At 10:00 this morning, Jio Platforms opened the largest share sale India has ever attempted. By 10:36 the book was full. Not "well subscribed". Full — 20.03 crore shares bid against 19.75 crore on offer, a take-up that took 36 minutes and beat Bajaj Housing Finance (2 p.m. on day one), LG Electronics India (end of day one) and Hyundai Motor India (18% at the close of day one) — while being larger than all three.
By noon the issue was 1.39 times covered, with the institutional portion at 4.10 times. The issue does not even close until 23 October. Allotment is due 26 October; trading starts 28 October.
I have waited a long time for this morning, and I want to be careful about what it proves. Two things are true at once, and only one of them is being celebrated on television.
First: the part I will allow myself
Since 2018 I have written about Jio not as a telecom company but as plumbing. In "Mukesh : the Enabler" I recorded Ambani's own warning that "data is the MOST IMPORTANT RESOURCE in this new world", and I argued that his real business was enabling millions of Indians rather than selling them minutes. In "Thus spoke Mukesh" I collected his insistence that data belongs to the individual and not to the corporation that stores it — the argument that later became DEPA and the account-aggregator architecture I wrote about in "If DEPA = Foundation, then SARAL = Superstructure". In May 2018, in "Reliance Jio : the Mighty Start Up", I described a company with the agility of a start-up and the balance sheet of a conglomerate.
And in December 2023, in "Akash reaching for the Sky", I quoted Ambani saying what the whole thesis rests on: that data has become "a new Factor of Production along with Talent and Money".
That sentence now has a market price, and it is the largest price India has ever put on anything: ₹30,213 crore of fresh capital, 27 crore new shares at ₹1,065–1,119, valuing Jio Platforms at up to $108.3 billion. The anchor book alone — ₹8,109.85 crore to 212 investors at the top of the band, with LIC taking the largest slice at ₹499.98 crore — drew bids of nearly ₹2.5 lakh crore, more than 30 times what was available. One person quoted by PTI put total anchor demand at about $26 billion. That is not enthusiasm. That is a queue.
So: the rails I said were being built have been priced. But let me state precisely what has been priced, because that is where the argument gets interesting.
Second: what is actually being sold is a pipe
Read the Reuters dispatch, not the ticker. In the year to March, Jio Platforms earned about $3.2 billion of profit on $15.5 billion of revenue — and roughly 90% of both figures came from the telecom business alone. It serves 533.3 million subscribers as of June 2026.
That is a magnificent business. It is also, by revenue, a pipe. Wires, spectrum, towers, a 5G core, a distribution machine that reaches a village. And we are buying it at $108.3 billion at the precise moment when, in Reuters' words, investors have "sharpened their scrutiny of AI-related companies" — a major test, as the wire put it, of global appetite for new share offerings in that climate.
Akash Ambani put the case better than any analyst on Tuesday, and I suspect he did not intend to prove my point: "There is no either-or at Jio: the network is our technology." Then, more precisely still: "Connectivity was the first product we built on that platform. Digital services, AI and cloud are what we are layering on top of it now."
Layering. On top of. That is the correct geometry of the thing. A brain needs a body. The question for a buyer of these shares is not whether Jio has a large language model. It is whether a rupee of capital put into connectivity compounds faster than a rupee put into a model lab. For the last decade in India, the answer has generally been yes — the returns came from connecting people, not from modelling them — and that answer is what is now listed.
The number nobody on television is reading
The subscription figure is a spectacle. The use of proceeds is the substance, and here the fine print is honest to the point of bluntness: ₹27,500 crore of the ₹30,213 crore goes to repay part of the borrowings of the subsidiary, Reliance Jio Infocomm. The balance is for general corporate purposes. Because this is a fresh issue — no offer-for-sale, nobody selling — every rupee lands inside the company.
I had expected, and half-hoped, to see the money earmarked for data centres, cloud and AI compute. It is not. The capital is being used to deleverage the pipe. PTI reports the company is focused on technology monetisation and is targeting debt-free status.
Is that a disappointment? On the contrary — it is the most reassuring line in the document. A company that raises ₹30,213 crore and spends it on its own balance sheet is telling you it intends to fund the next capex cycle out of a lower cost of capital. That is what utilities do. That is what infrastructure does. It is defensible, boring, and exactly the opposite of a story stock. The pipe is not pretending to be a brain. It is buying the right to become one later, cheaply.
| What was sold today | The number |
|---|---|
| Issue size | ₹30,213 crore, 27 crore fresh shares |
| Price band | ₹1,065–1,119 |
| Implied valuation | Up to $108.3 billion |
| Book fully subscribed | 10:36 IST (about 36 minutes after opening) |
| Anchor allotment | ₹8,109.85 crore to 212 investors; LIC largest at ₹499.98 crore |
| Anchor demand | More than ₹2.5 lakh crore, over 30x |
| Use of proceeds | ₹27,500 crore to repay Reliance Jio Infocomm borrowings; balance for general corporate purposes |
| Allotment / listing | 26 October / 28 October |
The "record" is a wholesale record
One more correction to the mood music. The queue was institutional. By 11:30 the QIB portion was 3.86 times covered while retail was at 0.33 times, non-institutional investors at 0.42 times, the employee reservation at 0.78 times and the Reliance shareholders' quota at 0.62 times. PTI noted that the institutional book crossed full subscription within the first ten minutes.
So India's biggest IPO was not a retail frenzy. It was a wholesale allocation decision by sovereign funds, insurers, pension money and mutual funds — GIC, Norges, ADIA, Qatar Investment Authority, PSP, CalSTRS, the Florida Retirement System, BlackRock, Capital Group, Fidelity, T. Rowe Price, alongside SBI MF, ICICI Prudential, HDFC MF, Aditya Birla Sun Life and Nippon. These are not momentum buyers. They are people who underwrite pipes for thirty years at a time and get paid for it. Retail will arrive by Friday, or it will arrive after listing. Either way, the price was set by patient money, and that is the more durable signal.
The Goldman note tells you where the next listings come from
Which brings me to the document I would frame above the subscription screen: Goldman Sachs' note of 19 October, reported today, which estimates that AI-led IPOs could add $50–100 billion to India's listed market capitalisation over two to five years, with 25–50% of nearly 70 mid-to-late-stage private AI companies potentially coming to market at a weighted average valuation of about four times their current estimate.
The composition matters far more than the headline. The largest segment by value is what Goldman calls "upstream AI enablers" — power generation and equipment, transmission, semiconductor supply chains, chip design, quantum, data centres, cloud and data management — at $35 billion of an estimated $55 billion across 84 "Private AI Contenders". Upstream is projected to supply 66% of the potential AI listings; downstream adopters 31%; core AI model builders a mere 3%.
Sixty-six per cent upstream. That is Jio's actual lane, and it is India's actual comparative advantage. We are not going to win the race to train the frontier model. We may well win the race to power, host and connect the world that runs it — because we already have the electricity problem, the fibre problem and the distribution problem, and because we have spent a decade building solutions to them at a price nobody else can match.
Which is why the honest verdict on today is not "India's AI moment". It is older and more useful than that: India has listed its digital plumbing before it has listed a single AI product. Given where the value in the AI buildout is actually accruing — power, transmission, silicon, data centres — that ordering may turn out to be correct rather than embarrassing. It is also, incidentally, the same insight that made the India Energy Stack interesting to me back in March: the rails are the market, and the applications are what you attach to them later.
The asterisk: today's rupee is a flow, not a trend
Now the part of the day's tape that argues against all of the above.
The rupee touched a six-week high around 94.5 to the dollar, up about 0.4% on the day, extending a recovery of more than 2% from levels that had threatened record lows earlier in October. Two things did that: foreign portfolio inflows and swap flows connected to this IPO, and a series of RBI measures last week — pushing oil-linked dollar demand away from the spot market, raising the cost of certain importer hedges — plus intervention. Foreign investors often sell dollars for rupees in the spot market ahead of an allotment and agree to buy them back later; when those positions unwind, the support reverses.
An IPO swap flow is a one-off. It arrives before allotment and leaves afterwards. It is not a structural surplus, and it does not fix a current account deficit — HSBC, per Reuters, expects the rupee to resume a "modest upward drift" against the dollar. Meanwhile the bond market is already voting the other way: the benchmark 6.94% 2036 paper yielded 7.4352% this morning, having touched 7.4397%, the highest since 9 March 2023. Brent is around $110. India imports nearly 90% of its crude. September CPI accelerated to 5.54% and is expected to breach the upper end of the RBI's band in October, with swap markets pricing about 100 basis points of hikes over the next twelve months, including a possible 50 basis points by December, after the first rate hike in four years on 7 October.
So: a $108.3 billion equity event, a six-week-high currency, and the worst bond yield in three and a half years — on the same morning. Anyone who tells you only one of those three is the story is selling something.
What I will watch on 28 October
- Whether the retail book fills by Friday. If it does not, the "record" belongs to institutions alone, and the stock's aftermarket will be thin and twitchy.
- The debt trajectory, not the listing pop. ₹27,500 crore of repayment against Reliance Jio Infocomm's borrowings is the actual transaction. Watch the interest line in the next two quarterly results.
- The revenue mix. The single number that would validate the $108.3 billion is the share of revenue that is not telecom. Everything else is commentary.
- Whether the next big Indian listing is another pipe. Goldman says 66% of the coming wave is upstream. If the market keeps paying up for enablers while discounting model builders, we will know which of the two this country is actually good at.
I have spent eight years saying that data is a factor of production and that someone was quietly building the rails for it. This morning the market agreed, in thirty-six minutes, at a price India has never paid before.
I would only add the caveat I have added all along: rails carry whatever you put on them. They do not, by themselves, decide what the cargo is worth. That is the next decade's work — and it will not be done by the listing bell.
No comments:
Post a Comment