AI crash risk, a US-Russia diesel deal, Black Sea drones and a split global economy

Broadcom's boss calling AI investors 'casino gamblers' is a striking thing for a chip-company chief executive to say, but the IMF's concern is more structural. The fund is worried about asset price bubbles: situations where the price of something rises far above what the underlying earnings or cash flows can justify. When a bubble is funded by borrowed money, a price correction does not just hurt the original buyers. It forces them to sell other assets to repay loans, which pushes prices down across markets, which triggers more selling. That chain is called contagion, and it is why a crash in one sector can hit pension funds, banks and bond markets that never owned a single AI share.

The labour-market warning is a separate but connected argument. The IMF describes AI as a 'slow-moving tsunami' for jobs, meaning the disruption builds gradually rather than arriving all at once. Young workers entering the job market now face the most exposure because the tasks AI can automate most easily, routine data processing, basic writing, pattern recognition, are exactly the entry-level tasks that used to give new workers their first foothold. If those entry points disappear, the long-run effect on wages and inequality could be large. Whether AI productivity gains eventually create enough new jobs to compensate is the genuinely open question, and one month of market data is nowhere near enough to answer it.

What I find odd is that both risks, the financial one and the labour one, point in the same direction for policy but pull governments in opposite directions politically. Regulating AI investment more tightly might reduce the bubble risk, but it would also slow the productivity gains that could offset the jobs problem. That tension is worth keeping in mind for any Paper 2 question on the role of government in managing technological change.

Sanctions on energy exports are a price-cap instrument: they do not stop the seller from selling, but they try to limit how much the seller earns by restricting who can buy. The logic is that fewer buyers means the seller must accept a lower price or leave oil in the ground. The US-Russia diesel deal chips away at that logic by adding a large new buyer, the US market, back into the pool. More buyers means Russia can charge closer to the world price again, which reduces the financial pressure the sanctions were designed to create.

For India, the deal cuts both ways. Indian refiners have been buying discounted Russian crude since 2022, precisely because Western sanctions pushed Russia to offer lower prices to whoever would buy. If Russia can now sell more freely to Western markets, the discount India enjoys may narrow. Indian refiners could save less per barrel than they have over the past two years. The rupee's recovery, which has been fragile, depends partly on a manageable import bill, so a narrowing Russian discount is one of the real risks the rupee faces in the weeks ahead. Whether the discount disappears entirely depends on how much Russian supply actually reaches the US market and how quickly.

The IMF's 'K-shaped' framing maps directly onto the idea of divergent growth paths, where two groups of countries face structurally different conditions rather than a shared cycle. The mechanism driving the split is two-sided. On one side, countries with strong tech sectors or low energy import dependence are capturing the productivity gains from AI and facing less inflation from the oil shock. On the other side, energy-importing developing economies are paying more for oil, borrowing at higher rates because global interest rates are still elevated, and seeing their currencies weaken as capital flows toward higher-return markets.

India sits in an interesting position in this split. The 7% growth forecast puts it firmly in the faster-growing group, and its large IT sector means it captures some AI-related demand. But India is also a heavy oil importer, so the energy side of the split works against it. The rupee has been under pressure from capital outflows and the oil bill simultaneously. That combination, strong growth but a vulnerable currency, is exactly the kind of tension the RBI has been managing all year. Whether India ends up in the upper arm or the lower arm of the K depends largely on where oil prices settle and whether foreign investors keep treating Indian assets as worth the currency risk.

Attacks on merchant shipping raise what insurers call war-risk premiums: the extra cost added to a ship's insurance policy when it enters a zone where military action is possible. When premiums rise, shipping companies either pass the cost on to buyers of the goods, or they avoid the route entirely. Either way, the goods become more expensive or harder to get. The Black Sea route matters for European grain and steel imports, so sustained attacks do not stay as a regional military story for long. They become a supply-chain and inflation story for European consumers.

The precedent that comes to mind is the Red Sea disruption of 2024 and 2025, when Houthi attacks on shipping pushed container costs sharply higher and rerouted vessels around the Cape of Good Hope. The parallel holds in one important way: uncertainty about who fired the drones today makes the insurance calculation even harder, because underwriters cannot model a risk they cannot attribute. Where it breaks down is geography. The Black Sea is a semi-enclosed sea with limited alternative routes, so rerouting is harder than in the Red Sea. If attacks continue, the pressure on European food and commodity prices could build faster than it did during the Red Sea episode.

Hammack's concern is about inflation expectations: what people and businesses believe prices will do in the future. This matters because expectations are self-fulfilling. If workers expect prices to keep rising, they demand higher wages. If firms expect higher wages, they raise prices to protect their margins. Once that cycle starts, it is very hard to stop without a sharp rise in unemployment. The Fed's job, as Hammack sees it, is to break that cycle before it starts, even if it means keeping rates higher than the economy would otherwise need.

The tricky part is that the evidence for an 'inflationary mindset' taking hold is not yet clear-cut. One or two months of sticky inflation data is not the same as a wage-price spiral. The obvious pushback to Hammack's position is that keeping rates high for too long risks tipping the economy into a recession that does more damage than the inflation it was meant to prevent. She is half right: the risk of entrenched inflation is real, especially with oil prices volatile. But the IMF's own data today showed the global economy already under pressure, which makes the cost of a policy mistake on the tight side higher than it was a year ago. Which risk is bigger is a genuinely live debate, and the answer depends on data that does not exist yet.