I had looked at Asian energy services few weeks back in the midst of the oil crisis. However, for some reason i could not find the valuation very attractive. It would be very helpful if @phreakv6 sir could help me out on the valuation part. You have put 15x EBIDTA as a fair price for the standalone business. Could you help me through your thought process here. Secondly, i understand that based on guidance on oilmax for fy29-30, a low to medium multiple will give the valuation that has been arrived at by the independent valuer. However, can we independently verify with decent probability. What i am unable to think through here is how to value oil assets. Are they like buildings, or land or cash, or something else? What are the things (and how many ) that can go wrong before it takes oilmax to pump enough oil to justify its valuation. And how much time will it take. I had written off this business based on my lack of understanding on valuation but since i read everything you write and deeply respect your opinions, it would be helpful to me if you could throw some light on it.
@sumit680 - One year forward, it should have 450-550 Cr topline (which is backed by healthy order book and execution is ~12 mos + AIC consol), and I think doubling of PAT is a possibility in FY27 (maybe too optimistic but they can land somewhere there and I think its probably trading ~30x 1 yr fwd). In a year, it will get ready for a mainboard listing and possibly also a merger with group companies which will strengthen overall capabilities and also remove the RPT overhang. In general companies improve corp gov at that point as mainboard requires better discipline.
So at 2500 Cr mcap, for the capabilities and growth, I was thinking its not bad. I would like to be convinced about a business first and if I am convinced, I would buy on 20% dips, so somehow overpaying doesnât bother me too much, because if I find something expensive, I donât buy too much of it anyway. However, there are some caveats I should add here - though margins look great here, the working capital is bad - both receivables and inventory - so essentially this company is putting money upfront and working on the orders and collecting money much later - so it has almost nil pricing power. I have seen this in some companies where margins look rosy (I think tejas networks also used to be like this) but the devil hides in the balance sheet. Here however, the company is small but if things do not improve in a year or two, market may not be kind. So far there is no debt on the balance sheet until last quarter and that seems to have been taken to finance to acquisition. I think there is no cause for alarm (yet). Growth opportunities like this are rare so I am willing to err on the side of optimism and be wrong with equanimity if it happens.
@Ravi.0104 I dont have great understanding of it but looks like North-east gas grid (NEGG operated by Indradhanush Gas Grid Ltd) is getting ready and phase-1 was operational in Mar '26 and I think Amguri can then ramp up comfortably (so far I think they are selling within the state to Assam Gas Company (AGCL). Mevad/Indrora are in Gujarat so donât have this problem. Duarmara is currently not operational as production is yet to start but my understanding is that they will evacuate through DNPL (duliajan-numaligarh pipeline) operated by Assam Gas Company. This pipeline appears to connect to NEGG, so I am assuming this shouldnât be a problem and they can sell to Numaligarh refinery. Only thing to track then I think is NEGG progress which should be available by the time Oilmax fields ramps up in a year or more. My understanding could be naive - please double-check. I am not too worried about this because - we are not overpaying here.
@Vineet_Dubey - I am no expect in valuing oil assets. I saw that the independet assessment had landed at 2,582 Cr which is inclusive of Oilmax holdings in AESL. AESL stake is worth 1012 Cr as per market value, so Oilmax oil assets are worth 1570 Cr. Is that too much too little? I donât know. I know they made a PAT of 50 Cr in FY25, so its 30x roughly. I saw in some concalls that this was claimed as 50x but I think the math is wrong (they have used the full 2582 Cr as value for oil assets instead of subtracting the AESL stake). I verified this based on swap ratio, the shares extinguished and per share value and it works out neatly. There are three more oil assets which are not producing and Duarmara alone can dwarf the rest when it starts producing. I used claude to work out NPV and it worked it out to ~$180m. The valuation done by independent valuer also is around this ballpark which is why I think its fair. However, Oilmax is not the only one owning these assets (only 65%), the other 35% guys did not have a exit liquidity so far - considering market offers that, the valuation should have discounted another 10-20% perhaps. Irrespective of this, at 330, its a safe price that provides that discount to you which is why I feel chances of losing money here is low. It might remain cheap and a value trap though.
Hi Bharani, Newbie here and trying to follow your thoughts and been reading about your AI local runs with qwen and other models. Just wanted to understand what kind of analyis can be done using the model with local model do you set the context with datasets and is this less prone to hallucinations as the dataset is limited. Iam no expert on the LLMS and this rather my thoughts from general reading articles. But interested to set something locally to explore the usage of local models. Any opinions is appreciated to get going. Thank you in advance :)
Consolidated Phreakâs investment strategy, portfolio, and macro views (2024-2026) pulled together from his posts across this thread to help newcomers catch up without scrolling through years of history.
Note: Treat it as a starting point and cross-check anything specific (multiples, order books, growth guidance) against Phreakâs original posts.
Analysis of Phreakâs Investment Strategy, Portfolio, and Macro Outlook (2024-2026).pdf (460.8 KB)
Thanks for making this. Though it summarizes what I have done in the past, I donât think i have a coherent framework. The only thing i believe in is that frameworks need to be flexible and adapt. Most times I am bunch of contradictions and lack clarity of thought. I donât see this is as a drawback to be dealt with but as carefully nurtured cognitive dissonance of a healthy mind that can hold multiple contradictory thoughts, ideas and value systems on par without trying to forcefully reconcile them. It is for this reason I refuse to go into podcasts or do interviews, because I truly have nothing to offer thatâs of any use.
Few things I would recommend though which are timeless
-
Read books and read a lot. Read outside of your comfort zone into multiple disciplines and expand your circle of understanding. More you read, more you can read. Donât read with the objective of making returns. Read to understand and become better. To simply make returns, you donât really need to read at all beyond some basics. Real knowledge though will keep you grounded for life.
-
Temperament likely correlates more with returns than intellect. It is worth developing a good temperament and it can help with multiple things. I see temperament as the equivalent of compression in music production (look it up if not familiar), so develop thresholds and be aware when going beyond them have ways to bring yourself back down or up within range.
-
Work on your health with good focus on food, breathing, sleep, exercise. Read books on these or watch some good sources to understand best practises.
-
Mental wellbeing is very important as physical. Develop hobbies, preferably at least 2 or 3 that are complimentary. Spend more time with family and friends and in general be kind.
-
Information is cheap, truth is expensive. Not everything you see or read can be taken as truth. Watch Kurosawaâs Rashomon - its sometimes impossible to know the truth - not because people are lying but because each personâs ego and self-image can reshape his/her memory itself (so they are not lying in the conventional sense). Truth might be unrecoverable once it passes through human perception. Take everything everyone says with a grain of salt. You could choose to be skeptical of everything always or trust until you canât and be prepared to be the fool once or twice (I prefer to be the latter nowadays, though being skeptical is what comes naturally to me - I had to unlearn it)
-
Avoid extreme views, political or otherwise since itâs impossible to know truth. It makes it very hard to change views as well and itâs very important to be able to do so, on anything
so much true. agree completely. information everywhere no truth
You are providing life mantra to all the readers of Valuepickr! Thanks a ton for that. Your thought clarity (or lack of it in your eyes) is a refreshing breath of fresh air! I am always eager reader of your musings (although I dont çlaim to understand all of it..
)
thanks once again for your insight. Have been following and reading your insight, with admiration and adoration. read a few of your recommended book and enjoyed them. please continue and contribute but a bit more frequently.
DIvyesh
Thanks to you n Bharani sir too. I hold a special place for him.It would be a guiding beacon for the newcomers. You have done a wonderful job and respect to Bharani sir too. Kindly check the notes on Qpower n Aeroflex enterprises too.You may add these too. Just a suggestion. Thanks
Please make a list of books mapped to theme and share in this thread or X wherever u prefer.
The line triggered the memory of a post written by Professor Bakshi, aka Fundoo Professor.
Saw this post on the Quality Power thread. I have been trying to learn how to use AI tools (Claude) to research companies faster and in a better way. Took a subscription of Claude Cowork and implemented a workflow. Made a short ppt on my progress so far and the workflow that I have implemented.
Previously I used to spend days, if not weeks, going through ARs, concall transcripts, corporate announcements etc. to research a single company. With the new workflow I am able to do that in a couple of days. Someone with a high degree of focus and grasping power should be able to do the same in a couple of hours.
While researching another company, it not only saved me a lot of time but it helped me unearthed redflags by digging out discrepancies between management commentary between 2 concalls. Its worth spending 2k per month on this.
Claude_Cowork_Stock_Research_Workflow_Subodh_Malgonde-21-Jul-2026.pdf (419.3 KB)
@fabregas - thank you for sharing the workflow, this is helpful.
I have been following Anil Tulsiram for sometime now (through his blogs - https://contrarianvalueedge.substack.com/). He has a very detailed foundational, first principles approach on how to use GenAI for equity research. He has shared some of his detailed presentations on this topic (150-250+ slides - available on his blog) which are genuinely useful and deeply insightful on how to use GenAI in general and specifically for equity research. Incase anyone is interested, sharing his latest presentation. All presentations are available on the blog as well. Warning: his thought process and approach runs deep, so if you do not have patience and genuine curisoity to learn, you will get bored beyond first 10 slides. But if you can persist, there is wealth of knowledge and insights to gain on this topic.
Happy learning!
P.S: if uploading this presentation violates any forum guidelines or copyright issues, please let me know. Will delete it. Beyond reading Anilâs blogs, i do not have any commercial connect or relationship with him.
@phreakv6 sir, may i know your views on TECHERA, tech era engineering india ltd. Supposedly they have won the bid to be direct vendor for IAF. They are doing tooling for HAL.
@phreakv6 Iâm a techno-funda investor, and one challenge I repeatedly face is position building.
Suppose I identify a fundamentally strong company with high growth potential and a bullish technical setup. I initiate a small position after the breakout, but I struggle with deciding when and how to add more.
Earlier, I tried adding on pullbacks to the 50 EMA, but Iâve noticed it isnât reliable. Sometimes the stock respects the 50 EMA and resumes its uptrend, while at other times it breaks below it and corrects much further.
For example, Iâm bullish on Viyash Scientific because of its recent merger and believe it has the potential for a turnaround. It has already given a breakout and is now retracing. I currently have only a 1% allocation but want to build it into a meaningful position. However, Iâm worried that if I add now, the stock may continue to fall.
I face the same dilemma with stocks like Sakar Healthcare, Aditya Infotech, and Ather Energy. My fundamental thesis remains intact, and I believe these businesses can continue growing, but Iâm unsure about the best technical approach to increase my allocation.
My question is: What is your framework for adding to winning positions? Do you use moving averages, swing lows, volatility (ATR), volume confirmation, staged buying, or some other technical method to build a position while managing downside risk? How do you decide when a retracement is a buying opportunity versus the start of a deeper correction?
if we compare aimtron with companies like syrma, avalon, Kaynes, etc, their margins appear high due to reasons mentioned by you: the working capital differs between these companies + aimtron is getting benefit of smaller batch size based on SKU presented in the IP. But wouldnât this mean, as they scale up, the margins may trend downwards? The sectors they are dealing with are not very different from the listed peers, so couldnât get my head around the sustainability of the margins.
Also, their disclosure have been a bit weird.
E.g.,
- Yesterday they disclosed their consolidated revenue jumped 94% YoY, but wouldnât this include the revenue from their acquired unit which did ~1.6 million in 2 months last year and they were expecting 170 cr for FY27?
- A disclosure made last week where the headline read âAimtron Secures Pilot Box-Build Manufacturing Order from a Leading Global Semiconductor Test Solutions Companyâ, but the order was for âBox-Build manufacturing of a next-generation intelligent fibre network test and certification platformâ which may not be as sophisticated as what other EMS players are doing for semiconductor equipment players. Not sure, if the understanding is correct or not, but would love to hear your feedback.
@phreakv6 How do you see the aeroflexâs results?
Thank you all for the kind words and feedback. Thanks @fabregas for sharing your research process with AI. Nice to see Anilâs in-depth slides too.
@robin1 - No views on TECHERA. I dont know anything about the company
@mugni_shaik - I think you should improve the quality of your conviction. If you are bullish on Viyash, and feel it in your gut, you would not have stopped at 1% allocation. It means you are not yet convinced enough to scale it. As I mentioned in my prev post, the best ones that have made money for me are the ones where I couldnât get sleep because I did not have enough in the position - even at 15%. It should feel slightly insane. But I wouldnât recommend blindly scaling it and feeling nervous. Its like how all pain is not same - there is good pain after exercise or trek and then there is pain of a sprained neck or migraine. You should feel a good nervousness and not anxiety.
To get there to that level of conviction, you must work deeply in the companies and sectors you own and have a fair idea of where the sector is headed and where your company stands in that sector. You must ignore breath - depth always comes at a cost to breath. A willingness to say no to lot of other ideas is just as important. Doesnât mean you become an ostrich in the sand with your positions but cultivating a detached sense of attachment (strong convictions loosely held) that allows you to remain committed for a period of at least a year. So before you commit you must know enough to decide which stocks deserve a commitment.
Imo, a whole lot of stocks do not deserve a commitment - all cyclical commodities like paper, sugar, financials, textiles and sectors that survive on govt subsidy etc. are worth avoiding unless you have a fairly good idea where we are in the cycle and how past cycles played out and how long you have left - I never can figure these out with any reliability so I have learnt to pass on all these. Once you have decided a stock needs a commitment based on sector tailwind, overall TAM for the business and companyâs capability (tailwind, tam and moat - in that order), you must look at valuation. If valuation is in favour, start with at least a 10% position and be ready to scale it up to 20% when valuation comes in favour. To do this, you must have enough conviction to not find reasons not to buy when price has come down 20-30%. Most people cannot do this because they never developed enough conviction in the first place. You must also be ready to scale up higher if required. Starting with a 10% position allows you to add another 5% quickly if required - this is my preferred method (pyramiding up) and I have done it in almost every position of mine.
To develop conviction, I recommend reading all research reports on the sector. Reading books is my preferred way of doing it. Being immersed for hours understanding the industry through a book is unparalleled in developed conviction. Before investing in Sai for eg., I read several books like Ten Drugs, Billion dollar molecule, Genentech, Magic Pill etc. so I had a fairly good idea of small molecules, biologics, innovation, biotech funding, approval process etc. A book like âThe gridâ gave me great understanding of US power sector and so on. There is simply no substitute for reading when it comes to building conviction.
If I wanted to understand the animal health sector - I would split it up into companion animals and farm animals and read what type of products exit in each (broad categories like products for external parasites like ticks, internal stuff like dewormers, cosmetic products, neutraceuticals, food additives and so on) and find books like zoobiquity (commonality between human and animal diseases), spillover (how animal diseases spread to humans), big chicken (on antibiotic use in farm animals), rabid (tracing rabies through history) and so on and immerse myself in robust sectoral knowledge. Reading latest annual reports of Zeotis, Boehringer Ingleheim would also be a must. Once I do that, I will look at what are the top selling products - both innovator and generic and who makes them and what their sales are. Which cdmos cater to this demand and where exactly does Viyash stand and what products they make etc. Without doing all this, I cannot buy a single share in the company at this point where I am.
I donât know if its a good or bad thing. I will stick to it as long as it produces returns.
@Karan_Sanwal - I donât think they will scale to doing things like Dixon. I feel they will continue doing what they are currently doing until they canât grow further. I think next 2-3 years (which is what my holding period will be at most) there is enough of the same growth available going by demand trends and their order book. Yes the growth will incl AIC since its now a subsidiary. The key to me in both Aimtron and AESL is the valuation of the acquisitions. This is exactly how Varun beverages has grown for years by simply acquiring territories and 1x p/s and then when it stabilises margins and consolidates, the same gets valued at 10x p/s. M&A works rarely but the easiest way to make it work is to acquire capabilities or acquire cheap or do both (AIC was acquired at 5x EBITDA and has great complementary capabilities in ruggedised electronics and market access to equipment sector - John Deere, Caterpillar). I do not know enough to comment on their recent order.
@Rakesh_Choudhary003 - Aeroflex results are very good. Scaling of skids business is going as planned and if they can get to 60% util as planned by Q4, we might be seeing 70 Cr levels of revenue from skids business alone and good thing now is base business is also growing 15-20% so we can expect a 180-200 Cr topline in Q4! Considering 40 odd skids make up 1mw, the capacity is only good for 350 MW or so which means there should be a capacity expansion coming by the time of Q4 concall in May '27. My guess is 15k skids capacity will be fully sold out in FY28 and we could see close to 900-1000 Cr topline with 220-250 Cr EBITDA in FY28. Everything seems to be going on track so far. Considering how long it takes for machines to be delivered, I just hope they dont delay announcing expansion.
Recent quarter results also show 200-300 bps margin compression from inc in staff cost (for the expansion) and digital initiatives (erp/crm) and also cost of expansion itself. Considering hoses is 16-20% margin and hose assemblies and fitting is 22-26% margin and 65% of latter, I think ex-skids business might be having margin of ~22% or so which implies that skids margin must be closer to 30% and promoter doesnât want to disclose this in a public call. Overall I see thesis playing out as expected here.
On overall AI/DC thesis, few things have however changed. Open weights models have bridged the gap strongly with Kimi k3 now on top of fable 5. Thereâs a coalition of support among tech companies and hyperscalers (Everyone except anthropic) towards open weights AI which implies USG might be thinking of regulating AI so that Chinese models are not used by US corp. This would be a dumb thing if it happens so I think all these companies are pre-empting it and acting to show support to open weights models.
This changes things materially because no longer do DCs have to be in US jurisdictions predominantly (which is what OpenAI and Anthropic were preferring - to avoid theft of their models). DCs might get built all over the place now to run open weights models. Imo, HCL Tech, TCS etc. need not sign any deals with US AI companies at all and can build their own DCs and run Chinese open weights models themselves. With DCs in middle-east under threat as well, I think sovereigns that can protect their DC assets with cheap labour and power will be preferred zones for hyperscalers, so India might see a favourable shift in all this.
I donât see this as a big negative for Alphabet, Amazon and Microsoft as most of the open weights models will still be served from AWS, GCP and Azure. I think this trend is seen in prices of stocks as well as US AI infra companies are taking a hit in sentiment whereas MAG7 was in healthy green last night. In India as well Aeroflex was strongly in the green vs TD/Mtar. This sentiment shift however I think is overdone and there should be a strong bounce in the short term in US DC stocks too (lot of it is kospi correlated leverage unwinding I felt). Fundamentally though there is a shift and if more DC capex announcements happen in India, that would be a good indicator of it. I also see more smaller DCs being announced (5-25 MW) as its easier to raise few hundred million dollars than few billions and being able to serve open weights models assures a decent return on capex for smaller providers. We are bound to see fragmentation but a bulk of the demand will still be served by hyperscalers. As long as thereâs a strong spread between their cost of capital and rental yield on DC assets, they will continue to bid up memory and gpu prices. The ROIC will keep reducing though (the payoff period was just 1.5 yrs last yr and is bound to drop to 6-7 yrs going forward) - for neoclouds as well this is a negative as these should be valued on a P/B basis and so as RoE drops, the P/B multiples should compress.
Thank you for such a detailed and insightful reply. Your framework on building conviction through deep research and disciplined position sizing adds immense value.
@phreakv6 it will be good if you can share your thoughts on Holding companies , i see lot of Good companies trading at abnormal discount to their holding value i.e Bombay burmah (Mkt cap 10.5k) , BUT controlling 50.1 % of britannia , similary Naperol investment (MKt Cap 368 Cr) , however holding around 9 .1 percent of Bombay burmah etc . are these bad bets should one exit from these type of Holding Company stocks?