Why the Next Decade Isn't Just About India ft. Saurabh Mukherjea

ELI5/TLDR

Saurabh Mukherjea’s pitch is that every Indian investing decade has had one defining theme — real estate (2004-14), domestic equities (2014-24) — and the next one is global diversification. The reasoning: 70% of an affluent Indian’s spending is effectively in dollars (iPhones, flights, foreign degrees, holidays), the rupee gives up roughly 40% to the dollar each decade, and Western mid-caps have spent 25 years de-rating while sitting on the biggest CapEx boom since the 1960s. His suggested split for an affluent portfolio: a third India, a third global equities, a third precious metals — routed through GIFT City mutual funds for tax efficiency rather than direct brokerage accounts.

The Full Story

The decadal pattern

Mukherjea’s frame is mechanical. Two ingredients have to line up for an asset class to have its decade: valuations get pounded long enough to be cheap, and a new financial need shows up in the middle class. Mumbai real estate halved between 1995 and 2003, then the IT-services and private-bank boom produced a salaried middle class that needed homes and could finally get loans. 2004 to 2014: even amateurs made 20% CAGR. Indian equities went sideways from 2007 to 2013, valuations got attractive, the NDA arrived in 2014, and the Nifty printed ten consecutive positive years.

Now he claims the same pattern is loading for global equities. Need plus valuation. The need is dollar liabilities masquerading as rupee ones — your iPhone bought in Mumbai is a dollar item, your domestic flight is a dollar item, your YouTube subscription is a dollar item. The valuation case is that US and European small and mid-caps have been de-rating relative to mega-caps for 25 years. He’s finding companies in the West at 15 PE growing earnings at 15% — PEG under one, which in the Indian context is unicorn territory.

Why not just stay in India

He doesn’t deny India will keep doing well. The arithmetic is just that the country generates close to a trillion dollars of household savings a year, and Indian large caps (which are 70% of the market) can probably absorb only $20-30 billion of that flow at current valuations. The rest leaks into gold, real estate (also looking punchy), and — increasingly — abroad. He estimates $100 billion plus will leave India for global equities over the next decade. BlackRock has already shown up in India for exactly this reason.

The diversification argument runs alongside. India and the US are the only two markets that have generated double-digit dollar returns over 10, 20, and 30-year horizons. They have low correlation with each other. Build a portfolio half in each and one of the two is usually firing. Outside Lehman 2008 and COVID 2020, the two markets have rarely been hammered in unison.

What the West is actually doing

The bit Mukherjea admits he didn’t see coming when Marcellus started investing globally in October 2022 is the CapEx boom — roughly $2 trillion flowing into aerospace, defence, and power. Airbus has a 10-year order book. GE Aerospace has 7-8 years. ChatGPT shifted global power demand growth from 2% to 4% per year, which works out to 500 gigawatts of new capacity needing $700 billion of CapEx annually. Alstom and Siemens have 7-8 year order books on the back of it.

His punchline: less than 20% of the biggest share-price generators in America over the last decade have been tech companies. The mental model of “America equals seven tech stocks” is wrong. The action has been heavy industrials, defence, healthcare, ultra-luxury — and crucially these are not trillion-dollar names.

What the Marcellus Global Compounders book looks like

He spends some time naming holdings, which is useful because the abstraction “global equities” obscures what the portfolio actually does:

20% tech (mix of European and American — ASML in Holland, plus US names)
20% aerospace and defence (Airbus, GE Aerospace, Safran, Raytheon, Lockheed Martin)
15-20% industrials (Sweden, Germany, US — for the power and grid build-out)
10% financial services (Berkshire, listed PE)
15-20% ultra-luxury (mostly Paris and Milan — French and Italian houses dominate the handbag and shoe trade)

Only 20% of the book is mega-cap. The rest is 5-100 billion dollar companies. The reasoning: if you want mega-cap, you can buy a Nasdaq ETF on your phone. You don’t need an active manager for that.

The plumbing — GIFT City vs Interactive Brokers

The tax point is the most actionable thing in the conversation. Mukherjea is explicit that direct routes — opening an Interactive Brokers account, buying US stocks through an Indian brokerage app — are a bad idea now, even though they were the only option a year ago. Two reasons. Short-term capital gains get taxed at the maximum marginal rate. And US estate tax is brutal: if you die holding US stocks directly, your heirs lose roughly half. His phrasing: “When you’re alive, the Indian government will tax you. When you die, the Americans will tax you.”

The cleaner route is mutual funds domiciled in GIFT City and regulated by IFSC, which then invest into global stocks. Marcellus is launching one. There are several other providers. Minimums are small ($5,000 in their case). Capital gains treatment is similar to Indian mutual funds, and the US estate-tax problem doesn’t apply because you don’t hold the underlying.

For body corporates there’s also the Overseas Portfolio Investment regime — any private/public limited company or LLP can park up to 50% of its net worth abroad through GIFT. Mukherjea mentions HDFC and other Indian banks now offer 4.5% dollar fixed deposits via their GIFT City branches, which combined with rupee depreciation effectively yields something like 9% in rupee terms. Useful for family offices and treasury cash.

The China question

He’s not buying China proper, only Taiwan via TSMC. Three reasons. Despite faster GDP growth, Chinese equities have given single-digit dollar returns over decadal cycles versus 10-15% for India and the US — Chinese companies are capital-intensive, cash-poor, and dilute shareholders constantly to fund growth. The Communist Party controls everything, including stock prices (the Jack Ma episode being the textbook example). And the market is on-off — two years of returns followed by two years of disappointment, with patchy governance underneath.

The AI book — the part worth marking

This is the most interesting digression. Marcellus has a book coming out 28 March 2026 arguing that AI is a bigger disruption for India than for the West. The historical pattern: the first industrial revolution destroyed jobs for 90 years before creating new ones (1750 power loom to 1850). The second took 30 years (1880 electricity to 1910 Hollywood). India sat out both under colonial rule. This time it lands directly on what India does — IT services, banking, media, retail, law — the white-collar professions that absorb 80 lakh graduates a year.

For Sweden, AI is heaven because they’re labour-constrained. For India, IT services have effectively stopped recruiting, and Mukherjea expects the same to spread to financial services, retail, and media. His three predictions for how India has to pivot: a more competitive currency (he reckons closer to 100 to the dollar), a manufactured-export-led model rather than a consumption-led one, and lower cost of capital and land to make manufacturing competitive. The next 2-3 years are the pivot.

Allocation

Final number. A third India, a third global equities, a third precious metals. He hedges on gold given current prices — hasn’t put a third in himself. And he concedes most listeners will find a third in global equities a stretch and should start at 15-20%. Worth noting his framing: India is 3% of the global market, so even at one-third of your portfolio you are massively overweight India. At one-third global, you’re still underweight global.

Key Takeaways

Decadal themes rotate when valuations and a new middle-class need converge. He thinks the third decade (2024-34) is global diversification.
70% of affluent Indian spending is effectively in dollars; the rupee gives up ~40% to the dollar each decade.
Western small/mid-caps have de-rated for 25 years and are sitting on the biggest CapEx boom since the 1960s — defence, aerospace, power.
Less than 20% of the biggest US share-price generators of the last decade were tech. The rest is industrials, healthcare, luxury.
GIFT City mutual funds beat Interactive Brokers for tax efficiency — direct US holdings expose heirs to ~50% US estate tax.
For corporates, the OPI regime allows up to 50% of net worth abroad; HDFC and others offer 4.5% dollar FDs via GIFT City.
Avoid China proper — capital-intensive, cash-poor, Party-controlled, governance patchy. TSMC fine.
AI book (March 2026): India is uniquely exposed because the disruption hits white-collar services where 80 lakh graduates land each year. Pivot needed: weaker rupee, manufactured exports, cheaper land and capital.
Suggested allocation: 1/3 India, 1/3 global, 1/3 gold. Start at 15-20% global if one-third feels like too much.

Claude’s Take

Mukherjea is a clear thinker with a self-serving thesis — Marcellus is launching a GIFT City fund, so a podcast about why every affluent Indian needs global exposure via GIFT City funds is, let’s say, on-message. That doesn’t make him wrong, but it’s worth flagging. The decadal framework is a retrospective fit dressed up as a forecast. Real estate had its decade because of one specific liquidity unlock (private banks, home loans) and one specific demographic (the IT middle class). Equities had theirs partly because of those same forces continuing, plus a once-in-a-generation political mandate. Calling global equities the next decadal theme requires assuming Western mid-cap valuations stay cheap and the CapEx boom keeps running — both of which are exactly the consensus he’s selling.

The strongest argument is the boring tax-and-estate point. If you’re going to hold meaningful global exposure as an Indian, GIFT City genuinely does look like the right wrapper, and the math on US estate tax exposure is real and underappreciated. The weakest argument is the rupee depreciation point treated as iron law — 40% per decade is roughly the historical average but not a guarantee, and the rupee can sit flat for years. The most interesting thing in the whole interview isn’t the global diversification thesis — it’s the AI side-comment, where his point about India being uniquely exposed because the disruption hits exactly the jobs that absorb 80 lakh graduates a year is sharper and more uncomfortable than anything else he says. The book on it is worth waiting for. Score 7 — useful frame, sound plumbing advice, but partly a sales pitch and the headline claim is more of a directional bet than a derived conclusion.