vishwas

Breaking down of investment philosophies | Stock market | CSE graduate - NITK '25 | Software developer at Arista Networks

[vishwas @gpvishwas20]

Hi guys,
Thought of sharing a project which I had been working for 3 months on how do we solve the problem of able to do stock analysis on future potential rather than on past performance.

The dashboard is so rich that it cannot be covered in one session and neither can an investment framework be built in one lens, so do give it a look.

I will be covering a lot lenses in future posts and I bet there will be many who would want to be familiar with the terminology or how to use this view to pick stocks that fit your investment style and later do your own research.

Open to answering questions in here - Actually would be a nice way to share how we invest.

This dashboard is like a smart screener you apply based on your investing style on estimate data.

[vishwas @gpvishwas20]

Evidence that you are not paying a premium if you are buying the stock at current valuation and there is enough left in the table.

ROCE expanding in the future, better oncology mix, better payor mix, better ARPOB would surely rerate the stock to better multiples and they have enough cash balance sheet strength to improve operational metrics.

[vishwas @gpvishwas20]

Yatharth Hospitals seems like a good bet if someone is interested to have exposure into hospitals sector:

This would not be a deep-dive but I will be mentioning a few thesis forming metrics on which I would place my analysis on:

i) Ideal Hospital stocks operational metrics
ii) Which cycle is the company at?
Capex heavy or capex utilization phase
iii) Have the management delivered on their promise previously?
iv) What are the things to watch out for?
v) Small video on peer comparison on EPS growth and valuation against the universe.

[vishwas @gpvishwas20]

Guidance outlook to monitor:
i) Maintain revenue growth above 35%, EBITDA margin above 24%, ARPOB growth of 8-10% over 3 years.
ii) Reduce government mix from 35% to 25% over 2 years.
iii) Add 5000 beds within 3 years with 1500Cr capex outlay.
iv) Two hospitals ramp up in H2 FY27 and become margin accretive.
v) Bring back ROCE from 12% currently to 20% range over 3 years.

[vishwas @gpvishwas20]

@Sachan8574 I would choose Option A as at no point will I want all my picks to underperform.

[vishwas @gpvishwas20]

@vishan_29 Thanks Vishan,
I will research the stock and when I understand the nature of the business I will get better clarity.

[vishwas @gpvishwas20]

@ThePupOfWallSt Thanks,
I ran through your list of stocks in my dashboard and these are the aggregate current and next fiscal numbers. The value to growth seems most lucrative for Microchip Technology, On, Infineon, Flex and Vertiv Holdings.

Will be studying these companies.

[vishwas @gpvishwas20]

@WealthEnrich That explains why SIPs work better for most people. They do not have to think everytime every month. There is no instant gratification buying a stock, buying a gadget gives some joy or comfort for a few days and hence we lean towards.

[vishwas @gpvishwas20]

I think that's why CDMO companies are trading at a higher valuation.
The market opportunity is so large that it's all upon individual companies to execute well.
No question on the TAM or industry growth right now.

I think we have transitioned from manufacturing APIs pre 2020 to speciality molecules post 2020 and now on the verge to enter into complex molecules like those used in oncology, gene therapy which carry better $/kg.

[vishwas @gpvishwas20]

@vini546 Memory components have more than quadrupled over the past year.
They have now become significant share of the COGS in electronics.

[vishwas @gpvishwas20]

@connectgurmeet This is what happens when you speculate not when you go by true numbers.
AI is different from commodities.
Have you seen companies trading at 10-20 times Forward PE and still call them overvalued delivering 50+ growth now and in the coming two years.

I don't get the argument.

[vishwas @gpvishwas20]

@StockMKTNewz What are the 5 stocks that are exited from Nasdaq 100?

[vishwas @gpvishwas20]

It was so disturbing to read a thread where software engineers who could choose teams previously are being stuffed into what is called a "data labelling job' - Post training phase during the RL phase where the model is taught what is good and what is bad.

Meta almost deprioritizing their existing ADs business and going all in on building a frontier model.

[vishwas @gpvishwas20]

I will just make one argument that makes memory detrimental to AI success:

Higher the HBM capacity, higher the DRAM capacity the longer your session's cache remains.
Currently with Anthropic models the cache TTL is 5minutes by default and prompt caching therefore is no better.

GPT models on the other hand have cache TTL upto 24 hours meaning it could be a few hours as well but way better than 5 minutes.
This means you will see higher cache hit rate while using codex as compared to claude code.

Prompt caching is very underrated and not many people are talking about it.
They just talk about the base price , input and output per million tokens rather than cache reads and cache writes cost per million tokens.

[vishwas @gpvishwas20]

I will just make one argument that makes memory detrimental to AI success:

Higher the HBM capacity, higher the DRAM capacity the longer your session's cache remains.
Currently with Anthropic models the cache TTL is 5minutes by default and prompt caching therefore is no better.

GPT models on the other hand have cache TTL upto 24 hours meaning it could be a few hours as well but way better than 5 minutes.
This means you will see higher cache hit rate while using codex as compared to claude code.

Prompt caching is very underrated and not many people are talking about it.
They just talk about the base price , input and output per million tokens rather than cache reads and cache writes cost per million tokens.

[vishwas @gpvishwas20]

@StockSavvyShay I was wondering what Micron will do with so much cash on the balance sheet.
$25B in capex for the year but $70B + in profits this year.
Micron is following the Nvidia playbook now, invest cash back into the ecosystem.