Every morning at about 11:05am Melbourne time, my trading bot wakes up and works through the same checklist. Each step has to pass before it moves on to the next.

Read more

• Market Cap: ~$10.4 Billion

• 7.5x Price-to-Operating Cash Flow (P/OCF)

• 18.8x Trailing P/E

The note will be available to read for free on website after 6 months to read now

Subscribe now

Read more

By 15 October 2026, the earnings threshold has been passed: FY25 profit was ₹17.78 crore, followed by ₹20.04 crore in FY26. Cash conversion improved in FY26, although the net margin narrowed.

The story has continued since those annual results. A 6 October disclosure describes a substantial RLNG purchase commitment, and yesterday’s filing announces a new refinery supply award. These developments make execution and funding more immediate questions than they were in the original small-company study.

Read more

Read more

Plan was simple retire by drawing 4% every year fixed 16L every year for rest of live

The 4% rule didn't just survive here it never even got stressed. Starting at ₹4 cr and pulling out ₹16 L a year, the man withdrew ₹2.88 cr over 18 years and still ended up with ₹43.66 cr. The corpus roughly 11×'d despite the withdrawals.

As reader you would say this not realistic as

• The withdrawal is flat ₹16 L, not inflation-adjusted. The textbook 4% rule (Bengen / Trinity) raises the withdrawal with inflation every year. To hold 2009 purchasing power, that ₹16 L should have grown to roughly ₹45–50 L by 2026.

• 2009 was a near-perfect entry — right off the GFC bottom. Sequence-of-returns risk is brutal when you retire into a crash; here he retired into a recovery. Someone who started the same plan on 1-Jan-2008 would look very different.

So fix the issues, we withdraw a 10% inflation escalation, Even pulling 2.5× more cash out, the corpus still grew ~8.7× (₹4 cr → ₹34.68 cr)

The real change is if we make it from near-perfect entry to near worst entry, what if Mr Lucky was Mr Unlucky and had retired in 2008 instead of 2009

The one-year timing difference swings today's corpus by 2.8× (flat) and a staggering 7.3× (inflation-adjusted) — ₹34.68 cr vs ₹4.76 cr.

What the 2008 retiree lived through

In year one his ₹4 cr halved to ₹1.99 cr as the NAV fell ₹196 → ₹100 — and he still pulled his withdrawal out at the bottom, locking in the loss by selling units cheap.But the inflation-adjusted version essentially never recovered: it spent 19 years grinding between ₹4–6 cr while the withdrawal climbed to ₹89 L.

With 10% inflation and 2008 start, Mr unlucky is looking to get a bloody shock in his retirement life at 74 in 2031/2, the risk of running out money

So what can Mr Unlucky do?

When winds are turbulent behave properly if things go south you adjust and play the game handed to you

One simple fix is below rule

The blue line never rises above either dashed line that is the rule. In good years (2008, 2011, 2015, blue dots) your indexed target is the lower number, so you spend exactly what you planned. In every other year the orange cap is lower, so your corpus sets your budget.

The big red dot in 2009 is the whole point: the crash drags the cap down to ₹11.9 L, so you spend that instead of the ₹17.6 L you “wanted.” That one voluntary cut is what stops the death spiral.

And notice you’re not getting poorer the orange cap keeps rising because the corpus keeps growing. Your spending still climbs from ₹16 L to ~₹59 L; it just grows slower than the fantasy 10%-indexed line, which runs away to ₹89 L. The shaded gap is exactly the spending you defer to stay solvent.

You can’t control markets but you can control how you plan and react, if you like data drive retirement goal setting done via our mutual fund distribution arm get in touch

Subscribe now

Is Mr Market giving us the 2022 kind of valuation after recent fall

Sep 2026

Back in November 2024 I wrote a note on Tesla to expand my own knowledge of the business. The last paragraph is the one that keeps coming back.

Two questions follow from that sentence.

Has the temporary bad time arrived?

And if it has, is Mr Market pricing the business at 30 times those depressed profits, or is he doing something else?

This is a follow up note.

# What I said then, and what actually happened

In Nov 2024 the stock was a roughly $800 billion company. Traditional metrics already looked stretched.

• 8 times sales

• 80 times profits

• FCF yield less than 2%

On the other side of the ledger I had four things.

• Toyota still sold about 10 times more cars

• Energy storage was just starting to scale

• Tesla had a track record of delivering the unimaginable

• It could be the first company to bring human like AI to factory and homes

I also laid out the flywheel.

Cars collect data. Data trains FSD. FSD feeds insurance and robotaxi. Energy sits next to the car. Manufacturing gets more vertical. Humanoid robots sit at the end of the same AI stack.

That flywheel is still the bull case. What changed is the cost of spinning it, and the price Mr Market is asking you to pay while profits go the other way

Updates to thesis

Capex

In 2024 management was guiding a $10 to $11 billion run rate for the next 2-3 years. I wrote that most of it would not be on automotive. That part was right. The number was not. Capex in Q2 2026 alone was $5.8 billion. Full year 2026 guidance is now more than $25 billion The extra is AI compute, Cybercab, Optimus lines, data centres. The flywheel is no longer a cheap option sitting on top of a cash generating car company. It is the capex budget.

Energy

I expected energy to be a $20 billion division in 2025. It did scale, and it is now a real second engine, but 2025 energy revenue came in closer to $13 billion. Deployments are lumpy. Margins, which looked fat, got clipped in 2026 by warranty charges, expired tariff benefits and falling industrial storage prices. The direction I had was right. The 2025 dollar figure was wrong.

Cars

Automotive revenue had already started tiring in 2024. It tired some more in 2025. Deliveries fell two years in a row. 2026 volume has come back, including a strong Q2, but it came back the way I said Tesla would defend share. Lower sticker. Cheaper trims. Financing. That is exactly the strategy list from the original note. More cars, less profit per car.

BYD and China

Still the right watch item. Anyone tracking Tesla still has to look at how they defend Europe, USA and China against BYD and the rest of the Chinese pack. It is no longer a future risk. It is why average selling prices keep leaking.

Has the temporary bad time arrived?

Yes. On profits.

Peak year was 2023. Tesla made about $15 billion. EPS was $4.30.

Then the slope.

• 2024 profit $7.1 billion

• 2025 profit $3.8 billion

Trailing twelve months to June 2026 still about $3.8 billion

Trailing EPS about $1.08

That is a 75% haircut from the peak year. Revenue did not collapse with it. Trailing sales are about $104 billion. The company is selling enough. It is keeping far less.

Q2 2026 is the cleanest recent snapshot of the new shape.

• Revenue $28.2 billion, up 26%

• Deliveries 480,126, up 25%

• Operating income $398 million, down 57%

• Operating margin 1.4%, down from 4.1% a year earlier

• Free cash flow minus $1.1 billion

They sold more cars and made less money. GAAP net income in that quarter still printed $1.1 billion. Do not get hypnotised by that line. A large mark to market gain on the SpaceX stake sat underneath it. The number that tells you how the operating business is travelling is the $398 million of operating income.

So the first half of my 2024 sentence is on the board. Temporary bad time, yes.

Why profits compressed

Five things stacked on top of each other. None of them is a mystery.

1. Price for volume

Tesla cut list prices, launched stripped down Model 3 and Model Y variants, and leaned on cheap finance. Average revenue per vehicle in Q2 2026 was about $42,730 against $45,345 in the prior quarter.

Auto gross margin excluding regulatory credits was 16.3% in Q2. In the fat years of 2022 and 2023 it was mid 20s to around 30%. Profit per car has gone from something like $6,000 plus to a bit over $2,000.

Model 3 and Model Y were 97% of Q2 deliveries. The high ASP cars are not carrying the P&L. The company is doing what I wrote they would do, make the entry car cheaper so more people can buy it. That works for share. It does not work for 2023 style margins.

2. The credit cushion is dying

Tesla used to sell emissions and fuel economy credits to other carmakers. That line is almost pure profit.

• 2024 credits $2.76 billion

• 2025 credits $1.99 billion

• Q2 2025 $439 million

• Q2 2026 $146 million, down 67%

In 2025 those credits were still more than half of reported net profit. They are fading because the rule that created demand for them is being unwound. The $7,500 US federal EV tax credit expired on 30 September 2025. Fuel economy penalties that forced rivals to buy Tesla credits have been eased. This is policy, not a Tesla execution miss. It still leaves a hole in the P&L.

3. US demand was hit, then they discounted to refill the factories

Two years of falling deliveries, 2024 and 2025, came before the 2026 rebound. The US lost the $7,500 credit. EV rules were loosened. Brand became political. Europe bounced in 2026. The US did not fully. Tesla then did the rational factory thing, cut price, fill the line. Volume returned. Margin did not.

4. AI, robotaxi and Optimus are now a P&L tax

This is the biggest change versus the 2024 note.

In Q2 2026 Tesla said, in so many words, that heavy R&D to accelerate AI, software and fleet based profits will negatively impact profitability during this phase

Operating expenses up 47% to $4.35 billion

• R&D up 49% to $2.37 billion

• Capex up 142% to $5.79 billion

• Full year capex guided above $25 billion

• Free cash flow minus $1.1 billion

Robotaxi, Cybercab, Optimus and AI compute are real cash leaving the building. They are not yet a real earnings line. FSD subscriptions are growing and services margins have improved. That is the first small royalty from the flywheel. It is nowhere near large enough to replace 2023 auto profit.

5. Energy grew, its fat margin did not

Energy storage had a strong deployment quarter in Q2, 13.5 GWh. Gross margin still fell from 39.5% in Q1 to 20.4% in Q2. A $240 million warranty charge on vendor cells, $200 million of Q1 tariff benefits that did not repeat, and industrial storage prices coming down as competition showed up.

Energy is the second engine. It is not plugging the auto hole.

Factory costs did get better. That is why Tesla is still profitable and still prints the highest profit per car among volume automakers. Cost cuts just did not keep up with ASP down, credits gone and opex up.

Now the second half of the sentence. Is it 30 times?

No.

Late 2022 / early January 2023, the window I pointed to.

• Price $108 to $123

• Market cap roughly $340 to $430 billion

• Trailing EPS $3.62

• Trailing P/E about 34 times at $123, about 30 times at the $108 low

• 2022 profit was still $12.6 billion

That was a price crash on still high earnings. Musk, Twitter, demand scare, price cuts, rates. Profits had not yet collapsed. The multiple did.

Today, September 2026 close.

• Price about $372

• Market cap about $1.47 trillion

• Trailing EPS about $1.08

• Trailing P/E about 350 times

• Forward P/E about 160 to 210 times on 2026 consensus EPS of roughly $1.78

Apply the same 30 times rule I wrote in 2024.

• 30 times current trailing EPS $1.08 is about $32

• 30 times 2026 consensus EPS $1.79 is about $54

• 30 times peak 2023 EPS $4.30 is about $129

Spot price $372 is about 11 times the 2022 trough and about 10 times richer on trailing earnings than that setup.

Even the recent 52 week low of $297 in July 2026 was still around 275 times trailing earnings. Not 30 times.

Why the multiple refuses to compress

In 2022 the market briefly valued Tesla as a car company with growth risk.

In 2026 the market is valuing Tesla as an AI option that happens to ship cars. Robotaxi, Optimus, FSD, AI chips, the same flywheel I drew in 2024. That is why earnings can fall 75% from peak and the stock can still sit near $370, only about 25% off a $499 high printed in December 2025.

What I can do is refuse to confuse the option with the 2022 setup.

If robotaxi and Optimus work at scale, today’s earnings do not matter. That is the bull case, and it is internally consistent.

If they take longer, or capex stays above $25 billion for years, or FSD never becomes a large paid network, then you own a 16% auto margin car and storage company at 14 times sales and 350 times trailing profits. That is also internally consistent. It is just a very expensive way to own a car company.

The original note already had this tension. Traditional metrics said overvalued. The flywheel said do not use traditional metrics. Twenty two months later the tension is wider, not narrower. Profits went down. The multiple went up.

What I am watching from here

Same list as 2024, with the new numbers attached.

1. Auto ASP and margin excluding credits. If 16% is the new floor and volume keeps growing, the auto engine can still fund the flywheel. If ASP keeps leaking, it cannot.

2. Regulatory credits. Treat them as a melting ice cube. Do not underwrite 2027 earnings on 2024 credit income.

3. Energy deployments and energy gross margin after the warranty noise. This is the only current business that can become a second profit engine without waiting for robotaxi regulation.

4. FSD paid subscribers and services margin. This is the first royalty from the data flywheel. It has to start showing up as dollars, not just miles.

5. Capex versus operating cash flow. $25 billion plus of capex against a mid single digit billion profit base is a funding question, not a vision question. Management is already lining up large debt capacity. Watch dilution and interest.

6. BYD and China in each of Tesla’s three big markets. Still the right competitive test.

Conclusion

The flywheel I wrote about is still the reason Tesla is an interesting business.

Cars → data → FSD → insurance and robotaxi

Energy next to the car

Manufacturing getting more vertical

Humanoid robots at the end of the same AI stack

What I underweighted in 2024 was how expensive it would become to keep turning that wheel, and how little of the 2023 profit pool would survive the price war, the credit fade and the AI opex.

Tesla has a track record of delivering the unimaginable. I wrote that then. I still believe the organisation can do hard things.

I may or may not have position in Tesla, This note was written to update my own understanding of the business. It is not a recommendation to buy or sell.

reason is this from page 4 of their FY26 presentation

A better price graph is this

Read more

Prepared for: personal research use. Date: September 2026 Price: INR 961.15

Business Snapshot

Shilpa began as an API manufacturer and has long been associated with oncology. The current company is trying to become a broader specialty pharma and integrated CDMO platform across small molecules, formulations, biologics, ADCs, peptides and recombinant albumin.

Where the business is today

Q1 FY27 was the fourth consecutive record quarter for revenue and EBITDA. Consolidated revenue was INR 469 cr, EBITDA INR 139 cr, and EBITDA margin 30%. Reported PAT was INR 101 cr, helped by a tax reversal; management said normalized tax should move toward about 25%.

Segment evidence

Q1 FY27 segment momentum was broad: API revenue was INR 260 cr including captive demand, formulation revenue was INR 198 cr with more than 100% YoY growth, and biologics revenue was INR 52 cr with 42% YoY growth. The official presentation’s consolidated revenue breakup shows API INR 216 cr, formulations INR 198 cr and biologics INR 52 cr.

Why Capability Could Be Real

Shilpa’s positive case rests on breadth plus integration. Many Indian pharma companies can claim one piece of the puzzle: API chemistry, formulation development, injectables, biologics, or manufacturing services. Shilpa is unusual because it is attempting several of these under one operating architecture.

Read more

Source of all data is RBI

1 Credit growth bottomed at 9.0% in May 2025 and has almost doubled to 17.7% in the twelve months since.Deposit growth moved too, but far less — from 9.9% to 12.2%. The result is a 5.4 percentage point wedge that banks have to fund from somewhere other than retail deposits:

The banks least hurt are those with the highest CASA ratio (current and savings balances as a share of deposits), because their funding cost barely moves. Banks plugging the gap with bulk money see margins compress first.

2

At 82.7% the system has limited room to keep growing loans faster than deposits without paying up for funding. Historically, a high CD ratio resolves one of two ways: banks compete harder for deposits which raises funding costs and squeezes net interest margins or credit growth slows to meet deposits. Both are margin events; only one is also a growth event.

3 The pools that grew fastest, personal loans and lending to non-bank finance companies, are the least seasoned.

Expect continued growth in personal loans

4

unsecured retail is the highest-yield and highest-risk slice of that growth, and the RBI has already intervened once by raising risk weights on it — regulatory action is a live variable, not a tail risk. Second, the mirror image of industry's decline is that corporate credit is now the underpenetrated pool. If a private capex cycle does arrive, the banks positioned in corporate lending have the most room to grow into it.

5

Falling policy rates cut both ways for a bank. They lift the value of the government bonds it already holds treasury gains, which flow to reported profit while lowering the yield on everything it lends from here.

6 savers are being paid very little to stay.

savers respond: a 2.5% savings rate against 12% broad money growth is a strong push into mutual funds, equities and insurance.

7 credit expansion in Figure 1 is being driven by commercial bank lending, not by central bank liquidity.

8

UPI is king and any new business has to ensure UPI is offered to stay relevant

Putting it all together

Credit demand is running ahead of deposit supply (Figures 1 and 2), and banks are managing the gap by paying savers very little (Figure 6) rather than by slowing lending. That is why margins look healthy today.

It also creates the two things worth tracking. Savers are leaving the deposit shortfall and the flows into mutual funds and insurance are the same phenomenon seen from two sides. And the loan book absorbing all this funding is now a household book, not a corporate one (Figure 4), which means the system’s asset quality is tied to jobs and wages rather than to the capex cycle, on loans too young to have been tested (Figure 3).

The near-term picture is good: rates are still falling (Figure 5), liquidity is ample (Figure 7), and growth is real. The questions are what deposits cost when the competition for them starts in earnest, and how the 2025–26 vintage of retail loans performs when it seasons.

Before you leave

Tremendous amount of work goes in producing content like this consider upgrading.

It frames the addressable market as growing from ~$3bn toward ~$21bn over six to seven years.

A paid subscriber requested this report via DM now available to all subscribers

Read more

Just marginally above S&P500 since start

A lot of churn happened folio snapshot is below only VAS and IVV were untouched or added

Note above is snapshot and things may change after writing this post, I have 20% in benchmark and 27% in muted Aus ETF so if I have to outdo benchmark then I have to find the alpha in rest of folio. I am giving thought to this and will share in my next year right up as you will see changes to align to such philosophy.

My rationale for current lot may change on a whim as I warned you I churn when facts change its a very small portfolio and we dont have behave like fund managers or billionaires it is our money and it should be positioned to do best

1. HEUR - General revival of Europe region after jolt from Trump, exporters tracked in this fund will benefit

2. DRGN - Chinese AI tech would be adopted everywhere except western democracies just like Chinese good as they are cheap

3. META - read this ,a punt of zuck

4. GRIN - most hated emerging market i.e. India will surprise many

5. COH - bought on the day when it dropped 40%, already laughing :)

6. UBER - Cheap and will benefit from autonomous vehicles on streets

7. AINF - Suppliers to AI ecosystem

8. Others - some space industry and cyber security ETFs

Where I killed it last year

1. Silver

2. GOLD

Where I got killed last year

1. BTC

NB: because they way retirement funds is designed here in Australia , you get continuous liquidity drip feeder that allows easy rebalancing

Not financial advise just my yearly notebook to keep myself accountable

Connect with me

Insta

X

Courses

BTC Wealth Builder

GOLD Investing System

Trading View Indicators

Internet Companies

Silver Indicator

Data Driven Mutual Funds

Tankrich MFD

The course covers below

Some students requested, so we added SMS delivery of signals for small fee to cover our Infrastructure cost

I added an update to signal last year they help in identifying if the market is losing momentum.

The code is simple and thats why it works, participate in trending market and stay away in side ways and its not for daily traders its for positional traders.

Any updates I make to course will be available lifetime on Gumroad at no cost to you.

General information only, not financial advice. Automated indicator output.

▎ Crypto is volatile — you can lose everything. Trade at your own risk.

82 companies:

- 102 Indian entries

- 11 non-Indian entries

All captured here - search sort and enjoy reading

Happy Independence Day

We are running an independence special, life time access to newsletter get it here

offer will close on Monday 17th August

Bonus* I will thrown in

• BTC course

• GOLD course

• Silver Trend following indicator

All of above are worth more than $1000 USD, see you inside

*Activated after 72 hours

In the 1930s, while the Golden Gate Bridge was being built, one safety decision changed the psychology of an entire construction site.

A safety net was installed under the bridge during deck construction. It saved 19 workers who fell and survived; those men later became known as the “Halfway to Hell Club.” The net did not remove the danger, but it changed how people behaved around danger. It gave workers the confidence to keep doing hard work in a high-risk environment.

Investing in small caps works in a similar way, except for one uncomfortable difference.

There is no net.

When small-cap investors fall, they feel every inch of it. The long-term numbers may look beautiful, but the path to those numbers is psychologically brutal.

On a spreadsheet, these numbers look almost unreal.

Nippon India Small Cap shows a 20.78% 10-year CAGR. Quant Small Cap shows 20%. Axis and HSBC are both above 18%. HDFC, Kotak and DSP are all in the mid-to-high teens over a decade.

This is why small caps remain so attractive. At 17–20% annualized returns, money does not merely grow. It transforms.

But this is also where investors misunderstand the product. They see the long-term return and assume the experience will feel like the long-term return.

It never does.

The invoice arrives in the short term

The same funds that look impressive over 10 years can look ordinary, frustrating or even wrong over 1–2 years.

HDFC Small Cap has delivered 17.25% over 10 years, but its 1-year return in this dataset is -1.97%. Nippon has a 20.78% 10-year number, but the latest 1-year number is only 5.26%. Kotak shows 16.64% over 10 years, but only 1.11% over 1 year.

This is the emotional toll.

When you look backward at a 10-year chart, the bad periods look like tiny bumps. But when you are living through them, they feel enormous. A flat year feels like failure. A negative year feels like betrayal. A two-year period of weak returns makes investors question the fund, the manager, the category and themselves.

Small-cap investing is not mainly a return problem. It is an endurance problem.

Size itself becomes a variable

The other invisible issue is capacity.

A small-cap fund is supposed to hunt in smaller companies. But when the fund itself becomes huge, the game changes. It may still perform well, but it cannot behave like a nimble boutique fund forever.

Nippon’s AUM is ₹74,604 crore. HDFC is at ₹38,809 crore. Quant is at ₹31,774 crore. Axis and Bandhan are both above ₹27,000 crore.

This does not mean these funds are bad. It means the question changes.

For a small fund, the question is: Can the manager find winners?

For a very large small-cap fund, the question is: Can the manager still find enough winners at scale without diluting the portfolio into a more mid-cap-like exposure?

That is why the most interesting research is not always in the obvious large names.

Four funds worth studying more closely

Read more

Capex is eating up FCF, This is not going to stop, Meta now expects to spend $130–145 billion on capital this year.

We have seen this film before Zuckerberg has take huge bets with operating cash flow

His last big bet Metaverse by all means of imagination is failed venture at least till now . His next one super intelligence can make or break company

It’s a big investment and it’s a big bet. We see the technology working… My personal bet is that the people who invest in this are going to be rewarded and feel very good over time.

— MARK ZUCKERBERG, Q2 2026 EARNINGS CALL

However the distinction this time is the change to balance sheet, yes META is borrowing and that too aggressively

To fund capex, Meta doing something it had never done as a public company. It is no longer cash fortress. As it now carries $83.7 billion of debt against $90.3 billion of cash, and it is reaching further afield for money this quarter read( data-centre venture with BlackRock)

One thing remain constant, its one best business on earth, Total revenue has jumped from $48B to $60B in a year

The operating margin however is now getting impacted with AI build out

The market is not euphoric about this spending; it is sceptical. It is pricing in a real chance that the hundreds of billions going into AI never earn their keep

From technical on above monthly chart its broken its 10 month EMA and last time when that happened in 2021 price went from 380 to 100

Will the bet pay off?

There were number of call outs made in their Q2 transcript , what I have done is largely organised them in short, medium and long term prospect

Near-term payoff: AI is already lifting the core business

“On a dollar basis, our ads business is reporting faster year-over-year revenue growth than any other company’s reported ad business -- so these AI investments are paying off.”

“9 million small businesses on our platforms are now using at least one of our AI ad creative tools, and we’re rolling out new end-to-end creative solutions that help advertisers translate performance data into their creative decisions. Muse Image is going to supercharge this.”

“Since we rebuilt Meta AI and integrated Muse Spark, we’ve seen a 60% increase in the number of people interacting with the assistant each day and that continues to grow quickly week-over-week.”

“On Instagram, global time spent this quarter grew double digits year-over-year, largely driven by improvements to our Feed and Reels recommendations. On Facebook video time spent increased 9% globally year over year and over 10% within the US & Canada.”

“In Q2, we also advanced our user understanding models to analyze ads and organic activity and simultaneously improve both user experience and advertiser performance. Combined with our GEM model for ads ranking and sequence learning, these advancements generated an 8.3% increase in ad clicks and a 15.7% uplift in conversions on Facebook.”

“Our AI-powered Advantage+ end-to-end solutions continued to grow, reaching over $75 billion in annual revenue run-rate this quarter.”

Medium-term payoff: New products that leverage the same infrastructure and distribution

“Soon, we will have agents that can work 24/7 on your behalf to help achieve your goals and improve your life, your health, your relationships, your finances -- whatever you want. … this needs to be a great consumer product that just works out of the box and is easy enough for billions of people to adopt and use.”

“We made Meta Business Agents available globally this quarter on WhatsApp and Messenger, and there are already more than 1 million businesses using them to talk to their customers or complete sales every week. We’re rolling business agents out on Instagram now too. … Over time, we’d like to build this into a business-in-a-box service that can help you start and run a whole business using Meta’s platforms.”

“In terms of how we will monetize these, we have a mix of subscriptions, volume-based pricing, and I expect that we’re going to evolve more of these products to be like our ad systems where businesses only pay us when we achieve results for them. Over time, that will let us run an efficient auction over our compute, similar to how we do that for our advertisers today.”

“We’ve built our API, we’re rolling out business agents, we’re getting a lot of offers for compute at a significant premium over what we paid for it, and we have more coding and productivity tools on our roadmap as well.”

“We just released our own line of Meta Glasses in collaboration with EssilorLuxottica, including a style that we designed with Kylie Jenner. They’re the first glasses to ship with Muse Spark out of the box so that they can understand what you’re seeing and give even more helpful answers. Early sales have been strong, exceeding our expectations.”

Structural / long-term payoff: Full-stack + distribution moat?

“I think that we’ve definitely shown at Meta that when we have a product in a format that works, I would go as far as to say that I think we’re probably the best company in the world at scaling those experiences to billions of people.”

“We are really a full stack technology company. We built our own data centers, our own infrastructure, our own chips, our own low-level software. … It just seems to me pretty clear that having kind of sovereignty over building your own models is going to be an important part of that stack going forward.”

“I believe that building the kind of full stack model is going to be a lot of the durable advantage over time in how we build personal superintelligence agents, business agents, all these different use cases for all of the different customers that we want to serve.”

How the economics are supposed to work?

“Overall, we expect that a significant portion of our compute is going to go towards training our models, growing our core business, and delivering personal agents and new products. But we also expect to grow a large business serving large customers as well.”

“We believe that there will continue to be a significantly higher margin on selling intelligence rather than selling compute directly. But we think that there’s a big opportunity obviously to sell compute as well.”

“Our current plans are geared towards maximizing 2026 and 2027 capacity. … Longer-term capacity strategy aims to give us the flexibility to continue growing compute in 2028 and beyond by laying down data center and network foundations to accommodate future server decisions.”

“Infrastructure planning remains highly dynamic and even this year, there are a range of outcomes embedded in our outlook.”

“We are today, and expect to be in the sort of foreseeable future, supply constrained, and that really includes our core business, too, where there are -- we still have numerous ROI positive places that we would put compute toward if we had it.”

“AI is improving our core business -- it’s making our apps more relevant, and delivering better results for businesses. We’re starting to deliver more novel products, and we’ll have a lot more there soon as well. We’re investing aggressively because the potential is huge and we know that there are many ways to deliver value here.”

I can do DCF math and tell you fair value of share is $700 or $500 by you using aggressive or conservative assumption but main question as investor you need answer it will Zuck get it right this time?

For me I am interested and tracking

Written by hand typing,

Vivek

The chart of ITC is horror story with lot of crosses

How about fundamentals?

Read more

Twenty or so women sit on a mat. A field officer in a company shirt arrives on a motorcycle. Each woman hands over ₹300, or ₹500, or ₹1,200. The officer marks a register, the group signs off, and everyone goes back to work. Nobody has pledged gold. Nobody has mortgaged land. The only collateral in the room is that these women have to live next to each other for the rest of their lives.

That is the entire technology of microfinance. Not algorithms, reputation, and the Tuesday morning meeting.

Sometime in 2024, in large parts of India, that meeting stopped happening.

The consequences ran to roughly ₹55,000 crore of bad loans by March 2025 — about 14.8% of the industry’s gross loan book, per MFIN’s own data. Two years of contraction followed. Yet outside the sector, in an equity market busy arguing about defence stocks and PSU rallies, it barely registered.

Now the numbers have turned. So the question worth sitting with: are the green shoots real, or is this the same optimism that got everybody into trouble the last time?

Let me try to answer that properly, because the answer is more interesting than a yes or a no.

Part 1: How the fire started

In March 2022, the RBI did something quietly liberating. It scrapped the interest rate cap on NBFC-MFIs and replaced the old rules with a household-income and indebtedness framework. Lend to whoever you can assess, price for the risk you’re taking, just don’t let a household’s loan repayments cross half its income.

It was good regulation. It was also a starting gun.

Capital poured in. Fintechs and ordinary NBFCs, who weren’t classified as MFIs at all and therefore weren’t counted the same way in the industry’s own overlap data, started writing small unsecured loans to the same households. Ticket sizes crept up. The industry grew its book 29% in FY24.

And underneath the growth, three things rotted quietly.

One: the borrower collected lenders. The share of microfinance borrowers dealing with five or more lenders rose to 5.36% by December 2024, from 3.6% in September 2021. That understates it, because bureau reporting lagged and non-MFI unsecured loans often didn’t show up in an MFI’s overlap check at all. A woman could be repaying four institutions and none of the four could see the other three cleanly.

Two: the group stopped guaranteeing anything. As ticket sizes rose and field officers churned, centre meetings thinned out. Officers began collecting door to door because it was faster. The joint liability group the actual credit enhancement in the model became a formality on paper. What was left was unsecured lending at 20–24% with no security and no social pressure.

Three: nobody had ever tested it. The book had grown through a benign patch. There had been no serious income shock since COVID.

Then 2024 delivered one.

Part 2: The break — four shocks in twelve months

Shock one, mid-2024: income. A brutal heatwave, floods in parts of the east, and a general election that pulled field staff and borrowers away from the weekly rhythm. Collections dipped. In a book where the borrower is servicing four lenders out of one income stream, a small dip is not small.

Shock two, August 2024: the guardrails. MFIN, the sector’s self-regulatory body, tightened lending norms — capping lenders per borrower and household indebtedness. In January 2025 came Guardrails 2.0, effective 1 April 2025: a maximum of three lenders per borrower, a ₹2 lakh household indebtedness ceiling now inclusive of unsecured retail loans, and no fresh disbursement to anyone more than 60 days overdue on any loan above ₹3,000.

This was the right medicine and it hurt exactly the way right medicine does. A borrower who had been rolling loan four to repay loan one was now cut off. Rejection rates jumped. Loan originations in Q3FY25 fell 41.7% year-on-year by volume. The music stopped.

Shock three, February 2025: Karnataka. Following a spate of borrower suicides and complaints of coercive recovery, Karnataka promulgated an ordinance carrying jail terms of up to ten years and ₹5 lakh fines. RBI-regulated lenders were formally exempt. It made almost no difference in the field. Word spread that microfinance recovery was now a criminal matter, and collection efficiency for regulated lenders in the state fell below 90%. Karnataka held around ₹60,000 crore of the industry’s book. Tamil Nadu another ₹50,583 crore, roughly 13% of the national portfolio passed its own version, notified in June 2025.

Shock four: the arithmetic. Everything above showed up in the P&L in FY25 and FY26.

CareEdge’s reconstruction of what happened to the loan book that existed in March 2024 is the number that should stay with you: 14.1% of it was written off during FY25, another 3.6% sold to ARCs, 5.4% sitting in Stage 3, 3.4% in Stage 2. By the end of FY26, they estimate total stress touched roughly 30% of the March-2024 book.

Nearly one rupee in three of the pre-crisis loan book went bad.

And the human number, the count of borrowers with a loan outstanding from an NBFC-MFI stood at 3.4 crore in March 2026 down 15.8% in a single year. CreditAccess alone wrote off around 19 lakh borrowers between FY21 and FY26. Those women simply fell out of formal credit.

Part 3: One slide, one whole cycle

The CreditAccess Grameen slide is worth reading slowly, because it is the cleanest picture of a credit cycle you will find in an Indian investor deck. They even label the three phases themselves: Prioritising Collections → Portfolio Maintenance → Portfolio Growth.

Look at the third row and the fourth row and then look at the second row again.

Credit cost went up nearly 5x. PAT went from +₹397 crore to a loss. But pre-provision operating profit — what the franchise actually earns before absorbing losses — fell about 9% and then made a new high.

That is the single most useful lesson in this whole episode. In a lending crisis, the P&L breaks but the franchise often doesn’t. The branches kept operating, the customers kept coming, the spread kept getting earned. What changed was how much of that spread had to be handed over to cover losses.

Which is why the correct question during the panic was never “what are earnings this quarter.” It was: is annual PPOP large enough to absorb the losses without eating into equity, and is the funding line still open? For CreditAccess the answer was yes — they never posted an annual loss (FY25 PAT: ₹531 crore), capital adequacy never fell below 24%, and gearing stayed around 3x. No rights issue, no dilution, no covenant waiver.

For others, the answer was no. Spandana Sphoorty lost ₹1,035 crore in FY25 and another ₹699 crore in FY26, wrote off ₹1,155 crore, saw revenue fall 56%, needed a ₹400 crore rights issue, and breached loan covenants (later waived). Fusion Finance lost ₹1,224 crore in FY25.

Part 4: The green shoots, measured

Now the good news, and it is better documented than most recoveries.

The industry stopped shrinking. MFIN’s Micrometer for the March 2026 quarter recorded the first sequential portfolio growth in seven quarters — up over 3% to ₹3,25,174 crore. Quarterly disbursements of ₹77,524 crore were the highest in seven quarters, though still below the FY24 peak.

Asset quality is back to pre-crisis levels. PAR 31–180 fell to 2.0% in March 2026, from just over 6% a year earlier.PAR 31–90 is 0.8%.

The new book is clean and it is now most of the book. This is the statistic I would put above all others.

CRISIL estimates that loans originated after the August 2024 guardrails now account for around 80% of MFI AUM, and PAR 90+ on that book is below 1%. The problem has largely been written off and replaced.

Lending discipline is visible in the mix. Around 95% of exposure now sits with borrowers linked to three or fewer lenders. Roughly 66% of AUM is to repeat borrowers in their second cycle or later, up from 53% two years ago.

Funding came back. NBFC-MFIs raised ₹77,870 crore of debt in FY26, up 30.9% year-on-year.

Policy turned supportive. In June 2025 the RBI cut the qualifying-asset requirement for NBFC-MFIs from 75% to 60%, letting them diversify a much larger share of the balance sheet into secured and individual lending. In March 2026 the government launched CGSMFI-2.0, a ₹20,000 crore credit guarantee facility covering 70–80% of default losses on bank lending to MFIs, since extended to 31 August 2026.

And the company-level evidence is now stacking up. CreditAccess did ₹340 crore PAT in Q4FY26 against ₹47 crore a year earlier, on 4.4% ROA and a quarterly credit cost of 1.21%. The June 2026 quarter update was stronger still: loan book ₹30,319 crore (+16.4%), record first-quarter disbursements of ₹6,107 crore, PAR 90+ down to 1.5%, collection efficiency 99.68%.

Northern Arc’s rural book tells the same story from a smaller base full-year credit cost down from 6.7% to 4.9%, the March quarter at 1.3%, X-bucket collection efficiency at 99.6%, and Karnataka recovering from 94.5% to about 99.5%. Note what they did with the book itself: cut it from ₹1,500 crore to about ₹900 crore during the storm, then rebuilt to ₹1,009 crore with 84% of it covered under CGFMU guarantee.

Even the wounded are stabilising. Fusion Finance turned a ₹1,225 crore FY25 loss into a ₹14 crore FY26 profit. Spandana posted ₹5.27 crore in Q4FY26 its first profit in six quarters.

CRISIL now expects sector AUM to grow around 20% in FY27, against 4% in FY26. ICRA sees credit costs falling to 3–3.5%.

I think the shoots are real

Part 5: Three things that stop me short of celebrating

First, the scar tissue is still on the books. PAR 180-plus including write-offs was still 16.3% in March 2026. Borrower count is still falling down 15.8% year on year.

So where is growth coming from? Larger loans to fewer, better-vetted customers. Average ticket size rose 18.3% to about ₹61,500.

Sit with that for a second. Rising ticket sizes to repeat borrowers is exactly the right strategy today and it is also precisely how the last cycle began. Every credit boom in history has started as a quality-led recovery. What matters is whether it survives the third good year, when a competitor starts writing ₹90,000 loans and your branch manager’s targets aren’t being met.

Second, the macro that produced this recovery has already turned. FY26 was a gift: inflation near 2%, the repo rate cut to 5.25%, two consecutive good monsoons, borrowing costs falling for every lender in this piece. FY27 opens differently. The IMD’s seasonal forecast is around 92% of the long-period average “below normal” with El Niño returning.

July was forecast under 94% of LPA after one of the driest Junes on record, and kharif sowing was running roughly 23% behind schedule in late June. The Agriculture Ministry has flagged 315 districts at risk, 111 of them with less than a quarter of farmland irrigated.

In June 2026 the RBI held rates at 5.25%, cut its FY27 growth forecast to 6.6% and raised its inflation forecast to 5.1%, citing energy prices and the West Asia conflict. MFIN itself, in the same release announcing the recovery, warned members to plan for a weaker monsoon and the conflict’s effect on rural livelihoods.

Microfinance borrowers are rural, cash-flow-driven, and weather-linked. This is the first monsoon that will actually test the new underwriting. Everything before this was tested against good rain.

Third the banks have not come back. Banks’ share of the microfinance portfolio has fallen to 26.4% from 32.6% a year ago.

CGSMFI-2.0 was designed precisely to fix this, with ₹20,000 crore of guarantee capacity. As of the June 2026 extension announcement, loans of just ₹770 crore had been sanctioned under it. The plumbing exists; the water isn’t flowing.

Misra’s own comment was pointed: the sector has done its part, and it is now time for banks to come forward.

A lending recovery is only as real as the liability side. We need to watch bank funding, not just collection efficiency.

Part 6: Five names, four different businesses

The whole sector is quietly converting itself from unsecured group lending into small-ticket household finance, partly secured. If you’re underwriting these companies on their 2021 identity, you’re valuing the wrong business.

Part 7: What you’re paying for the recovery

The uncomfortable part.

CreditAccess bottomed at ₹750 on 27 January 2025 the same week it reported a ₹99.5 crore quarterly loss. A global brokerage had by then cut its target to ₹564 and called the de-rating structural. Maximum reported pain and maximum analyst pessimism arrived in the same fortnight, as they almost always do.

It is around ₹1,490 today. Roughly a double in eighteen months, against a Dec-2023 high of ₹1,794.

At that price you’re paying roughly 3x March-2026 book and about 30x trailing earnings but those trailing earnings are cycle-depressed, still absorbing FY26 credit costs of 6.8%.

If management delivers its FY27 guidance of 20–25% AUM growth at a 4–4.8% ROA, that same price works out to somewhere in the mid-teens on forward earnings.

The entire bull case sits in the gap between those two numbers. And it requires guidance to be met in a year when the rain may not arrive, funding costs have stopped falling, and the RBI has just trimmed its own growth forecast.

That is not an argument against owning it. It is an argument for knowing precisely which bet you’re making: the operating recovery is evidenced, but a meaningful share of it is already in the price. The easy money the de-rating reversal from distressed to normal has been made. What’s left is execution money.

The lesson that outlives the cycle

Go back to CreditAccess’s ten-year table. Across FY17 to FY26 they disbursed ₹1,37,763 crore, collected ₹1,03,235 crore back, and wrote off ₹4,938 crore net of recoveries.

Roughly 3.6 paise lost for every rupee lent, over a decade that contained demonetisation, a pandemic, and the worst credit cycle the sector has ever seen.

The business works. It just doesn’t work smoothly. Three crises in ten years, four if you count Andhra Pradesh in 2010 roughly one every four years. Anyone underwriting microfinance as a serene 25% compounder is underwriting an asset that has never existed.

The right question was never “will there be another crisis.”

It’s “will this lender still be standing, funded, and taking share when it comes?” which is a balance sheet question, asked years before you need the answer.

Which brings me back to the mat in the village, and the twenty women, and the register.

At CreditAccess, 99.68% of what was due in June 2026 came in on time. It always does, right up until the year it doesn’t. Both of those facts are true at once, and holding both of them in your head simultaneously is the whole job.

Watch the rain in August. Watch whether banks start lending. And watch ticket sizes because that is where the next one starts.

Written July 2026. All figures from company disclosures (CreditAccess Grameen and Northern Arc Q4FY26 and Q1FY27 filings and earnings calls), MFIN Micrometer, CRIF High Mark, and published reports from CRISIL, ICRA and CareEdge Ratings.

This is market commentary for education and discussion, not investment advice, and not a recommendation to buy or sell any security. I am not a registered research analyst. Please do your own work or speak to a SEBI-registered adviser before acting on anything here.

as someone tracking Affle this should be tracked at least annually along with trends

Affle 3i is India's largest listed ad-tech company and a rare global technology platform built from India.

Read more

Read more

BTC

Read more