India Markets Jump After Nvidia CEO's Bold AI Software Call | Govindraj Ethiraj

ELI5 / TLDR

Indian stocks started the day down, then flipped green after Nvidia’s CEO said AI will grow the software business, not kill it. That sent IT stocks like TCS and Infosys up more than 5%. The same episode also covers a quieter worry — IBM’s CEO thinks the world is building roughly twice as many AI data centers as the revenue can justify. Closer to home, rising fuel prices threaten to nudge inflation up, foreign investors have pulled money out of India to a near-decade low, and jewelry shops are quietly becoming one of the fastest-growing categories in Indian retail.

The Full Story

The two-sided AI debate driving the markets

The whole episode pivots on a single contradiction at the heart of the AI trade: massive enthusiasm sitting right next to a quiet fear of overbuilding.

On the enthusiasm side, three things happened in one week. Alphabet (Google’s parent) is trying to raise $80 billion for AI data centers and compute. Anthropic — the company behind Claude — confidentially filed to go public, meaning AI labs are graduating from venture-funded startups to public companies. And Nvidia is pushing its chips beyond data centers into ordinary PCs, betting that AI will increasingly run on your device rather than only in the cloud.

The fear side came from IBM’s CEO Arvind Krishna, who walked through the arithmetic carefully. Think of a “gigawatt” as a unit of data-center scale — one gigawatt of power needs roughly $60–80 billion of chips to fill it. Companies have committed to over 100 gigawatts.

“If you say that that’s got a five to seven-year payback, you are going to need an extra 1 to 2 trillion a year of revenue… So that much incremental revenue I don’t believe is there and so that’s why I think it’s a bit ahead.”

His second worry is subtler. He thinks the biggest AI models will become commodities — interchangeable, with low cost to switch between them. And a commodity with low switching costs can earn a margin, but not a protected one.

“Some will disappoint, many will thrive but not all will thrive.”

Why India’s market actually went up

So why did Indian stocks rally rather than worry? Because of one line from Nvidia’s Jensen Huang. The popular fear is that “agentic AI” — software that acts on its own rather than waiting for clicks — will wipe out software companies. Huang flipped it:

“It’s exactly the opposite because there are going to be so many agents. The world is no longer limited by the number of people. Therefore, those agents are going to use more tools than ever.”

More agents means more software being used, not less. That logic helped Indian IT — TCS, Infosys — jump over 5%, because technology is the third-heaviest sector in the Nifty50. The index, down earlier in the day, closed up 100 points; the Sensex up 382.

The less cheerful India backdrop

A few facts cut against the rally. South Korea just overtook India to become the world’s sixth-largest equity market — its chipmakers (Samsung, SK Hynix) rode the AI wave while India’s market cap slipped to $4.8 trillion. Foreign investors’ cumulative net equity bets in India have fallen to a near 10-year low after relentless selling — the lowest since 2016. The counterweight: actual company earnings have been steady, with Nifty50 profits up 6.6% year-on-year, ahead of a forecast 2%.

Fuel, inflation, and the gold shuffle

Rising fuel prices are the inflation worry. Crisil (“Crystal Ratings” in the transcript) estimates a 7–12 rupee per liter petrol/diesel hike adds about 36–48 basis points to consumer inflation. The mechanism is transport: freight is 54% of India’s logistics cost, road is 71% of freight, and fuel is 42% of road transport cost. So fuel prices ripple through almost everything.

Separately, the Reserve Bank appears to have sold about $12 billion of gold while buying $7.5 billion of foreign currency — a defensive reshuffle to cushion its reserves against fallout from the West Asia war. The rupee ended at 95.26 per dollar.

Jewelry, quietly, is eating retail

The episode’s long interview is with CBRE’s Anshuman Magazine on a retail shift. Jewelry’s share of organized retail leasing has roughly quadrupled — from about 2% in 2019 to 8% in 2025 — putting it among the top three demand drivers after apparel and food. The reasons:

Stores are getting bigger (8,000–14,000 sq ft) and more experiential.
The buyer is changing. Weddings used to dominate; now wedding jewelry is only about half, with the rest going to gifts, anniversaries, and casual occasions. The target is shifting from millennials to Gen Z, who want lighter, cheaper, more fashion-driven pieces.
High gold prices, a 15% import duty, and rising labor costs are pushing brands toward lab-grown diamonds and lighter products to keep the market affordable.

On broader retail, Magazine flagged two durable trends: retailers expanding into tier-2 and tier-3 cities (because the big metros are saturated), and digital-first brands opening physical stores — settling the old “will brick-and-mortar die?” question with a firm “no, you need both.” Food, beverage, and entertainment, he argued, hold up best in a downturn because options in most Indian cities are limited.

Key Takeaways

IBM’s Krishna estimates over 100 gigawatts of committed AI data-center buildout implies $6–8 trillion of spend, requiring $1–2 trillion of new annual revenue to justify — revenue he doubts exists yet.
His second thesis: frontier AI models will commoditize, leaving margins without a protective moat; only two or three model-builders may ultimately survive.
Jensen Huang’s bull case: more AI agents means more software tools consumed, expanding rather than shrinking the software market — the line that drove the Indian IT rally.
TCS and Infosys rose over 5%; tech is the third-largest sector weight in the Nifty50, so the rally lifted the whole index.
South Korea overtook India as the world’s sixth-largest equity market, powered by AI-linked chipmakers; India’s market cap slipped to $4.8 trillion.
Foreign portfolio investors’ cumulative net equity investment in India hit a near 10-year low (~₹730,000 crore, lowest since 2016).
Nifty50 net profit grew 6.6% YoY for the quarter ended March, beating a 2% forecast — the eighth straight quarter of single-digit growth.
A 7–12 rupee/liter fuel hike adds an estimated 36–48 bps to CPI; transport is the transmission channel (freight = 54% of logistics cost, fuel = 42% of road transport cost).
The RBI likely sold ~$12bn of gold and bought ~$7.5bn of FX in two weeks through May 22 to defend reserves against West Asia war fallout.
Jewelry’s share of organized retail leasing rose from ~2% (2019) to ~8% (2025); stores are larger, buyers younger, and occasions broader than weddings.
A new Producer Price Index (PPI) covering both goods and services launches June 15, aligning India’s inflation measurement with advanced economies.

Claude’s Take

This is a daily market brief, not a deep dive, and it does the brief job well — fast, sourced (Bloomberg, Reuters, Kotak, Crisil, CBRE), and honest about the contradictions. The smartest thing in it is the editorial decision to play Krishna’s bubble math right next to the Huang quote that moved Indian markets. The viewer gets the bull and bear of the AI trade in the same breath, which is rare for a market wrap.

The IBM segment is the genuinely useful five minutes — the gigawatt-to-revenue arithmetic is a clean mental model for sizing whether the AI capex wave is over its skis, and the “models become commodities” point is the kind of thing that’s obvious in hindsight but underweighted now. The jewelry interview runs long and gets repetitive, but the lease-share data (2% to 8%) is a concrete, verifiable signal about where Indian retail demand is actually shifting.

Scoring it a 6. The content is solid and well-sourced, but it’s daily ephemera by design — most of the specific numbers (index levels, rupee, daily oil moves) are stale within 48 hours. The durable value is the AI capex framework and the structural retail trend, both of which you could get more cleanly elsewhere. Note: the auto-transcript garbles names — “Crystal Ratings” is Crisil, “Anuman magazine” is Anshuman Magazine, “Jensen Wang” is Jensen Huang.

India Markets Jump After Nvidia CEO's Bold AI Software Call | Govindraj Ethiraj | The Core Report