India wants to be a global AI hub. But AI runs on physical infrastructure.
Servers. Power. Cooling. Land. Fiber.
I mapped the entire AI value chain. Here are the listed companies building India's AI backbone.
Thread 🧵👇
Anchoring bias cost me a good trade last year.
A stock I was tracking had been at ₹1,200 for months. It corrected to ₹850. My brain screamed "CHEAP! 30% discount!"
But cheap compared to what? To an arbitrary past price? Or to what the business is actually worth?
I ran it through my quality filters. Fair value was closer to ₹700-750.
₹850 wasn't cheap. It was still expensive. But my brain was anchored to ₹1,200 and everything below it "felt" like a deal.
The antidote in my system: I never compare a stock's current price to its past price. I compare it to what the business fundamentals justify. https://t.co/i1J3N9hrYS, not memory.
(Still catch myself doing it. The rule exists because I need it.)
#Psychology
Earnings Quality Scan — Apollo Hospitals (Q4 FY26)
₹529 Cr profit. +33% YoY. Revenue at ₹6,606 Cr (+18%).
Clean headline. But Apollo is actually three businesses wearing one stock ticker.
Healthcare Services (hospitals): The core. EBITDA ₹1,011 Cr, up 31% YoY. Occupancy strong, ARPOB rising. This is the cash-generating engine. No complaints here.
Apollo HealthCo (pharmacy + digital): Revenue ₹10,808 Cr for FY26, first full year of profitability. PAT ₹324 Cr. A year ago this was a loss-making drag. Now it's contributing. But the margin is thin. Very thin. ₹324 Cr profit on ₹10,808 Cr revenue is a 3% PAT margin. One bad quarter and this flips back.
Apollo 24/7 (digital health): Still burning ₹96 Cr/quarter. Down from ₹160 Cr. Improving, but not there yet.
The EQ question: what portion of consolidated profit comes from the hospital business vs the HealthCo turning profitable vs accounting treatment of the upcoming demerger?
(The demerger is the real story here. Once HealthCo gets listed separately, the hospital stock will be a completely different valuation animal.)
Verdict: REAL — hospital earnings are genuine and cash-backed. HealthCo profitability needs a few more quarters before I trust it. The 3% margin doesn't have much room for error.
Disc: Not a buy/sell recommendation. DYOR.
#EQS #APOLLOHOSP
Earnings Quality Scan — LIC (Q4 FY26)
₹23,420 Cr profit. Up 23% YoY.
Here's the uncomfortable part about that number: for a life insurer, reported PAT is one of the least useful figures on the page.
An insurer's profit is shaped by how much surplus gets transferred from the policyholders' fund to the shareholders' account, plus investment gains booked in the period. It's an accounting outcome, not a clean read on whether the business sold more profitable policies.
LIC also revised a key accounting policy for its standalone statements this year. Any time the policy changes in the same year the profit jumps, you slow down before celebrating.
So what actually measures earnings quality for an insurer? Value of New Business and VNB margin. And this is where it gets genuinely good:
VNB ₹14,179 Cr, up 41.6%. VNB margin 21.2%, up 360 bps from 17.6%.
That margin jump is the real story. LIC has been shifting from low-margin par products toward non-par: non-par APE share went from 27.7% to 35.1%. Higher-margin mix, written into new business. That's earnings quality improving at the source, years before it shows up cleanly in PAT.
(They also announced a 1:1 bonus. Nice, but it's a sweetener, not a fundamental.)
Don't read LIC like a normal company. The PAT is noise. The VNB margin is the signal, and the signal is good.
Disc: Not a buy/sell recommendation. DYOR.
#EQS #LIC
@WeekendInvestng The analyst is the last person I'd boot. By the time a position this size exists, it was a top-down allocation call made years ago, and no junior analyst was ever going to override that. Real question is who signed off on the audit for five straight years
@warikoo Beautifully put, and true. The only thing I'd add for anyone reading this earlier in the journey: The freedom to ignore the number comes after you've made enough to ignore it. At 33, checking obsessively was probably the right instinct.
@BaluGorade The bank doesn't work for free and neither does the seller. You're not getting a loan for nothing. You're getting a discount you'll never see, converted into a schedule.
We already have a coaching culture that optimises for rote learning over understanding and fundamentals. Drop AI into that and you don't get smarter students, you get faster shortcut-takers.
During my UPSC preparation years the people who cleared weren't the ones with the best notes, they were the ones who'd struggled through making their own. The struggle was the learning.
@REDBOXINDIA The under-reported win here isn't the strait, it's the diversified sourcing, strategic reserves, refining capacity. India leaned hard on discounted non-Gulf barrels over the last few years, which is exactly why a Hormuz crisis hurts less than it would have in 2022.
@microcp2mltibgr I disagree on your assessment of the upper end. Price holding for years assumes demand just keeps increasing. Every prior super spike killed its own demand and pulled in supply. 2008 hit 147 then collapsed to the 30s. The spike I believe. The plateau, less so.
@InvestRepeat You answered your own question without noticing. No mutual fund bought it. No brokerage covered it. That silence wasn't an accident, it was the verdict. Every active manager who looked at the numbers walked away. LIC was the lone domestic institution holding it.
@IndianGems_ Calling it corruption at the highest level almost makes it sound exotic. It wasn't hidden. Revenue in lakhs of crore, profit in a few hundred crore. I'm surprised how no one at LIC saw it. Must be corruption and connivance of LIC officials.
The lesson from Deep Dive #1: the strongest consumer moats aren't built by winning customers. They're built by owning whoever the customer trusts to choose.
Pidilite owned the carpenter. That's the whole game.
Next deep dive, I'll take a very different moat type off the index.
Which monopoly should I dissect next? Drop the name in the comments 👇
🔖 Bookmark the Monopoly Mavens index if you missed it.
Disc: Moat analysis, not a buy/sell recommendation. DYOR.
#TheSystemsTrader #StockAnalysis #PIDILITE
There's a company in India so dominant its brand name replaced the word for the product.
Nobody says "give me adhesive." They say "give me Fevicol."
That's Pidilite. And that one fact is worth more than most balance sheets.
Monopoly Mavens, Deep Dive #1 🧵
Moat verdict: Wide and durable.
Pidilite is close to a textbook M3 monopoly. Brand became category, the decider is locked in, distribution is everywhere, and the numbers prove the pricing power is real.
To be clear, this is a verdict on the MOAT, not a buy call. The moat is excellent. Whether ~60x earnings is a fair price for it is a separate question, and that one depends on your entry discipline, not Pidilite's quality.