Original Lens and publication context
Four things happened abroad last week. Only two of them will show up in Indian portfolios — and the biggest one is oil.
- Four things happened abroad: a record chip result was treated as bad news, Korean funds using borrowed money were forced to sell, oil jumped seven percent on Gulf attacks, and the US Federal Reserve held rates while long-term US yields hit their highest level since 2007.
- For India, oil is the one that matters most. We import roughly eighty-five percent of the crude we use, so every ten dollars on the barrel widens the trade gap, presses on the rupee, and reaches inflation with a lag.
- The second is the American thirty-year yield above 5.2 percent. When money earns that much safely in dollars, Indian assets have to work harder to attract it — and the RBI has less room to cut than domestic inflation alone would suggest.
- The selling in AI shares is funds shrinking their positions, not a verdict on demand. It is not a signal about Indian IT services yet; if it becomes one, it will arrive through client budgets two or three quarters later.
- For an India-based portfolio: stay with domestic-demand sectors, prefer refiners to oil producers, keep gold and some dollar exposure as insurance, and do not buy the global AI dip until the borrowing cost of the big AI spenders stops rising.
I. The Four Things That Happened
- A large memory-chip maker reported the best results in its history. Share prices fell anyway. Investors are no longer asking how much these companies earn. They are asking whether the growth is speeding up or slowing down.
- Korean shares crashed and stopped only at their 200-day average price. The cause was not weaker Korean profits. Funds that had borrowed to buy were forced to sell. Their borrowed exposure had reached about 3.5 percent of all freely traded Korean shares in June and has since fallen to roughly 1.5 percent.
- Oil rose about seven percent, above eighty-five dollars a barrel, after roughly thirty drone attacks over three days on a US base in Jordan and on Saudi energy sites. All were shot down. American and Saudi forces struck back at Iran-linked targets in Iraq.
- The US Federal Reserve held its rate at 3.50 to 3.75 percent. Three of its own officials wanted an increase. Yet the thirty-year US yield still rose to 5.21 percent, the highest since 2007, while the two-year fell.
Only two of these four reach an Indian portfolio in any meaningful way. Ranking them correctly is the whole job.
II. How It Reaches India
1. Oil is the main channel, and it is not close
India buys roughly eighty-five percent of the crude it consumes. That makes the oil price a direct tax on the Indian economy, paid in dollars.
Two consequences. Fuel-consuming sectors — paints, tyres, aviation, logistics, packaging — face a margin squeeze that appears in the next two quarterly results, not in the next headline. Indian refiners sit on the other side: the global profit from turning crude into fuel is near sixty-five dollars a barrel, close to the widest in a year.
The oil move was not a new shortage. It was the removal of hope that talks would hold. That widens the range of bad outcomes without changing the central case.
Figure 01 · Conceptual framework
From an oil shock to an Indian balance sheet
2. The American thirty-year yield sets the price of patience
This is the channel most Indian investors underweight, and it is the more durable of the two.
This is not a crash mechanism. It is a slow drag on the valuation Indian equities can sustain, and it falls hardest on sectors priced off profits that arrive many years from now.
3. The AI selling is not yet an Indian IT story
Last week’s selling was mostly global hedge funds cutting the size of their books — selling what they owned and buying back what they had bet against. It was the largest such episode since November 2022. That tells you who owns the risk today, not what demand will be.
The split inside AI still matters: software and platform companies rose while chipmakers, data-centre builders and networking companies fell. The market has stopped asking whether AI works and started asking who pays for the chips, the buildings and the power — all bought with borrowed money that has become dearer. For Indian IT the read-across is indirect. Our firms sell people and systems, not chips, but client budgets are set by finance officers watching their own borrowing costs. That reaches Indian deal flow in two to three quarters, not next week.
4. Korea is a warning about mechanics, not about Asia
The forced selling is now largely finished, which removes the forced seller — and the forced buyer on any bounce. The lesson is about market structure, not about Asia: wherever borrowed money builds up quietly, its unwind sets the price for a few days regardless of fundamentals. That applies here too.
III. Where It Could Bite
- Oil going higher, not staying here. The escalation risk is live and the market has stopped pricing a diplomatic outcome. An actual hit on export infrastructure is the scenario that hurts the rupee, the deficit and inflation together.
- Lenders disagreeing with shareholders abroad. The cost of borrowing for the biggest AI spenders widened through July to the highest of the year, while their share prices merely dipped and bounced. Over horizons longer than a quarter, the lenders have usually been right. Watch that borrowing cost as the early warning for global risk appetite, and therefore for foreign flows into India.
- Assuming the sellers are finished. Fast-rising global shares fell about twenty-seven percent in six days. Desks argue the selling is nearly over; across eleven similar episodes since 1980, such falls often continued another month or two. We do not know which applies, so we size for not knowing rather than pick.
IV. What We Would Do
All views below are at the level of sectors and instruments. No individual company is named or recommended.
- Stay with domestic-demand sectors over globally-linked ones. Domestic consumption, lending to households and small businesses, and infrastructure spending are driven by Indian conditions, not by American borrowing costs.
- Prefer refiners to oil producers. Refining margins near their yearly wides are the cleanest way to hold energy exposure while crude carries war risk in both directions.
- Be selective in oil-sensitive manufacturing. Paints, tyres, aviation, logistics and packaging carry the input-cost squeeze, and it shows in results rather than headlines.
- Keep gold and some dollar-linked exposure, sized as insurance rather than as bets. In rupee terms gold pays twice — once if the metal rises on escalation, again if the rupee weakens on the import bill.
- Do not buy the global AI dip yet. Wait for the borrowing cost of the large AI spenders to stop making new highs. That, not the share price, is the signal that the financing question has been answered.
- Treat Indian IT as a two-to-three-quarter question, not a one-week one. The real test is client budgets in the next two reporting cycles.
- Expect less RBI room than domestic inflation implies. Build a portfolio that does not need a rate cut to work.
The short version: the argument abroad is about the cost of money and who is carrying it, not about whether technology works. For India it arrives as an oil bill and a higher global discount rate. Both are slow, both are real, and neither changes the Indian growth story.
Horizon: three months. Review date: 30 September 2026.
