ISRAEL

Six hostages are found dead in a tunnel, and the country turns its fury on its own government

The bodies of six captives, killed shortly before soldiers reached them, brought the largest protests of the war and a general strike. For many Israelis, the grief curdled into rage not at Hamas but at a prime minister they blame for leaving the hostages to die.

The Israeli military recovered the bodies of six hostages from a tunnel in southern Gaza on September 1, killed by their captors, forensic evidence suggested, only shortly before the soldiers arrived. The deaths, of young people whose faces had become familiar over eleven months of captivity, including a 23-year-old American, broke something in Israel's public. Hundreds of thousands poured into the streets in the largest demonstrations since the war began, and the country's main labour federation called a general strike that briefly shut the economy, all of it directed not at Hamas but at Prime Minister Netanyahu, whom the protesters accused of sabotaging a ceasefire-and-hostage deal to preserve his own political survival. The grief had found a target closer to home. Families of the hostages, who had spent eleven months pleading for a deal, now openly charged their own government with choosing the war over their children's lives. Netanyahu, insisting that military pressure was the only path to freeing the rest, refused to yield on the terms a deal would require, and the war, and the captivity of those still held, ground on.

MARKETS

Nvidia loses $279 billion in a day, and the market asks whether the AI boom is a bubble

Nvidia, the company the entire artificial-intelligence boom runs through, lost about $279 billion of its stock-market value in a single day on September 3, the largest one-day loss in the history of any company, even after reporting results that would once have been called spectacular. The fall crystallised a doubt that had been building all summer: that the market's euphoria over AI had outrun the reality, that the hundreds of billions of dollars being poured into data centres and chips had not yet produced the profits to justify them, and that the whole edifice might be, if not a bubble, then at least priced for a perfection it had not delivered. Nvidia's business is booming by any ordinary measure; the question the market is now asking is whether "booming" is enough when the price assumes "miraculous." The answer will shape not just one stock but the trillions of dollars, and the vast physical build-out, that the AI story has set in motion.

IN BRIEF

Brazil switches off X; a school shooting; the iPhone gets its brain; a Pope in the islands

Brazil blocked the social network X across the country on August 30, after its owner Elon Musk defied a judge's orders to remove accounts and appoint a legal representative, a dramatic assertion that even a billionaire's platform must obey a nation's courts. A 14-year-old killed four people at a high school in Georgia on September 4, and in a new turn, his father was charged too, for buying him the gun. Apple prepared to unveil the iPhone 16, built around the "Apple Intelligence" features that will put generative AI in a billion pockets. And Pope Francis, 87 and frail, undertook the longest journey of his papacy, through Indonesia, Papua New Guinea, Timor-Leste and Singapore, drawing enormous crowds in the world's Catholic peripheries.

The Column

Booming is not the same as priced for a miracle

Nvidia lost a quarter of a trillion dollars in a day after excellent results. That paradox is the whole AI-bubble question, and I want to separate the two things everyone keeps confusing: whether the technology is real, and whether the price is.

By Rebecca Cho

Nvidia this fortnight reported results that were, by any normal standard, magnificent, revenue up enormously, demand for its AI chips insatiable, and its stock fell so hard it erased more value in a day than any company ever has. To a lot of people that looks irrational, and the temptation is to conclude either that the market is mad or that the AI story is collapsing. Both conclusions are wrong, and the reason they are wrong is worth understanding, because it cuts to the heart of the question everyone is now asking, whether the AI boom is a bubble, and the honest answer requires separating two things that the hype has deliberately fused: whether the technology is real, and whether the price is.

The technology is real. Let me say that plainly, because the bubble question tempts people into a lazy cynicism that dismisses the whole thing as hype, and that is not what I am arguing. The AI systems are genuinely capable and getting more so; the demand for the chips that run them is genuine; Nvidia's business is genuinely booming. If the question is "is this like the crypto froth, an empty thing dressed as a revolution," the answer is no. Something real is happening. But, and this is the entire point, a technology being real and transformative does not tell you anything about whether a particular price is justified, because the price is not a bet on whether the technology is real. It is a bet on a specific, enormous, quantified future, and that future can fail to arrive even if the technology is everything its believers say.

Here is the distinction that matters. Nvidia's valuation, and the trillions riding on the AI story, do not merely assume that AI is real and useful. They assume a specific, staggering trajectory: that the hundreds of billions being spent on data centres and chips right now will translate, soon, into hundreds of billions in profits at the companies buying them, that the spending is the early phase of a return that justifies it. And that is the part that is not yet proven. The chips are selling; the profits from actually using them, at the businesses pouring money into this, are so far mostly a promise. The market has priced in not just that AI is real but that the return on this vast investment will be swift and huge, and this fortnight, for a day, it got nervous that the return might be slower or smaller than the price demands. That nervousness is not a judgment that AI is fake. It is a judgment that "real and transformative" and "worth exactly this price today" are different claims, and that the gap between them is where a lot of money can be lost even if every optimistic thing about the technology turns out to be true.

So is it a bubble? Here is my honest answer, and it is deliberately unsatisfying. It is not a bubble in the sense that the underlying thing is worthless; it plainly is not. It may well be a bubble in the narrower, more dangerous sense: that the prices assume a speed and scale of return that the actual economics have not yet delivered and might not deliver on the assumed schedule, and that when a gap opens between the promised future and the arriving one, a great deal of value can evaporate fast, as it did for a day this fortnight. The dot-com crash is the cautionary tale precisely because the optimists were right about the internet and still lost everything, because being right about the technology and wrong about the timing and the price is a way to be ruined by a revolution you correctly predicted. AI is real. The build-out is real. Whether it is worth what is being paid for it, on the timeline being assumed, is a completely separate question, and this fortnight the market remembered, for a day, that it did not actually know the answer.

The Ledger

Nvidia's record one-day loss crystallises the bubble question. The technology is real; whether the prices, which assume a swift and vast return on the build-out, are justified is unproven. Watch whether the companies spending billions on AI start showing the profits to match.

Data centers & power

The bubble question and the power question are the same question from two angles: the vast data-centre build-out is a bet that the returns will justify not just the money but the enormous, grid-straining energy it demands. If the returns lag, the power was still spent.

Rates

A September cut is locked in after Powell's Jackson Hole signal; the debate is 25 versus 50 basis points. The turn of the rate cycle, awaited all year, arrives next fortnight.

Real estate

US 30-year mortgage 6.35 percent (September 5), the lowest in over a year, as the cut nears. The frozen market is finally, tentatively, thawing.

India tech

The AI-bubble question matters acutely for India's services sector, which has staked its next decade on building for these models. If the boom cools, the firms selling reliable, real, profitable AI implementation will outlast those riding the hype.

What we called wrong

Nothing to retract. This paper has tracked the AI build-out's disconnect from its power and profit foundations for over a year; Nvidia's wobble is the market beginning to ask the questions this Ledger has been quietly posing since No. 12.

The Back Page

The old man in the far islands

While the powerful world argued about wars and elections and stock prices, an 87-year-old man in a wheelchair spent this fortnight travelling to the ends of the Earth. Pope Francis, frail, in visible pain, undertook the longest and most demanding journey of his papacy, nearly a month and tens of thousands of kilometres, through Indonesia, Papua New Guinea, Timor-Leste and Singapore, the peripheries of the Catholic world, the small and poor and far-flung places that the centres of power never visit. In Timor-Leste, one of the poorest countries on Earth, nearly half the entire population, some 600,000 people, came to see him say mass. He went not to the capitals of influence but to the margins, to the jungle airstrips and the island churches, to people who would otherwise never in their lives be visited by anyone the world considers important. There is a message in the sheer geography of the trip, whatever you believe: that the man had decided, in his physical decline, to spend what strength he had left going to the people the powerful forget, the ones with nothing to offer him, the far islands. He is old and unwell and will not make many more such journeys, and he chose to make this one to the ends of the Earth, to be, for a moment, with the people at the very edge of everything. In a fortnight of the mighty grasping for power, an old man went as far as he could go to sit with those who have none. It is worth noting, whatever your faith or lack of it, which direction he chose to travel.