Roaring trades: oil majors’ secret success story

0:00This episode is sponsored by HeyGen. HeyGen turns a script, photo, or presentation into a polished video of you in minutes, no camera or crew required. It's how real estate agents, financial advisors, attorneys, and creators show up consistently on every feed and build a brand that brings in clients without spending their whole week making content. Rated the number one AI video platform for small businesses on G2 with the most realistic AI avatars in the industry. trusted by a community of over 30 million people and 85 % of the Fortune 100 and built to work in over 175 languages so you can reach your audience in theirs.

0:33Here's the wild part. Record yourself for just 15 seconds and HeyGen builds an AI avatar that makes professional videos for you on demand so you can post everywhere your audience is, as yourself, without filming every time. Your first three videos are free at HeyGen.com slash pod. That's H-E-Y-G-E-N dot com slash P-O-D.

0:58The Economist

1:05Hello and welcome to The Intelligence from The Economist. I'm Jason Palmer.

1:14Today on the show, America's capricious grip on frontier AI models and the changing business of concert tours.

1:29First up, though.

1:38The world's largest energy firms, collectively known as The Majors, they've done really well through this oil shock. And one reason is that they've drilled oil and they've refined it. And they've sold it at a higher price, which is not surprising to most people. But another reason is far less known, which is that they've done a lot of trading. Mathieu Favre is our commodities editor. And the Iran war has really shown the extent to which this activity trading has become central to these businesses. Let's start with what you mean when you say trading. So trading can be defined in opposition to marketing.

2:15So marketing is selling what you produce, distributing it to the global markets. Trading is buying someone else's produce and selling it to whoever wants it most at the highest price possible, at a profit. And whereas when you sell your own products, what matters is the level of the price at which you sell. When you trade, what matters is the spread. It's the difference between the price at which you buy and the difference in price at which you sell. And the spread is typically higher when the market is quite chaotic. So if there is a war somewhere, there's a supply shock. then in certain places there will be shortages there will be also a rush for getting the product now instead of later and so you'll have all these differences in price that traders can seek to benefit from and so over time the majors have learned to do that really well especially the Europeans So which companies are we talking about then?

3:07Companies like BP and Shell which are mostly British Total Energy also known as TOTO which is French Total, no? Like Total And they trade by our estimate 40 to 50 million barrels per day of oil and gas. Relative to? It's five to 10 times more than what they produce. So it's huge. It just shows how central trading is to these businesses. At this moment or in general? In general. Yeah, in general. And probably during this very volatile period, there's been more because they've shifted more barrels. It's a very, very lucrative activity. The majors don't publish data for their trading desks on profitability or anything else, really.

3:45It's very secretive. But the estimates we've got from a number of sources suggest that these trio, BP Shell and Total, this year could earn in between 15 to 20 billion dollars in profit, which might be one fifth of the profits. And this might add up to a third to their return on capital. So something that shareholders will like a lot. Trading profits alone explain why the Europeans have outperformed the American majors so far this year, since the crisis in particular, since February. Why should that be? Why aren't the American majors just as good at this secretive game? It's largely a product of history and geology.

4:24So the American majors, they've always had access to vast resources in the ground. They have a lot of oil and gas. They also have a vast domestic market that they can sell to, so they can take advantage of both. The Europeans, they don't have that, neither of the things. So what they've had to do pretty much since the formation is to go abroad and produce the barrels there and then distribute it to the global market, which worked fine until the 1970s when there was a wave of nationalizations in the Gulf, in the Middle East, and they lost access to this oil. So they had to go abroad and buy barrels from someone else.

4:59So BP did it first, essentially discovered that it could buy the barrels. it didn't need and resell them at a profit in the 1980s. And then in the 1990s, Shell and Total did it then because oil prices were pretty low. They needed to boost their returns and they thought that was a way maybe to diversify. So that's a lot of the history of this trading arm for the matrix, but you say that they're really good at it. What does that look like? How do they do it well? The secret sauce of these trading operations is the intelligence, the information they can harness from the vast operations of the mothership.

5:38So the majors are huge companies. They've got lots and lots of oil and gas fields, refineries, terminals, storage facilities, tanker fleets all around the world. As the majors operate this big global network, there's a lot of information that surfaces about the amount of supply, the amount of demand there is in the world, and therefore the direction of prices. And so using that intelligence, they can harness volatility. So that's how they do it. And that's the reason why when I speak to sources, they say these operations, they basically never lose money. They make bad bets sometimes, but on an annual basis, they find ways to correct that.

6:17And so in worse years, they just make a bit less. In good years, they make a lot of money. This being a good year. It's been a pretty good year, I think, for them. Yeah, I think so. Yeah, yeah, really, really good. But it's got to require quite a few people to be chasing all of that information and compiling it into a bet. That's what's really interesting because these teams, they are at the same time large and pretty lean. So for each trading desk, for each product that they trade, so it would be like crude oil or it would be what we call light ends, which are petrol and jet fuel or LNG, liquefied natural gas.

6:51For each of these products, you maybe have like 60 to 70 traders. So if you add all the decks together for the whole trading operation, it's maybe like a few hundreds. And then if you add shipping, financing, risk management, all the sort of support functions, then it would be, you know, one or two thousand, which seems quite a lot. But at the same time, these companies, they have like a hundred thousand maybe people. So it's a really small percentage. So if you take the profit that they make and you divide by the number of people that they employ, it's a really high profit per head, you know, in a good year, maybe like$10 million.

7:28dollars. So there's a lot of incentive then for companies that are not the majors, European majors, and even the American ones, to get in on the game, if you say it is so lucrative. Yeah, absolutely. Or is it something that just only works at scale in this way? It only works at scale. So the Americans have tried before, the American majors, and it didn't work because they did it half-heartedly. They didn't give enough money, basically, to their traders. So that limited the scale of the bets. And also they did not give them enough independence they were forced to sell the barrels of the company basically but they're much more serious about it now they're much more determined to do it and so you're starting to see that in the way you speak about it and also the national oil companies which are the big state-owned oil producers which are in the big oil producing countries so in the gulf for example saudi aramco or adnok which is the abu dhabi national oil company they also want to do it So I spoke to one recruiter for this piece, and he said that the Americans and the national oil companies, Exxon in particular, and Adnok, are among the biggest recruiters at the moment of traders worldwide.

8:36So they're really quite keen to do it. To do it properly this time with the right number of people, the right number of dollars to spend. That's right. But even if they do that, to your first point, closing the gap may still take quite a few years because it really is quite a finely tuned machine, a well-oiled race car that you need to put together to make it work. You know, you prepare the car, you do a race during a crisis, you look at the daytime when it comes back, you make a few tweaks and then you do another race. But this could take many, many cycles before you get very good at it. That is to say that the European majors will stay on top for some time yet?

9:15I would say the European majors still have a few years of really lucrative trading. The golden goose is still there. But competition is coming. So in a few years' time, maybe profits will shrink. And the engineers, not the dealmakers, will be on the ascendant once again. Maybe they'll take back the building. Mathieu, thank you very much for your time. My pleasure.

10:07Thank you. with the most realistic AI avatars in the industry, trusted by a community of over 30 million people and 85 % of the Fortune 100, and built to work in over 175 languages, so you can reach your audience in theirs. Here's the wild part. Record yourself for just 15 seconds, and HeyGen builds an AI avatar that makes professional videos for you on demand, so you can post everywhere your audience is, as yourself, without filming every time. Your first three videos are free at heygen.com. slash pod. That's H-E-Y-G-E-N dot com slash P-O-D.

10:50The AI future is not going to be won by hand-wringing about safety. It will be won by building. Again and again, America's Vice President J.D. Vance and other Trump administration officials laid out the same ethos about AI. Cut red tape, get government out of the way, let America win. And I'm going to go see President Xi in two weeks. I look forward to that, but I'll say I'm leading. And then what might reasonably be called hand-wringing about safety. Anthropic pulled access to its newly released AI models, Mythos 5 and Fable 5, after the U.S. government restricted who could actually use them.

11:30The administration now wants the makers of the most powerful AI models to seek government review before new launches, in an executive order that sounds more suggestion than law. The government here is asking those companies to give over those models for review for safety reasons, it says. Putting that into practice has been, let's call it uneven. The effects are likely to get in the way of that whole letting America win idea. The Trump administration assumed office railing against the regulation of AI. Shashank Joshi is our incoming Washington bureau chief. What's happened in the last couple of months is a dramatic reversal, a stance in which leading AI models are now under de facto, in some cases, pretty draconian regulation.

12:24And what's behind that turnaround, that complete about-face? Well, fundamentally, Jason, it's that the models have got very good. You recall the release of Mythos, the model by Anthropic, that turned out to have extremely good cyber capabilities, including being able to find and exploit vulnerabilities in software, some of which had been undiscovered for years and years. And so the administration was pretty worried about this, as I think any government would have been and said, hang on a minute, you can't release this freely. We're going to release this on a very careful basis. There was a public version of the model called Fable 5 that was then put out.

13:02It was pretty carefully restricted, but there was an alleged way to get around some of its safeguards. And that prompted, in a kind of panic, the administration to say, you have to now shut down everything. We've also seen in recent weeks OpenAI, which makes the ChatGPT series of models, release another model called GPT 5.6 or SOL, a particularly capable variant of that. And they have not been allowed to freely release it. Basically, the government has given them permission to give it to a relatively modest number of trusted partners. So what's your view on this reversal? There are clear motivations behind it, a matter of safety, national security and so on.

13:45but there are these questions around innovation and what have you. Well, there was an executive order on June 2nd, and it was a way of trying to impose some sense of governance of AI. And it said, look, we should have a framework in which private companies can voluntarily share their most advanced models with the government and discuss the risks. That was kind of happening anyway, right? We know that OpenAI was doing that. We know that Anthropic was doing it. But the interesting thing about that order is that it explicitly says nothing should be a formal licensing regime. This is not a compulsory or mandatory system for companies to follow because there are a lot of people in the Trump administration who basically think this should be unshackled to maximize innovation at a time when investment in AI infrastructure like data centers is really holding up some parts of the American economy.

14:33Effectively, we have a de facto licensing regime in place that is described as voluntary. But to me, it looks about as voluntary as, you know, paying your taxes. You can choose not to do it, but there are going to be some consequences if you don't. But what about the model that they're pointing at here? Trusted partners figure out what harms are possible and then one presumes a wider release. There is a certain sense to it. There is a certain sense to it. You want to exploit this window of superiority in particular ways without giving access to the model to every potential bad actor who would use it for nasty things.

15:04The problem is, though, it's been such an opaque process. It hasn't been clear what the rules are, and it hasn't been clear who gets special access. So for example, if you look at GPT 5.6 Sol, the open AI model, on June 26th, that was opened up to trusted partners. And the same day, America's government eased some of the export controls on Mythos 5, which was the anthropic model that, as I said earlier, had been completely effectively shut down. So the models are now accessible again to some degree, but it's taken a while to get there. And nobody really understands the exact criteria by which these things are being held back, restricted, opened up again.

15:46And that makes life very, very difficult for the companies in understanding how their models are going to be out there in the world in the future. So the principle is sound, but the pace of change is such that the implementation is just really messy. Implementation is really messy. There's not that much in-house expertise on cutting-edge AI inside the American government at all. The Commerce Department has been leading the implementation of policy, but I don't think it has very many AI experts. And there are AI experts in something called the Center for AI Standards and Innovation, or Casey. That body is a standards body.

16:22It's not a regulation body. That executive order I mentioned, that is supposed to create a process to judge which models need to be controlled and how they should be controlled. And that should happen before August. But it really, it isn't clear to me how that's going. And a lot of people I speak to say that it's really a complete mess at this point. And a mess that seems to stand in stark contrast to what's going on in China, where the other best models are. Chinese AI labs are really storming ahead. Those models are six to 10 months behind their American equivalents. So like they're not up there with, you know, Mythos or with 5.6 Sol, but they are getting increasingly good.

17:03They offer increasingly advanced capabilities. And here's the important bit. They do that at pretty competitive prices. And so there is this concern among many in America that if the administration holds back the American labs from releasing their models, the companies, including American companies, might look to using Chinese open weights models. And so we've seen Microsoft, for example, consider should they use Deepsea for their co-pilot, which is their own AI tool. And I think that sense that this could give an opening to the Chinese, particularly when this administration has allowed the sale of some pretty advanced chips to China that then can be used to train the next generation of Chinese models.

17:47I think there's a lot of concern about that, including in the Trump administration. So essentially, America is at grave risk of exceeding its lead in this just because it's bound up in regulatory hell? I think there's a question of its lead, particularly in adoption. But I think it's not just about the competition versus China. There are huge economic stakes here. Almost $800 billion is going to be spent on AI capital expenditure this year. And most of that's on data centers in America. That will grow to$1.6 trillion annually by 2031. The American economy is really, really reliant on this technology and its promise.

18:24And if you look at what happened to the ban on Mythos, it reduced Anthropik's expected IPO value by about 10%. So you can kind of see this economic knock-on effect. If there is uncertainty around whether these companies will really be able to release their next models, or will have to wait for weeks to do so, that's eating into that window in which they can basically sell access to models and defray the enormous cost of developing this technology. And so the economic stakes here, Jason, are also absolutely enormous. Chishong, thanks as ever for joining us. Thanks so much, Jason.

19:30When Harry Styles sang those lyrics on One Direction's hit Midnight Memories, he was already very well accustomed to the glamour and grind of a global concert tour. Vicky Jessup writes about culture for The Economist. This is a man who has been in one direction and now is striking out on a very successful solo tour.

19:52Across his boy band and solo career, Harry has performed with more than 600 gigs on six continents. His previous tour visited 79 cities. This is a man who is used to being on the move. But to paraphrase Harry himself, all is not as it was.

20:15The 68 concerts on his new tour, Together, will be hosted in just seven cities, which include several quasi-residencies, including 12 Nights in London, 10 in Amsterdam, where he kicked things off, and 30 in New York City. He is currently in a 12-night run at Wembley Stadium in London, which ends on July 4th. And this string of performances at Wembley, which take place across June and July, means that he'll actually overtake Coldplay and Taylor Swift as the artist with the most performances at the venue in a single year. And that's before August, where he goes to Madison Square Garden to play on whopping 30 concerts.

20:57But it's not only Harry Styles who's travelling Lassbury's work. Many of the biggest pop acts have started to expect their fans to come to Dan. Take Beyoncé. The star spent half her Cowboy Carter tour in just three venues, including London's Tottenham Hotspur Stadium. Coldplay played extensively in certain cities. Oasis breezed through Wembley not once but twice. Olivia Rodrigo would do the same when she had that on the road in September. And Ariana Grande is only appearing in three countries on her tour, the US, Canada and the UK. There are a couple of reasons that musicians are actually travelling less.

21:39first of all touring is grueling these tours hit multiple cities sometimes night after night if you stay in fewer places for a longer period of time that's less exhausting for the artist and it's more lucrative touring is more expensive than it's ever been and fans expect elaborate sets these days transporting them across the world is not cheap so staying in one place for longer is a good way of keeping outlays down and products up in addition to making it easier for artists to secure good support pack.

22:15For the chosen cities who get lucky, often London, often New York or Los Angeles, this means huge windfalls. So stars and fans are expected to spend£1 billion on tickets, accommodation and other expenses in London alone, according to Barclays. Similarly to Taylor Swift, who is a multi-night residency in Wembley, injected a predicted£1 billion into the UK economy. But there are losers in this too. For the smaller regional cities bumped from the schedule, such as Manchester and Coventry, it means losing access to megastars and fans spending power, as well as smaller artists. The way we gig is changing.

22:52Instead of dropping£50 on a Wednesday night gig, concerts have become modern Super Bowl-style events. People split out and are willing to spend the money, but they demand spectacle in return. Still, despite the expense, fans don't seem to mind going on the road themselves.

23:13Around 25 % of the fans watching Mr. Siles at Wembley are travelling from outside London, and 28 % are turning the occasion into a mini-break. Last year, a survey of 3 ,000 UK adults showed that more than a third said they'd booked overnight accommodation to attend a gig. One person who was at Harry Sardis opening night at Wembley travelled 1 ,500 miles from central Romania, whilst another travelled 3 ,900 miles from Indiana in the USA to be one of the first in line when the doors opened. These days, concertgoers are cultural tourists as well as giggoers. In an age where music is cheap and increasingly easy to access, live shows have acquired a new social value.

23:59there's scarcity value there of course because gigs sell out very quickly but these gigs also offered the chance to connect with other fans as well as the actual artist in that sense concerts have become a kind of modern pilgrimage journeys made out of devotion to an idol one thing's for certain when it comes to live music the future of touring involves a lot less of it than they might suggest.