Pinned Tweet
🧰We manage 200 cr of client money (their money not ours) to help them with their financial needs.
While doing so - it is uber critical for us to be careful (& not reckless). 🧑✈️
Graph below shows an X-Ray of our entire Mutual Fund Portfolio across different scheme categories
A Thread on our investment philosophy
We had sent this message on 29 Jan'26 to our clients when precious metals were firing on all cylinders. These moments can be v challenging for trying to steer client portfolios. A baggage of 20 yrs of investing experience tends to give a bit of butterflies in the tummy.
2026 is packed amazing learnings. How asset classes went up to the stratosphere and then going towards the patal lok ! Market darlings are now being perceived as a hot coal.
Gold / Silver / Semi Conductors - Kospi / Taiwan.. list goes on
Is the opportunity building ?
1
205
CA Puneet Agarwal retweeted
At BanyanFA we too deal with similar offering. But we aren't going to publicize it at this point of time. This is perhaps the worse time to attract capital seeing overseas valuations. And when we compare with India after 2 yrs of valuation correction- this is unfair comparison.
6 months since we created our model portfolio of UCITS ETFs. Up 14% vs Nifty return of -6% over the same period. Will it always beat the Nifty? Of course not! Is it a good diversifier? Hell yes!
thefynprint.com/global-portf…
Yesterday's market decline in the US was perhaps the first major one in the memory of people post COVID. Many joined the bandwagon post COVID & for them this was the first ever.
If you are in the right place - stick on to it. And for future - take a learning - respect valuations
All the proponents of global investing will all of a sudden have deep heart burns. If not them, their clients / followers. It a great learning to see how mean reversion happens & how valuations need to be respected.
Big Tech falling like a pack of cards.
1
155
All the proponents of global investing will all of a sudden have deep heart burns. If not them, their clients / followers. It a great learning to see how mean reversion happens & how valuations need to be respected.
Big Tech falling like a pack of cards.
We learn with time. Almost all our clients who have been with us for long duration - they aren't bothered. They have seen the cycles with us. They have learnt with time - lower points in cycles are opportunistic times rather than threats. It can't be learnt by seeing a reel
4
336
My issue with AI Infra companies is not whether AI has a use case & it's viability. The issue I have is similar to what India went through when Reliance entered into Telecom. Over a decade of consolidation and massacre of lower order.
Survival of the fittest. Not all !
AI COSTS ARE COLLAPSING:
The cost of achieving a given level of AI performance has dropped -47% per quarter since 2023, equivalent to an annual reduction of ~92%.
This drop is 6 times faster than the decline in computing costs from 1940 to 2001, and 18 times faster than the decline in lithium battery costs from 1991 to 2024.
Meanwhile, US and Chinese hyperscalers and neocloud companies are projected to spend nearly $1 trillion on AI infrastructure this year, almost twice last year's amount, with investment expected to rise further in the coming years.
Falling costs are making AI more accessible, accelerating its democratization, with wider adoption and intensifying competition pushing prices lower and compressing profit margins.
The key question is whether productivity gains, new revenue streams and wider AI adoption will generate sufficient returns to justify record spending.
AI is becoming dramatically cheaper, putting pressure on pricing power and future returns.
151
Does this sound similar to India Telecom rush when Reliance entered. 1.5 decades of
Consumers enjoying more choices and lower prices
Investors seeing no returns.
Let's see how this transpires for US hyperscalers.
Falling token prices + sticky-high GPU prices is classic Jevons paradox playing out in real time.
Inference is getting dramatically cheaper (efficiency, competition, Blackwell-scale gains). That doesn’t reduce compute demand - it explodes it. Cheaper tokens unlock agentic workflows, more coding, more automation and more users. Total token volume is growing far faster than unit costs are falling, so the scarce resource (GPUs, HBM, power, data centres) stays tight.
Implications for AI infrastructure stocks:
- GPU / accelerator makers (NVDA and peers) still benefit. High utilisation + sustained scarcity supports pricing and volume.
- Memory (HBM/GDDR), networking and power infrastructure see durable demand as inference scales.
- Hyperscalers and neoclouds face a margin squeeze if token price declines outrun volume growth, but capacity constraints and rising absolute spend so far argue demand is still winning.
- The shift to inference-heavy workloads favours specialised silicon and high-throughput systems over pure training clusters.
Bottom line: AI is proliferating faster than efficiency is reducing the need for hardware. Unit prices collapse, total spend and infrastructure demand keep rising. That’s bullish for the physical layer as long as volume growth remains the dominant force. alphatarget.com
202
Make a capex sound as OPEX. Helps in bolstering your ROCE. If this isn't a creative accounting, what else would be ?
If I am the signing Audit Partner for such firms, I would be getting really worried. Costs understated, Revenues overstated.
This is the dodgiest and most-desperate thing I’ve seen in the bubble so far. Amazon selling their chips to a group of investors and leasing them back - an off balance sheet arrangement done in plain sight. Really scraping the bottom of the barrel
ft.com/content/97d8d346-519e…
3
267
One thing which I agree materially - the transparency is lacking in Insurance sector. A lot gets pushed under the carpet. Bigger you are, better you are remunerated. Special products / payouts / terms are crafted for you.
Good to see regulations catching up.
Many investors hoped that PB Fintech would get a Managing General Agent (MGA) license - basically a subcontractor - who can underwrite & sell policies on behalf of insurers.
The IRDAI Chairman explains why this is unlikely.
Think this is partly why stock fell another -8% yday.
#Nifty #BankNifty #IRDAI #Insurance
1
228
The world where we live in today - this kind of circular trade / financing is being so nicely accepted. In the world of credit risk - the term was Wrong Way Risk. The lessons of The Credit Crisis of 2007-08 are now being evolved. Let's see how this transpires.
Broadcom has agreed to lend Anthropic up to $42B, and Anthropic would spend that money renting AI chips that Broadcom helps build with Google.
The loan could cover about 1/3 of Anthropic's $125.2B, 5-year chip lease, and the debt could later convert into Anthropic shares.
Broadcom can bring in a financing partner, and the debt could convert into Anthropic shares, though Anthropic expects no note sales before its IPO.
Much of that money would flow back to Broadcom, which expects Anthropic to become its largest compute customer in 2027.
Like Nvidia, Broadcom is using its balance sheet to sell chips, having already backstopped senior debt in a $35B Apollo-Blackstone vehicle that leases chips to Anthropic.
Anthropic ofcourse warned that Broadcom's dual role as supplier and lender creates "potential conflicts of interest" that could affect its access to compute.
1
259
A salaried person gets paid a salary - Fine
Book writer a royalty - fine
Speaker - honararium, flight tickets and nice hotel - fine
Commissions - these are remuneration for agency services and support. Why to demonise this ?
Replying to @banyanfa @monikahalan
Than what's your view on third party motor commission, you might be working in that motor insurance too, and criticizing this is ok but accept you get hefty payments in motor, health and life , accept that too
A good start but worst way to conclude. Does the publisher / editor / research team or the re-poster have on ground experience of claims. Reality on ground is diff vs nicely designed graphs.
Who will help with reality and assistance ?
1
3
315
And how will Anthropic pay Nvidia?
Anthropic has now contracted 2.6GW of $NVDA AI infrastructure to date, with shipments through 2028
Across CSPs/neoclouds, its reported contracted value now exceeds $180B
It also has ~15-year data-center leases starting FY28-29 that are expected to be reassigned to 3rd parties
1
250
Reading that AI majors are launching cheaper models. Their current pricing is resulting in deep losses. With lower prices, longer will be the runway to profits - courtesy Chinese models. How does that transition to Hyperscalers and the downstream ecosystem. Valuations !
2
291
Insurers cry foul - citing high healthcare inflation
Hospitals call it justified - citing quality
Regulator - lets apply band-aid - cut commission.
Customer - poorer post regulatory decision. Now lesser agents to service and lesser push / sales.
An IV infusion set procured for Rs 11.05 carries an MRP of Rs 325. A syringe procured for Rs 6.75 has an MRP of Rs 57.2. FDA Commissioner @Tukaram_IndIAS’s recent post on X has once again spotlighted hospital-insurer disagreement.
And even if the bill is approved, it comes with deductions. Insurers blame hospitals for lack of standardisation; cost break-ups are opaque, and there is no hospital regulator (which is a big pain). Hospitals say insurers arbitrarily question hospitalisation and make random deductions from the claim amount.
But caught in this crossfire is the consumer who, despite cashless insurance, is often made to wait up to 6 hours at the hospital at checkout. And even then, the claim money comes with multiple deductions.
@preeti_kul speaks with two relatives of patients who are knocking on many doors to get their rightful claims. And speaks with experts on how to break this logjam.
This is a story that must be repeated.
2
236
We love Hybrid funds. More than a third of our AUM is Hybrids. For me and Aggressive Hybrid over long duration gives a better outcome vs Nifty at lower risks. Depending upon a person's need, there is a hybrid.
Mutual funds are understood as only equities
Hybrid funds have played their role of cushioning flat to negative returns in least 3 years
High PE markets - hybrid funds and SiP
Low Pe markets - equity funds and Lumpsum
This is an easier framework than predicting future euphoria or extrapolating current pessimism
Where are we today
Neither cheap, nor expensive
Moderate risks - moderate returns - immoderate time horizons
2
315
I have a few contacts - they don't want to add earnings to their incomes as they would need to pay a tax.
Some - I don't want to pay higher cost irrespective of higher gains.
I get requests from investors everyday to review their portfolio. Upon discussion, I find the common set of funds
Nifty 50 Index Fund, Nifty Next 50 & A Flexicap fund
Most of them are in early 30s who need their investment only after 10-15 yrs. They have higher risk appetite.
I ask why stay with Nifty 50 & Next 50 when the tenure is long & risk appetite is high?
Answer is Low Cost! 😀
2
510

























