MacroVoices #551 Michael Every: Decoding The Global Geopolitical Puzzle (Part 2)

Anthropic

Across the shows: 5 speakers in other episodes, 6 bullish and 11 bearish calls, past 30 days

• Erik Townsend · MacroVoicesBearish

  Anthropic's training on user data could push enterprises toward free Chinese open models.

  Erik relays Matt Berry's view that Anthropic's terms let it keep and train on personal and business information shared with Claude, such as strategy discussions. Berry thinks that once users tire of their data being mined, they will move to free Chinese models run on their own hardware. Berry expects this to eventually break the business model of Anthropic and of US frontier labs more broadly. Erik presents this as a clever argument and asks what it does to the US AI industry.

  “So Matt thinks we get to a point where people get sick of having their data mined and Anthropic is training on your... Matt thinks eventually that blows up, and everybody says, I'd rather just get these Chinese models that are free and put them on my own hardware.”at 39:33

Oil & fuels

• Michael Every · Rabobank

  Hormuz shipping risk and Iranian escalation keep crude flows and oil prices elevated despite Saudi shuttling.

  Saudi Arabia and the UAE are successfully shuttling more oil through Hormuz, but live firing continues and ships are still being hit. The Saudis are proceeding anyway, so oil is more expensive from a security perspective because of higher transit, insurance and risk costs. Iran is escalating after a one-week US ultimatum and is threatening to strike in the Indian Ocean as well as via the Houthis in the Red Sea. Iranian missiles have already reached Cyprus, so ship-to-ship transfers off Oman could also be hit. Michael therefore sees risk of lower crude flows on top of higher costs.

  “the Saudis have been prepared to go ahead with it anyway, which just means that oil is that much more expensive from a security perspective.”at 4:51

• Michael Every · Rabobank

  A US diesel export ban would backfire, raising crude and fuel prices and causing shortages at home and abroad.

  The global energy system is integrated and relies on counterparties, so stopping US diesel exports solves one problem and creates others. The blowback would be more expensive crude, which feeds into higher fuel prices, plus shortages internationally and possibly inside the US. Michael expects the ban to be narrowed to specific countries and supported by further measures such as refinery run mandates or subsidised diesel storage. He is deeply cynical that this integrated system strengthens the US's position relative to China and others. He says a private-sector energy executive would oppose it as a headline measure.

  “if you suddenly stop The US exporting diesel, it doesn't just solve one problem. It creates lots of others in equal measure. And the blowback can be more expensive crude”at 7:15

• Michael Every · Rabobank

  A 90-day US diesel export ban is plausible and was always on the cards.

  Every says a potential 90-day US diesel export ban is being discussed, and that some people were surprised by it. He says it has always been on the cards for the same reasons: energy and key commodities are being fought over in zero-sum competition, and the US wants to keep supply at home. He treats this as part of a wider pattern of states restricting the flow of key commodities during rivalry.

  “some people are surprised we've got a potential ninety day diesel export ban being discussed. Who thought that wasn't possible? And that's always been on the cards for me too for the same kind of reasons.”at 48:58

US–China relations

• Michael Every · Rabobank

  Serious national security circles see Xi possibly moving on Taiwan next year, though this is unconfirmed.

  Michael says there are serious discussions in national security circles that Xi Jinping is deadly serious about past statements and would consider moving on Taiwan next year. He notes the idea is not new and that no one knows whether it is true. If it happened, the worst case would be a US–China clash in Asia on top of the Ukraine and Middle East wars, which he describes as a small-scale world war. He links this to US weakness: interceptor shortages and limited bunker fuel for the Pacific fleet, which give China home advantage. He frames the US choice as either a deal with China or making a Chinese move strategically unviable.

  “That Xi Jinping is deadly serious about past statements he's made and that he would consider moving on Taiwan next year.”at 24:14

• Michael Every · Rabobank

  Trump's refusal to extend the China trade truce past early January suggests the US sees a coming card to play.

  At the last summit the US agreed a trade truce under which tariffs on China are lower than before, and China was not punished as harshly as expected. China asked at this summit for the truce to run through the end of Trump's term, but Trump would only extend it until early January. Michael asks why the US would refuse, and suggests it may expect a shift in its position over the next few months. He speculates that new AI or mineral processing technology could reduce China's leverage, though he admits this is outside his expertise.

  “And this time, China asked to have that trade truce extended all the way through to the end of Trump's term. And Trump turned around and said, no. We only want to do it until early January.”at 32:48

Open-source AI models

• Erik Townsend · MacroVoices

  Free Chinese open models run on own hardware could upset US frontier labs' business.

  Erik relays Matt Berry's argument. US frontier labs such as Anthropic and OpenAI produce the best models, but they are expensive and their payback is unclear. Chinese open-source models can be run on a company's own hardware and air-gapped from the lab's cloud, so the company's data is not used to train the next model. Berry expects this to upset the market. Erik says the free Chinese models are about as good as US frontier models were three or four months earlier, and he sees China possibly running a strategy of letting US labs spend heavily on research, copying their models, and releasing them for free to undercut the US AI boom. He calls it a clever argument and says it is speculative.

  “what you end up with open source for free out of China is a model that's about as good as the frontier labs in The US were delivering three or four months ago.”at 39:10

• Michael Every · Rabobank

  Chinese open-source models trail US frontier models by about three to four months because they clone them.

  Every argues that much of China's open-source AI performance comes from standing on the shoulders of US models. He says Chinese open models are always three or four months behind the American frontier because they are effectively cloning them, including by using US AI systems through intermediary tools to do tasks. He hedges that he is not an AI expert and does not know whether this holds for all Chinese models. He reads it as an attempt to undermine the US model race because China fears being left behind.

  “a lot of the Chinese open source performance has been because they're effectively standing on the shoulders of The US models, that they're always three or four months behind... I don't know whether that's true. It may not be true for all of them.”at 41:32

Canada, Australia & other developed economies

• Michael Every · Rabobank

  Canada's pivot away from the US toward Europe makes no economic sense and would carry severe downside risk.

  Canada's high-value industrial exports are deeply integrated with the US, and about 90% of its population lives within a two-hour drive of the border. Its tradable goods are mainly raw materials, which Europe can already buy from Canada without any new trade deal. Europe does not want Canadian cars, trains or jets because it has enough of its own. Disentangling from the US would therefore mean trading more with a block that wants only Canada's raw materials. Michael adds that the US would not accept a Canada drifting toward China or Europe with an open, undefended border, so the downside tail risk is staggering and the upside is unclear.

  “Europe can buy those raw materials from Canada right now. There's no need to have any kind of new trading relationship... Europe does not want to buy Canadian cars. Europe does not want to buy Canadian trains.”at 13:38

Russia & Ukraine

• Michael Every · Rabobank

  Russia's jet-powered Iranian-design drones are outpacing Ukrainian interceptors, so Ukraine cannot match Russian strikes.

  The drones now striking Ukraine were originally an Iranian design. Russia has improved them and added jet engines, making them much faster than Ukrainian drone interceptors. Russia is hitting Ukraine hard and Ukraine has no matching response at the moment. Michael also suggests these Russian-engined drones may return to Iran to be fired at US, Israeli and other forces in the region, which shows deep Russia–Iran military collaboration.

  “The drones, are currently striking Ukraine, are originally an Iranian design. The Russians have improved them and put jet engines on them, and then are much faster than Ukrainian drone interceptors.”at 20:32

Defense

• Michael Every · Rabobank

  The Iran war exposed US shortages of interceptors and Pacific bunker fuel, weakening it for a sustained China fight.

  The war against Iran has shown the US does not have enough interceptors. Bunker fuel for the US Pacific fleet is also short, and the US is short of several inputs it would need for a sustained war across the Pacific. China would have home advantage in any such conflict. Michael uses these shortages to explain why Washington would want to avoid the risk of a Taiwan clash, either by negotiating with Beijing or by making a Chinese move strategically unviable.

  “the war against Iran has already exposed that we actually don't have enough interceptors... We don't have enough bunker fuel in Hawaii anymore for their Pacific fleet.”at 25:36

Manufacturing & industrial policy

• Michael Every · Rabobank

  US industrial and military capacity is far weaker than assumed, and it cannot easily match China's shipbuilding.

  Michael argues the US had far fewer resources than it believed and did not realise how out of shape it was, like a fit-looking man in his fifties who discovers he cannot keep up. He contrasts this with China's industrial complex, which can build a new ship in five minutes rather than five years. He asks whether the US could take on China in this context, and if not, how it could slow China down. The argument is a broader view that US industrial weakness constrains its strategic options.

  “my viewpoint is somewhat different, which is The US already had far fewer resources than it liked to think it did.”at 30:51

Frontier AI labs & model race

• Michael Every · Rabobank

  US may let private AI labs struggle, then seize them on national-security grounds.

  Every argues that if AI could deliver a super-intelligence that solves problems, it would be a strategic gain for the US if private AI labs fail. The US cannot afford to fund the buildout with more debt. The model he sketches is that the US lets international capital, including Chinese capital, buy shares in the labs, then the government takes the technology on national-security grounds once it works, as with the Manhattan Project. He says this is not a forecast but a realpolitik reading of the incentives.

  “I would actually argue it's a genius strategic move by the Americans to say, look, we can't afford X number of trillion dollars borrowing on top of the deficit we've already got to develop this technology”at 43:08

US–China AI & chip race

• Michael Every · Rabobank

  US will likely bar Chinese open-source AI models, as it did Chinese EVs, soon.

  Every points out that Chinese EVs, which record data, are sold in the West but not in America, which he says shows a template of US exclusion of Chinese technology on security grounds. He argues the same logic should extend to Chinese open-source AI models, and he expects that to happen imminently. He frames this as a national-security reaction to a competitor that is 'better' and that the US does not want to have.

  “How many can you buy in America? None... Why isn't that gonna be extended to open source stuff from China too imminently”at 48:34

Wealth, inequality & billionaires

• Michael Every · Rabobank

  Exit taxes and unrealised-gains taxes are spreading and will likely be extended, as cash-strapped governments need funds.

  Every says many jurisdictions already impose ruinously high exit taxes, and some are adding taxes on unrealised capital gains, so people owe tax each year on assets they have not sold, and half of their holdings can be lost on leaving. He says more countries are adopting this and he expects it to extend because governments are cash-strapped and need money for AI infrastructure, defence and energy stockpiles. He says this is high taxation rather than a formal capital control, but it matters for people with large capital.

  “there are many jurisdictions where you already have ruinously high exit taxes... There are more and more countries bringing that in. And I expect that to be extended and extended because governments are so cashed out everywhere.”at 52:47

Middle East & Iran

• Michael Every · Rabobank

  Iranian tanker attacks and Houthi escalation could push the US to try to topple Iran's regime.

  Every says that if Iran is so entrenched that it can cause chaos along supply chains, keeping the Middle East destabilised becomes too painful to sustain. The US would then have to pivot to its main objective, which he describes as flipping Iran like Venezuela to control the key commodity complex. He sees a path where escalation against tankers in the Indian Ocean and Houthi attacks make the diesel export ban insufficient. He does not predict nuclear use and says he does not want it, but he does not rule out tactical strikes on Iran's underground nuclear sites.

  “In that case, you have to try and pivot back to your primary objective. Let's flip Iran like Venezuela so I'm sitting pretty in terms of that key commodity complex.”at 56:17