Can anyone help me here : when I tried to buy this stock for tracking purposes , broker didn’t allowed me to buy .
Read a Nvidia newsletter on similar line of business, link appended Using AI, Upstart Boosts Loans and Lowers Rates | NVIDIA Blogs. This raised my curiosity and found few more software startups working in a similar field, applications built using AI/ML APIs for lending.
Few weeks ago when I first saw the post here I thought this is a plain NBFC and I had all the apprehensions every investor would have. Invested 2%, with a hope the management acts diligently. When I studies the company I was excited with the technology back bone.
As per me, the biggest risk is the recommendation engine (AI/ML application) they are using for approving loans, how efficient is it? What kind of testing did this go through, etc?
Since these kind of solutions are implemented in several countries and did not encounter any large Indian company in similar business I wanted to try my luck with Ugro. I am still excited about this investment
Disc: Invested 6 percent, very risky business
At a global level, use of ML models is very common for lending, approval and rate decision purposes.
But use of AI is new thing, global banks still shy away to use AI as in such banks every decision has to be backed with logical and ethical reasoning which is not possible with AI. so, startups and nbfc are able to use AI as their internal regulatory are not as stringent.
I agree with your point of view, but when I think of every major disruption started small, and there are million failed ventures for every successful story, time is the only answer for companies like Ugro capital.
Bottom-line, this is a very high risk and reward investment
Disc: invested, biased opinion
Stock is on Lower circuits today. Is there any particular news about the company? Nothing in the media and no updates by the company yet.
Normally for stocks belonging to X category, Angel does not allow orders from app. You have to call them to place orders
Even in india most lenders use ML based models. From my analysis and scuttlebutt ugro does not have much of a moat if any. Lot of fluff. It’s essentially an NBFC without a track record. Value it as you must.
Btw curious to know how you differentiate between ai and ml. All my college research was in ml and my work at my company is also ml this is the first time I have seen someone imply that ai and ml are different until now everyone acts like they’re the same so curious to know what you meant 
Will try to keep it short and simple.
Discl: I work at a global bank as analyst. So, whatever I say, not everything is baseless.
I myself am only familiar with ML only. But I can give my 2 cent knowledge on AI. ML is only a subset of AI. AI includes many other things like deep networks, neural networks etc.
A complete self-correcting and self-improving system is possible with AI, which helps create an automated system(saves company resources) and adapts real-time, giving customer the best service(imp for any retail and b2c business).
There are many other possibilities like chat bots, intelligent systems etc which are less relevant to lenders, but banks can narrate stories using transaction histories of customers. They know your behaviour better than your wife and parents(not literally).
But, AI has many black boxes. It doesn’t need algorithm/logic input like in ML, user needs to focus on AI architecture design more, the neurons will handle rest and predict for us without a well defined logic. Such a thing often raises ethics issue.
Link(3 min read only): 4 ways AI neural networks will disrupt banking - Bobsguide
Hope this post adds some value!
Please do not invest in this or any other lending stock on the basis of purported ML / AI systems. Particularly if they are SME lending cos - the necessary granularity of data is simply not there. These systems are good at picking out fraud but not necessarily more so than existing systems, and none of them are very good at underwriting credit - which depends on very few factors , primarily income, cash flows and access to liquidity. Credit can go wrong for many reasons , and the business is completely in the ongoing management of credit and collections. A lot of risk is mitigated not through underwriting but through structure i.e. modulate amount, loan tenor, security construct etc. Where an ML system can help is with process compression and reducing TAT on applications - TAT is a crucial metric in a competitive credit space like SME or consumer, so a good system that can weed out fraud and check 80% or so of what a credit analyst would do is by now a basic requirement. Plus there are several tech companies providing these tools for use, so there is not much proprietary IP here - what makes a difference is the lenders own credit policy and processes.
In UGro , the thesis imo ought to be the emergence of a new class of SME lenders that are supply chain driven , adopting a vertical approach using anchor relationships and technology to expand into new SCF asset classes. This is an emerging global asset class , with other prominent private NBFCs are also targeting in various verticals. Globally , it became infamous due to Greensil, which collapsed due to over expansion and fraud , but the basic tailwind for SCF is a secular one. For UGro specifically, it is young , and while the management is starting afresh, the Religare overhang is not to be dismissed. It is currently light on leverage , has good access to cheaper capital , and is competitive on rates for its target segment and is growing presence in digital ecosystems. Given the high equity capitalisation and largely secured nature of business, do not expect lasting impacts from lockdowns. However , business is young and much remains to be proven. Take note of the risk and size positions accordingly.
Disc: Invested
Which broker you suggest I can use to trade in catagory X smoothly ?
The stock is in ASM list. Zerodha allows you to buy illiquid ASM stocks using additional security measures, like enabling 2FA / TOTP. I did such to take positions this week.
Some more bad news on these kind of platforms As AI-based loan apps boom in India, some borrowers miss out | VCCircle
Employee cost in 2019 and 2020 is 58.44% and 46.19%. Before 2019 it used to be in single digits. Can anybody let us know why this changed so much?
Thanks in advance
Ashwath
The quality of this thread and speed at which the collective thought process has been built is exceptional. Proves the quality of Valuepickr as a platform. Wanted to add my overall thoughts on the business. I apologise if it’s repetitive.
Tech should not be the primary investment thesis here. Tech is just an enabler in this business and should be available in plug and play format to most industry participants in a few years. They will certainly have a head start over traditional NBFCs over the next few years due to their tech-focus from day 1, but it should not be a sustainable competitive advantage. The story here is two-pronged:
(1) Sector specific focus
(2) PE-backed hyperscaling
(1) SME sector specific focus - This is, I feel, the most exciting thing about Ugro. SME lending is a difficult business which most banks/NBFCs haven’t been able to crack due to high NPAs and difficult scale-up. But difficult businesses, if built with relentless focus, over a period of time create the biggest MOATs. They have taken 18 months for sector selection and the deep research put in to create sub-sector specific checklists is super interesting.
Moreover, distribution strategy is brilliant. They are trying to create touch points around each and every participant in a sector thus covering all funding needs. (1) Creating a pool of anchors in each industry and tapping their entire supply chain to provide funding (2) Going deep into the machinery finance ecosystem by forming relationships directly with the OEMs (3) Creating branches in relevant micro markets instead of financial hubs (Ex: Bangalore branches are in Peenya and Bommasandra which are the biggest industrial areas there). (4) Creating upstream and downstream partnerships with sector specific fintechs and NBFCs. Within a few years, they should have deep penetration in their selected sectors knowing most suppliers, vendors, customers, financiers.
Imagine the scale and the barriers to entry if they execute all this over the next 5 years. All this knowledge and data should lead to best-in-class NPAs thus solving the biggest pain point in the SME lending industry.
(2) PE backed hyper-scaling - Building a innovative and challenging business like this at scale usually takes 15-20 years. But Ugro is trying to blitzscale this in 5-7 years. Capital is a barrier to entry in startups and a $150m capital infusion (usually done at series E or F in startups) at the onset reflects the capability of the founder and the kind of head start Ugro has over other companies in the same space.
Moreover, the quality of the investors which have participated should give a lot of comfort to most of us. Private equities are not entirely passive investors and would have done deep due-diligence before entering and should also be involved in the future at strategy level. This should ensure good corporate governance as well.
They are aiming to double their disbursals every year for the next 5 years. Frankly, this kind of growth, is only possible due to PE money that they have raised and we are very lucky to be able to participate in such a growth story.
Disclosure: Invested
Not an investor but following the topic.
What about the concerns raised by fellow members? Haven’t you come across any negatives regarding the model, the management, execution risks etc?
I could see only the positive side from you, don’t you have any questions or concerns that could effect the story?
Hi Chaitanya - Being a early stage business with no historical performance makes it a risky investment. So it remains to be seen whether the management will be able to deliver. On paper, everything looks great but needs to be followed up with execution. That’s the one thing which should be considered before making an investment.
For me, the potential rewards currently outweigh the risks, hence I took the plunge 
Attended the Concall. It was 1 hour long and about 35 minutes were spent my Mr nath talking about ugro as a whole ie stuff we already know… This is understandable since its such a new company and its pretty obvious the participants don’t know much about it either. I couldn’t get my questions through (regards strategy at micro branches considering they don’t really use Google maps even now AND regards the increase in borrowing limits from 3000 to 4500 crores) since a lot of time was wasted by people asking basic questions and getting basic answers and then the time ran out. Q4 was barely addressed and I’m getting along to this line of thinking too and giving more weight to full year’s data since reading each quarters result of something so fledgling paints a very incomplete picture at this point of time. The only somewhat new information I got from the Concall was
Overall I’m just happy a small company like this conducts Concalls regularly and gives us presentations and replies to emails. They seem very confident long term and look like they have a clear understanding of the domain in the way they answer questions. So nothing has changed. This is as risky and fun a bet as it ever was. Expecting the next full year(or even years) to be a bumpy one and will continue to add to my position slowly but surely during this time and the next few quarters should offer some good buying opportunities until the opex costs go down below 50 percent, the interest margins slowly improve, borrowing costs slowly decrease, the final view on npas comes out and as leverage increases from sub 1.0 in what could be an upcycle but im not expecting any fireworks here for atleast another year or two.
Disc: same as my posts above. Invested… More of an educational bet on seeing the rise of an nbfc from scratch. Not a sebi advisor
Note: investor presentation used during concall
8cee4a1e-b156-49f2-9a75-3fd1f7d01633.pdf (3.8 MB)
IMHO, designating vertical heads as CEOs gives them the freedom to grow business while platform takes care of the governance aspects. Adding an extra layer sounds redundant
Ujjivan and …: majority of lending they do are not secure hence the problem
UGro sanctions loans to deserving borrowers by mortgaging machinery while other NBFCs don’t mortgage machinery, this is a big opportunity
Some borrowers pay anywhere between 20 to 40 percent when they borrow money from unorganized players, same borrowers who have adequate cashflow borrow from UGro Saathi at 18 percent improving their cashflow by at least 15%
My observations: young company, lot of energy, innovative ideas
Disc: holding 6 percent, not a recommendation
This is interesting… but what is the realisable value of such a mortgage in case it needs to b auctioned. i dont think such machinery have a huge market when it comes to auctioning…
whats their LTV in such a mortgage ?
Discl : invested. Wanting to know more about buisness model before sizing up.
Good interview of chief risk officer.
One thing to note from interview as they have disbursed around 50% loans in Sanjeevni scheme after pandemic and affected severely in may and june which would be reflected in this quarter.
if there is severe 3rd wave, they can get into big trouble.
Disc: tracking
