Reliance & Meta are together investing ₹855 crores to build AI Infra in India
RIL is also partnering with Google to build an AI data center in Jamnagar
At RIL’s 48th AGM, Mukesh bhai stated: “Jio promised and delivered digital everywhere and for every Indian. Similarly, Reliance Intelligence promises to deliver AI everywhere for every Indian”
“Reliance Intelligence” is RIL’s next bold bet after Jio ⤵️
Remember: RIL has partnered with both Meta & Google in the past for Jio
Jio raised ~₹150,000 crore from strategic & PE investors in early 2020 which included:
(1) Meta, who invested ₹43,574 crore (~$5.7bn) for 9.99% (announced April 22, 2020)
(2) Alphabet, who invested ₹33,737 crore (~$4.5bn) for 7.73% (announced July 15, 2020)
Btw, these two investments were strategic in nature e.g.
JioPhone Next was built with Pragati OS - a collaboration with Alphabet in Oct ‘21
Meta owned WhatsApp launched a deep e-commerce integration with JioMart in Oct ‘22
Jio is expected to IPO in the ~$110bn valuation range i.e. an acceptable 13% $ IRR for Meta & Alphabet
Meta & Alphabet may sell down some stake in the IPO (liquidity always helps drive re-investment decisions)
Btw, I had written about Jio’s upcoming IPO almost a year ago (nothing much has changed in the analysis) - bookmark for future reading:
RIL and Meta are forming a JV with 70:30 split to create a platform targeting Indian businesses to offer AI services billed on a monthly basis. Meta is the Tech partner offering its open-source Llama models & deep integrations with the Meta ecosystem products (e.g. WhatsApp)
Key features are:
(i) Local data residency
Indian user queries will hit local data centers & data won’t moved to SG/USA. We have already seen the RBI Mandate which requires payment related data to be stored within India; so there is clear biz rationale for the same
(ii) SMB focused products | WhatsApp focused
SMBs in India prefer Zoho & other SaaS partners who offer low cost products for tracking inventory, analyzing sales data & managing relationships.
RIL Intelligence plans to offer AI solutions like an automated sales analyzer which can scan bills & run customer campaigns on WA
(iii) India pricing
Since OpenAI has led the way here with ChatGPT Go - RIL Intelligence will aim to provide business solutions at an affordable India specific pricing (esp. since the underlying models are hosted in their own cloud)
RIL Intelligence is part of a wider RIL Group strategy - “Own The Stack”
The 3 principles of the “stack” are Intelligence, Connectivity & Power
RIL’s core business (>60% revenue contribution) lies in oil & petrochemicals
(1) Jio was core to building the Connectivity pillar for RIL
(2) Reliance New Energy was launched 4 years ago to be the “Power” pillar
PS: RIL has invested ₹70,000 crore to build a complex a 5000 acre complex to produce solar panels , batteries, green hydrogen and wind turbines (!)
RIL believes that data & energy is the “new” oil - hence investing behind this.
To emphasize the scale of the proposed Google X RIL Jamnagar 3 GW data center - it would be the biggest one globally (as of now) - would consume electricity equivalent to 6 Lakh houses
Akash Ambani mentioned ”We want to complete it true Jamnagar style in record time - as we have always done in Jamnagar - in 24 months.”
PS: I had explained the “Jio Stack” strategy 2 years ago in a post (again, bookmark for future):
(1) RIL Intelligence is incorporated as a subsidiary of the RIL Industries and NOT a subsidiary of Jio
(2) Unlike Jio Financial Services - for the Jio IPO, you won’t get allocated Jio shares - however you can buy Jio shares during & post IPO
(3) Given Jio’s journey, you can expect that RIL Industries will try to list RIL Intelligence in a similar fashion to Jio over a 3 to 5 year window - especially if it raises capital from PE investors
In short, if you have RIL Industry shares then you get proxy exposure to this new business unit & should benefit from the upside
It took me some time to go through the AGM notes since I’m in SFO right now - the “India question” (about both AI & geopolitics) has come up several times.
It is clear that AI Infra is a sovereign responsibility (hence IndiaAI Mission) and a hyperscaler opportunity (due to CAPEX requirements)
There are few deep pocketed & risk taking Indian conglomerates who can pursue this opportunity - RIL & Bharti Enterprises are two frontrunners here - expecting to see a LOT more in the next 18-24 months! Will write about the AI efforts from Bharti Enterprises in the coming days too!
Discl: Views are my own. Not an endorsement or promotion. Family & affiliates are shareholders of RIL.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
I read Ruchir Sharma’s new book “What Went Wrong With Capitalism?” earlier this week
Ruchir spent 25 years at Morgan Stanley & headed up their Emerging Markets investments ($20bn+ in AUM across funds)
He’s also a great writer on macroeconomics & history; there are 3 great points which he has made in this book ⤵️
First, I think it is important to set the premise of this book:
Ruchir highlights how constant intervention (bailouts, interest rate drops, welfare programs) from the Govt & Central Bank has altered the normal functioning of key economies:
(1) Recessions are less frequent but more severe
(2) Wealth is more concentrated (Top 1% of US controls 40% of wealth)
(3) Economies are becoming financialized (10% of income flows to Finance)
(4) Debt fueled economy: 80% of US lending happens in the shadow bank economy
(5) Big companies dominate: Low interest rates & high regulatory compliance requirements perpetuate oligopolies & monopolies
The funniest acronym used in the book was NINJA - a loan customer who has No Income (NI), No Job (NJ) and no Assets (A) - which is apparently Ruchir’s characterization of sub-prime lending in USA.
The 3 points most relevant to the common man are:
(1) Investing has changed
Gugenheim CIO said: "The Fed has made it clear that prudent investing will not be tolerated”
And, Ruchir adds to this by highlighting: "Investors get outsized gains during the boom times and benefit from socialization of outsized losses during the pain times”
Unfortunately, the current Govt approach (low interests + bailouts etc) stand to benefit the rich far more than the poor:
(a) Capital markets are overwhelming owned by the Rich | Bailouts help the stock market more than the economy
(b) The Rich have assets which can be used to access cheap debt to earn more income
(c) There has been a steady low of wealth from labor to capital (”ownership”) with automation, offshoring & financial engineering
The data shows this too - Share of income owned by the Top 1% in USA is back to its peak of 20%; top 1% own ~40% of wealth (still shy of the 1929 peak of 50%)
YOLO isn’t just for the retail pleb but also for the Manhattan institutional investor - you either make bank or get a bailout.
KFC’s sales per store in India is down 5.7% to ₹98,000 per day as compared to ₹1.04 lakh last year
The pressure isn’t just on KFC—McDonald’s, Pizza Hut and Domino’s are feeling the heat too.
“Maintaining last year’s sales numbers has become a big challenge” said Rajeev Varman, CEO of Burger King India
Yet, individual KFC franchise store operators struggling - KFC India has still managed to post ₹1,500 crore in profit (v/s ₹1,425 crore last year:
First, let us understand why this dichotomy has arisen: Store owners are struggling but the franchisor (KFC India) is making healthy profits.
(i) KFC India is primarily run by 2 franchises : Devyani International (DIL) Sapphire Foods India Private Limited
(ii) DIL runs 70% of KFC India outlets with around 700 stores. On paper they posted 6.6% YOY increase in revenue to ₹2,178.7 crores from KFC - but this was driven by new store opening
(iii) DIL’s avg sales per KFC outlet has fallen. The profit margins has fallen to 15.1% from 18.3% - primarily due to higher advertisement spending & discounts.
KFC rolled out the “Epic Savers Deal” — 9 fried chicken pieces (7 boneless strips + 2 leg pieces, the most popular cuts) for just ₹299.
And, they launched a “Wednesday Special” - for ₹100 extra , customers could upgrade to a bucket with 6 additional wings
“It’s one of our boldest & most compelling value propositions to date” - Aparna Bhawal, chief marketing officer, KFC India
And, they spent a BUNCH on advertising - roping in The Great Khali plus Mrunal Thakur and comedian Danish Sait!
Therefore, total Sales rose - store margins took a hit; but here lies the nuance of the franchisee business:
Franchisees pay 3–8% of Sales (NOT profits) as royalty and marketing fees to the brand
Therefore, if margins are hit but Sales go up - the franchisor (KFC India) doesn’t suffer as much as the franchisee (Devyani Foods International).
Of course, the franchisor also bears some part of the additional marketing expenses - but you can see in the numbers above that interests can diverge.
And, what other reasons are causing margins to get hit?
Karnataka has become the 3rd state to pass a bill to protect Gig Workers
Soon, every transaction you make on Swiggy, Rapido or Ola will result in ~1% to 5% of the total bill being deposited into a “welfare account”
India is projected to have 2.3 crore Gig Workers by 2030
Here’s what you need to know about this bill 👇
Firstly, do Gig Workers need a separate labor protection law?
(1) These riders are called delivery partners - they are NOT employees and collect their payments as “commissions” from the company for completing an order; therefore, the traditional employee laws do NOT apply
(2) These riders work for 12 hours/day - due to (1) they don’t get PF, ESI & other social security benefits given by government/employers as they are not classified as employees.
(3) There is some gray area: An Instamart delivery person completing 10-12 orders can earn ‘minimum guaranteed income’ of around ₹1000. But, it comes with lot of terms and conditions.
(4) Since their income is highly variable - so skipping a Sunday, taking a medical leave for family or relaxing on a public holiday leads to loss of income (unlike all of us who get a stable monthly salary)
(5) Their job is high risk - road accidents, dog attacks & injuries from carry objects - while they don’t have Govt provided insurance - most platforms today have a Personal Accident cover for their delivery partners
Therefore, I do think passing these Gig Worker Laws is akin to the Factories Act (1881) - a new class of labor needs a new set of protections
So, what does the Karnataka Bill cover?
(a) Every gig worker will be issued a portable digital ID - so if a rider switches from Uber to Rapido - they will still carry all past benefits
(b) Any change to payment terms, deductions or incentives should be altered only after 14 days notice. A partner cannot be terminated without 14 days of notice period except in cases of violence.
(c) Explicitly covers algorithmic accountability. Any auto-deductions that apps make e.g. penalizing a worker for rejecting a long-distance order, delivering late due to traffic or rain should be explained transparently in simple language
(d) Rest, sanitation, and clean facilities are a legal obligation for platforms, not a good will measure (which is what it currently is)
(e) A Welfare Board which will consist of 10 people - including 4 members from the delivery partner community!
What is a very tactical implication?
A logged-in delivery rider for 2 hours can take a break to drive 5–10 minutes off-route to find a toilet or rest stop. It will NOT be treated as delay, refusal, or inefficiency for not accepting the order.
According to MIT, 95% of organizations are getting zero return from GenAI pilots
MIT conducted 50 interviews, surveyed 150+ leaders & studied 300+ AI projects to highlight the “GenAI Divide” i.e. despite $40bn+ in spending on AI projects, only 5% of companies are able to extract real value.
Sharing 5 key takeaways for startup founders from this report ⤵️
Which functions are getting the largest AI budget allocation?
Per MIT, 50% of GenAI project budgets are going towards the “front-office” i.e. Sales & Marketing which includes:
Why? Outcomes in S&M impact top-line, output is measurable & deployments are visible (to customers & management)
The best anecdote (highlighted below) says:
"If I buy a tool to help my team work faster, how do I quantify that impact? How do I justify it to my CEO when it won't directly move revenue or decrease measurable costs? I could argue it helps our scientists get their tools faster, but that's several degrees removed from bottom-line impact."
As you’d expect - front office is a LOT of show - AI is unlocking real gains in the back office (more on this later)
Let’s start by seeing what Product Principles a startup needs to follow in order to succeed:
Context & Memory is the Moat - an Adaptive Flywheel is a must
This phrase is courtesy Bessemer Venture Partners in their “State of AI 2025” - MIT confirms this with the following data:
(1) ~60% of respondents rank “ability to improve over time” as a key selection criteria for GenAI vendors
(2) ~63% demand that their GenAI tool “retains context”
(3) ~66% want the GenAI tool to “learn from feedback”
So, while ChatGPT is great for quick iterations - true AI transformation is only possible if your product:
(a) Builds memory
(b) Retains context
(c) Improves via usage
(d) Integrates with existing tools & processes
And, there is a reason why Enterprises are acting:
India’s ₹70,000 paints industry has shrunk for the first time
Market leader Asian Paints has several problems at hand: Revenue remains flat, a legal suit from Birla Opus and the merger of JSW Paints & AzkoNobel (Dulux)
Analysts had dismissed Birla Opus saying it would get ~2% market share; the company is at > 8% market share in less than 2 years of launch.
Asian Paints has started to show some signs of recovery - Q1 FY26 PAT was ~₹1,000 crore i.e. over 50% higher than Q4 FY25
But, the story is yet to fully unfold⤵️
Asian Paints grew to dominate the industry with a 50%+ market share because it did a few things right:
(1) Installed Tinting Machines at distributor's shops
Asian Paints discovered early that installation of tinting machines helps increase throughput per dealer by 3x; they were the 1st to install these machines at distribution points
(2) Repeatedly won over hearts & minds with Marketing
From Gattu (1954) with cartoonist RK Lakshman to “Don’t lose your temper, use tractor distemper” to the legendary 2002 TVC :Har Ghar Kuch Kehta Hai” - Asian Paints dominated share of mind
(3) Enforced Working Capital discipline across its distribution network via the Regular payment performance discount (RPPD)
Asian Paints gave discounts for early re-payments by distributors; 3.5% rebate for 30 day payment & 5% rebate for 3 day payment. In a low margin industry (< 5%), this was a huge incentive!
Birla Opus took on Asian Paints by using the company’s playbook against itself:
Birla Opus has done the following:
Scaling from 0 to 50K distributors in 2 years by piggybacking on Grasim's 200K+ cement distribution points; and:
(a) Released a next-gen compact tinting machines (40% smaller footprint); which is where the legal case comes from - they believe Asian Paints is forcing distributors to NOT take the Birla machines
(b) Enabled 4-hour deliveries for 1,200+ SKUs; this same day delivery promise was a key driver of Asian Paints growth & NPS.
(c) Executed an ALL OUT price war which I had written about approx. a month ago (bookmark for future):
2007: HYD - Google’s 1st office in India
2010: HYD - Facebook’s 1st office in India
2015: BLR - Twitter’s 1st office in India
OpenAI’s 1st India hire is a Public Policy head and their 1st India office opening in New Delhi signals one thing:
Any global entity serious about AI in India will have to approach Public Policy / Govt relations in a formal manner.
Sam Altman has led on this front:
(a) He came to India in Feb ‘25 to discuss “OpenAI for Countries” with the Govt of India.
(b) In May ‘25 - OpenAI enabled local data residency for Indian ChatGPT users.
(c) Last week, OpenAI launched India focused pricing with ChatGPT Go
Unlike social media & ride hailing, AI has an upfront & clear public policy implication - no surprises here for the Govt; they're taking global & sovereign AI very seriously (e.g. the ₹10,300 crore IndiaAI Mission budget)
Today, OpenAI announced its 1st office in India will open in New Delhi;
Sure, you can argue that Google, FB & Twitter launched Indian offices in engineering hubs whereas Delhi is our capital city.
The wider NCR belt also has several Tech companies (Paytm, Policybazaar, Eternal etc) and the global MNCs have offices in NCR (e.g. Microsoft) - so this isn’t just a politically motivated move.
OpenAI is going down the same path as Google, Meta etc of localization - pricing is done, team will be built - it will be VERY interesting to see what product changes now happen given they are serious about India