buys elsewhere.

About one-fifth of FY26 profit came from old warranty cover

Reported EBITDA was about Rs 104.60 crore and PAT about Rs 50.90 crore in FY26. That includes a net write-back of about Rs 14.20 crore of warranty provisions made in earlier years, against net charges in the prior two years.

The release added about Rs 18.80 crore to EBITDA and about five points to margin. Without it, EBITDA would be about Rs 85.80 crore and PAT about Rs 40.20 crore. Write-offs of customer dues also fell sharply, to about Rs 3.30 crore from about Rs 19.10 crore, so trading did improve, but the headline margin overstates what current sales earned.

• This matters because the multiple in section 6 looks cheaper on the reported number than on clean earnings.
• Watch whether the warranty line reverts to a normal yearly charge.

Cash is locked in stock and late bills, funded by bank loans

Profit stayed positive in FY25 and FY26 while cash from operations was negative in both years and in the June quarter. Money sits in two places: stock built for the delayed export order and expected FY27 business, and bills raised late in the year.

Stock rose from about Rs 111.70 crore to about Rs 319.60 crore by June 2026, while receivables spiked to about Rs 211.70 crore at March 2026 because much of the year's sales were billed in the last quarter. Customers then paid down about Rs 114.80 crore by June, but stock grew again, so the June quarter stayed cash-negative on tiny sales. Working capital days rose from 152 to 280 to 517, with the June figure of 1,454 days on a 91-day base not comparable to full years.

The gap is carried by working-capital loans. Cash was only about Rs 7.10 crore at June against much larger short-term borrowings, with fund-based limits already heavily used. The offer brings no cash because it is only a sale by holders.

• Further delay in shipping the stocked order or collecting late bills means more borrowing.
• Conversion of stock into sales and then cash is the single cash test for FY27.

Buying the technology ended royalties but left a fixed yearly charge

In March 2024 Tonbo bought all core technology from CEAQ Singapore for about Rs 183.20 crore. Royalty of about Rs 10.50 crore in FY24 then stopped, but the bought technology is written off over ten years.

Amortisation on intangibles rose from about Rs 11.40 crore to about Rs 25 crore, and intangibles of about Rs 167 crore at June are about one-fifth of assets. Owning designs did not end buying parts from the same group. Project material from CEAQ rose to about Rs 79 crore in FY26, about 28% of purchases, with about Rs 17.20 crore in the June quarter.

• Technology ownership removes a variable royalty but replaces it with a fixed charge that hurts when sales fall.
• Supply still runs through CEAQ for critical parts, alongside bought detectors and displays.

Outsourcing and more research have not yet spread sales

Design in Bengaluru with bulk build by Kaynes and Avalon keeps own plant small, but deliveries rest on two makers. Kaynes made about 45% of FY26 units and Avalon about 28%, while research cost rose sixfold and the research team tripled to support platforms, seekers and gimbals.

Sales have not spread. The largest customer swung from 57% to 14% to 55% in the June quarter, tactical systems are still about 58% of sales, and the win rate slipped from about 41% to about 34%. Low fixed cost and rising research help build new products, but near-term sales still hinge on winning a few tenders and on two factories delivering on time.

6. Valuation Analysis

At the top of the band, on its post-issue base, Tonbo is priced at about 79 times its FY26 reported profit, using the earnings multiple that fits a profitable hardware maker. That is a small discount to peers priced around 84 times, but the discount is earned rather than a bargain given lumpy sales, warranty-helped profit and cash locked in stock.

The price looks fair only if delayed exports ship on time, collections speed up and domestic tenders convert without margin collapse.

7. Peer Analysis

Tonbo post-issue P/E is on FY26 consolidated profit; peers are October 7 closes on FY26 diluted earnings, with Data Patterns on a standalone basis and the rest consolidated, as disclosed.

Tonbo really compares with Paras, Data Patterns, Astra and Zen, mid-cap defence electronics makers, rather than Bharat Electronics, a Rs 27,610.10 crore Navratna prime. None is a pure thermal-imaging peer. Paras makes optics and payloads, Data and Astra make radar and electronics around sensors, and Zen uses imaging inside simulators and anti-drone systems.

Three differences matter. First, Tonbo shrank about 23% in FY26 while Paras, Data and Astra grew, because a large European order ended with no follow-on and other exports slipped to FY27. Second, clean earning power is lower than it looks, with about one-fifth of FY26 profit from a warranty release, which lifts the same band to about 100 times clean profit, a premium to the median. Third, cash conversion is weaker, with 517 working-capital days against 365 for Data Patterns, the only peer disclosing it, plus negative operating cash funded by bank loans.

These problems are largely Tonbo-specific, not industry-wide. Peers converted larger domestic books, held higher margins and returns, and kept growing, while Zen also fell in FY26 but from a prior surge and at far higher margins. The small headline discount therefore turns into a premium on clean earnings without superior growth, margin, return or scale to support it.

8. Moat

Owns its designs and can sell where US-controlled rivals cannot

Two things genuinely set Tonbo apart. It owns all its technology since the 2024 buyout, while many Indian peers still rely on licensed production or joint ventures. And its imaging stack is built outside US controls, so it can bid in markets where American, Israeli or European kit needs permission. Evidence is acceptance: about 28,000 systems across 24 countries, the largest domestic thermal-imaging share by value over FY25-FY26, and 94% of India's thermal-imaging export volumes by units over FY24-FY26.

Copying this is hard but not impossible. It needs years of field references, trials with armies and integrators, and an India-based design chain that stays outside US rules. Five US patents assigned from the group help, though some patent and trademark transfers are still pending approval.

Policy and conflict demand help order flow

Outside forces help. Make in India and self-reliance rules favour indigenously designed kit, with domestic-content thresholds that suit Tonbo's Bengaluru design and local assembly. The defence ministry's push toward 70% self-reliance by 2027 opens sight, seeker and vehicle programmes where Tonbo's gimbals and fire controls fit. Border tensions and fast-track emergency buys create sudden spikes for night vision and counter-drone kit.

How each reaches sales is through tenders. Indigenisation puts Tonbo on the qualified list, while emergencies pull forward orders that would otherwise take years.

Edge is real but only moderately durable

This is an edge in design and market access, not in manufacturing scale or supply. Tonbo outsources most build to two factories under non-exclusive terms, buys detectors, displays and germanium from few global suppliers, and faces price-led tenders with low switching costs once rivals qualify. Larger primes can outspend it on research and fund longer payment terms.

The moat lasts only if design keeps shrinking size, weight and power ahead of rivals and if export references keep converting into repeat tenders rather than one-off spikes.

9. Risks

• One customer moves the year. Top-ten share exceeds 77% every year and foreign-government share swung from 57% to 3% in a year. Losing or deferring one export tender cuts sales and leaves research and staff costs uncovered. This is Tonbo-specific, not peer-wide.

• Bought sensors and two factories can stall delivery. Top-ten suppliers were 69% of expenses in FY26 and over 100% in the June quarter on advance buying, while Kaynes and Avalon build most units. A detector shortage, price rise or factory slip inflates stock, now about Rs 319.60 crore, and delays revenue. Peers with broader sourcing and plants are less exposed.

• Exports face licences and local-content walls. Each export needs a licence, and European funded buys cap non-EU content at 35%. The Europe collapse shows how fast visibility fades when one order ends. Industry-wide for exporters, but sharper for Tonbo given its export swing.

• Cash strain forces more bank borrowing. With 517 working-capital days, negative operating cash for two years plus the June quarter, and no IPO money to the company, further delay means higher limits, more guarantees and financing cost. Guarantees outstanding already exceed Rs 31.20 crore, with bank guarantees of about Rs 78.90 crore counted in contingent liabilities.

• Profit quality and governance overhangs. The FY26 warranty release flatters margins, while a related-party technology buyout, a Singapore court suit naming the company, complaints to regulators, a promoter-leased office and promoter-guaranteed loans add governance noise. These are idiosyncratic and could weigh on listing sentiment even if operations recover.

10. Verdict

The call rests on four load-bearing facts developed above: sales fell while the order book grew on delayed exports; about one-fifth of FY26 profit came from releasing old warranty cover; cash is locked in stock funded by bank loans with no IPO proceeds to the company; and sales still rest on a few customers and two outside factories. Together they show owned technology and design strength, but weak near-term earnings quality and balance-sheet strain.

At about 79 times reported profit and about 100 times clean profit, the price demands a smooth FY27 rebound that the cash and concentration evidence does not yet de-risk. For the thesis to work, the Rs 314.24 crore book plus new wins must ship and turn into cash, cutting stock, receivables and short-term debt. It breaks if the delayed export slips again, domestic tenders price margins down, or warranty and write-offs revert, leaving a premium multiple on a small, lumpy, cash-hungry maker.