Story of a Stock · Field notes
Journal
Updates land in the Telegram group → (opens in a new tab)The flagship lane
Company Stories
44stories · 37 cos
Yash HighvoltageYASHHVOLTLatestStory 1 · 21 Oct 2026
Yash Highvoltage: A factory with two opening days
Read the story →- Avalon TechnologiesStory 1Avalon Technologies: Both workshops have a better room. The work keeps landing in the other oneRead →
- Bhagyanagar IndiaStory 1Bhagyanagar India: It keeps five paise of every copper rupee. The shop is moving to a new nameRead →
- Pondy OxidesStory 1Pondy Oxides: The lead press ran slow and earned more a tonne. The copper press that matters is still being builtRead →
- Unimech AerospaceStory 1Unimech Aerospace: It built a hangar for twice its fleet. The first planes in it were boughtRead →
The Notebook
- No.
01Sector mapsTwelve contract factories: earnings grew in both cohorts. The multiple moved in only oneTwelve Indian electronics makers, split by what they build. The five AC and consumer makers turn 4–10% margins into cash; the seven specialists earn 10–20% and are owed it. Both trade near 80–85 times earnings. Since January the specialists are 2.5x with the P/E doing most of it; the AC makers are flat. Then the chain, and which way each specialist is moving.EMSElectronics manufacturing - No.
02ProductWe scored 121 earnings calls. Then the market voted.We scored all 121 Q4 FY26 earnings calls, then watched the stocks for two months. Strong calls ran ahead of the market and kept going; weak ones fell behind and stayed there — plus an honest accounting of how much of that edge is hindsight.121-company testScore vs price - No.
03ProductA stock can rally and get cheaperTwo weeks ago I started keeping every valuation reading instead of overwriting it. The biggest movers since include a stock whose price didn't budge and one that rose 9% while getting cheaper — because valuation moves for two different reasons, and only one of them is the price.Valuation scoreImplied growth - No.
04ProductIs the quarterly score telling us anything?I stress-tested our quarterly scores against real profits and real stock prices across 18 companies. What the score is, what it isn't, and the one pattern that paid every time.18-company backtestScore vs returns - No.
05How I investWhat is my criteria?What I'm building is secondary. What I'm really trying to do is solve my investing decisions — and that comes down to one question: what is my criteria?My criteriaConviction - No.
06How I investHow the analysis becomes a decisionThe machinery behind a single call on a stock: seven elements, a trust-and-understanding gate, four component scores, one composite, and finally — with direction and reward asymmetry — a decision.7 elementsComposite score - No.
07How I investSwing variables: tracking the story between quartersAn experiment: tracking each company's story the way you'd track an EPS beat — expected vs delivered, beat/match/miss — and finding the few variables that can swing the whole trajectory.Swing variablesBeat / match / miss - No.
08ProductMaking it worth your timeAlmost every change to the portal right now comes back to one idea: a working investor can't read everything. The portal's job is to earn the time you give it.Attention budgetWhat to cut - No.
09ProductWhat is Story of a Stock?The wedge, the portfolio problem I'm building it to solve, and the ladder from raw data to an actual buy/sell decision.The wedgeRaw data → decision
Journal
What I’m building, how I think as an investor, and the companies I’m digging into — with every number traceable back to a transcript, a deck, or a broker note.
Company Stories
Plate 01 — Company Stories38 stories · 36 companies
Yash HighvoltageYASHHVOLTYash Highvoltage: A factory with two opening days
Bhagyanagar IndiaBHAGYANGRBhagyanagar India: It keeps five paise of every copper rupee. The shop is moving to a new name
Pondy OxidesPOCLPondy Oxides: The lead press ran slow and earned more a tonne. The copper press that matters is still being builtMore stories
Latest 5 of 35EarlierBefore 12 Oct5 stories
- Unimech AerospaceUnimech Aerospace: It built a hangar for twice its fleet. The first planes in it were boughtUnimech Aerospace makes the tools jet-engine makers build with, at a 70% gross margin, and used its IPO to triple capacity. FY26 revenue fell 1% with the plants about half used, most of FY27's step comes from Hobel Bellows, bought in April 2026, and the price implies about 72% earnings growth.
- ideaForgeideaForge: It farms on the Army's rain. The tank it filled lasts until DecemberideaForge makes drones mostly for the Indian state, and its revenue follows the Army's emergency orders: ₹314 Cr in FY24, ₹161.2 Cr in FY25 and ₹226.1 Cr in FY26, 62% of it in one quarter. The ₹310 Cr book that carries FY27 is due by December, the cash cycle has stretched to 369 days, and the price sits above our model value even at 100% growth, against a 23.5% base case.
- Aeroflex IndustriesStory 2 of 2Aeroflex Industries: The pipe-fitter found one big contractor, and now works to the contractor's clockAeroflex Industries' liquid-cooling skids went from 1.9% to 22.3% of revenue in three quarters, all sold to one buyer, while the line that makes them has moved its finish date twice. It has met 3 of its 4 closed financial guides, return on capital fell from 36% to 19%, and the price implies about 54.5% earnings growth a year, above our 50% bull case.
- KRN Heat ExchangerKRN Heat Exchanger: It built a hall six times the old one. The first season filled about 15% of the seatsKRN Heat Exchanger ran its first plant near full, used IPO money to build one six times larger, and filled about 15% of it in FY26 against a 20-25% guide. It has met 4 of 5 closed money guides, and the price implies about 73% earnings growth a year for two years, right on our base case.
- Acutaas ChemicalsAcutaas Chemicals: One crop paid for two new fields. The harvest is no longer weighed in publicAcutaas Chemicals doubled its EBITDA margin from 17.9% in FY24 to 35.9% in FY26 on contract (CDMO) work, and is putting the proceeds into battery and chip chemicals that earned almost nothing in FY26. It has met 8 of its 12 closed financial guides while 21 project dates slipped, and the price implies about 44% growth a year for two years.
Plate 02 — Head to head
Head to head
Two listed peers in the same business, read side by side from their own filings.
Same industrySpecialist electronics manufacturingAvalon TechnologiesAVALON versus Centum ElectronicsCENTUMHead to headBoth workshops have a better room. The work keeps landing in the other oneAvalon Technologies and Centum Electronics both blend to about 12% EBITDA from one room that pays well and one that pays less. Avalon's loss-making US plant went from 12% to 28% of revenue in two years. Centum's design work is 51.5% of its order book but 28.5% of revenue. Avalon grew 46% in FY26, Centum 25%. The market values Avalon at 9.6 times sales and Centum at 6.3.Read the comparison →Same industryPharma contract manufacturingLaurus LabsLAURUSLABS versus Neuland LaboratoriesNEULANDLABHead to headLaurus Labs is building the kitchen Neuland already runs. Neuland's runs on five chefsNeuland Laboratories already earns 67% of its revenue making other companies' molecules. Laurus Labs is at 43% and wants half by FY30. Laurus grew revenue 7% a year over five years to Neuland's 17%, met 11 of 12 financial guides and slipped 26 dates. Neuland kept its dates and missed FY25. Both prices sit above our bull case: about 54% growth a year at Laurus, about 37% at Neuland.Read the comparison →Same industryMachine toolsJyoti CNC AutomationJYOTICNC versus Macpower CNC MachinesMACPOWERHead to headJyoti CNC Automation vs Macpower CNC Machines: Two bakeries on one street. The bigger one is building an oven for plainer breadJyoti CNC and Macpower CNC make the same machines in the same city. Jyoti grew revenue 31% a year against 18%, at a 25% margin against 16%, yet our board ranks it 98th to Macpower's 28th and its price implies about 45% earnings growth a year against 38%. Neither turns much of its profit into cash.Read the comparison →Same industryAI infrastructureE2E NetworksE2E versus Netweb TechnologiesNETWEBHead to headE2E Networks owns the hotel. Netweb Technologies builds the rooms and hands over the keysNetweb builds GPU systems and is paid once per box, at a 13% EBITDA margin. E2E owns GPUs and rents them by the hour, at 51-75%, and closed FY26 at a loss. Both prices already assume more growth than our bull case. Our own leaderboard ranks them 8th and 53rd, and about half of that gap is our model filling a blank.Read the comparison →Same industryOptical fibreSterlite TechnologiesSTLTECH versus HFCLHFCLHead to headSterlite Technologies and HFCL wired India's internet and earned nothing. Now both are bidding for its AISterlite Technologies and HFCL make the same thing, did about the same revenue last year, and are chasing the same new customer — AI data centres. Neither has grown earnings in five years. One is priced at 129 times earnings, the other at 57. Both have met 2 of their last 8 and 10 financial guides.Read the comparison →Same industryInstant coffeeCCL ProductsCCL versus Vintage CoffeeVINCOFEHead to headCCL Products has the kitchen. Vintage Coffee has the queue. Dinner costs the sameIndia has exactly two listed instant-coffee suppliers. CCL Products is eight times the size and runs its plants two-thirds full. Vintage Coffee is full, with orders it cannot cook until a ₹550 crore plant opens in FY28. Both trade at about 32 times earnings.Read the comparison →
Plate 03 — The Notebook
The Notebook
What I’m building, how I invest, and the method behind the scores.
- 01Sector mapsTwelve contract factories: earnings grew in both cohorts. The multiple moved in only oneTwelve Indian electronics makers, split by what they build. The five AC and consumer makers turn 4–10% margins into cash; the seven specialists earn 10–20% and are owed it. Both trade near 80–85 times earnings. Since January the specialists are 2.5x with the P/E doing most of it; the AC makers are flat. Then the chain, and which way each specialist is moving.EMSElectronics manufacturingSector mapSemiconductors
- 02ProductWe scored 121 earnings calls. Then the market voted.We scored all 121 Q4 FY26 earnings calls, then watched the stocks for two months. Strong calls ran ahead of the market and kept going; weak ones fell behind and stayed there — plus an honest accounting of how much of that edge is hindsight.121-company testScore vs price
- 03ProductA stock can rally and get cheaperTwo weeks ago I started keeping every valuation reading instead of overwriting it. The biggest movers since include a stock whose price didn't budge and one that rose 9% while getting cheaper — because valuation moves for two different reasons, and only one of them is the price.Valuation scoreImplied growth
- 04ProductIs the quarterly score telling us anything?I stress-tested our quarterly scores against real profits and real stock prices across 18 companies. What the score is, what it isn't, and the one pattern that paid every time.18-company backtestScore vs returns
- 05How I investWhat is my criteria?What I'm building is secondary. What I'm really trying to do is solve my investing decisions — and that comes down to one question: what is my criteria?My criteriaConviction
- 06How I investHow the analysis becomes a decisionThe machinery behind a single call on a stock: seven elements, a trust-and-understanding gate, four component scores, one composite, and finally — with direction and reward asymmetry — a decision.7 elementsComposite scoreDecision rule
- 07How I investSwing variables: tracking the story between quartersAn experiment: tracking each company's story the way you'd track an EPS beat — expected vs delivered, beat/match/miss — and finding the few variables that can swing the whole trajectory.Swing variablesBeat / match / miss
- 08ProductMaking it worth your timeAlmost every change to the portal right now comes back to one idea: a working investor can't read everything. The portal's job is to earn the time you give it.Attention budgetWhat to cut
- 09ProductWhat is Story of a Stock?The wedge, the portfolio problem I'm building it to solve, and the ladder from raw data to an actual buy/sell decision.The wedgeRaw data → decision