Metabolic Competition

Countries compete with each other economically, militarily and politically. While we typically measure GDP, number of aircraft carriers or alliances to understand global power, soon we may be measuring waistlines.

My framework for analyzing an unruly world insists we look at how politics, law and technology crash into each other to create new risks and opportunities. Breakthrough medicines have the potential of being just as impactful as breakthrough AI.

“Skinny jabs” - a somewhat insensitive British shorthand for Ozempic, Wegovy, Mounjaro and other GLP-1s that restrict appetite-may not seem a likely candidate for a geopolitical lever. But their widespread usage and potential to reshape the diet and fitness of large parts of the global population makes them so.

GLP-1s offer the promise of weight control, diabetes management, lower cardiovascular risk, and decreased cholesterol—all of which can extend longevity. Though these drugs are expensive, have nasty short-term side effects and unknown long-term health impacts, 12% of Americans are currently using them. When you consider that 70% of Americans are obese or overweight under some definitions, the audience for these medicines involves most of society, who will increasingly take such medicine as prices go down and side-effects come under control. Patent protection will soon fall away in China, India and Canada which opens the door to generic versions at very low-cost for billions of people.

So, we need to start to ask ourselves: What happens to food producing countries if large parts of the global population eat less sugary or processed food because their appetite is suppressed? Will public health systems invest to subsidize GLP-1s today to reduce long-term treatment costs? Will healthy populations that live longer come to provide a military advantage? Will voters who live longer skew conservative, as data has shown in the past? And, finally, will GLP-1 precursor and finished product production become a new geopolitical chokepoint that gives producers leverage?

Taken together, we can start to grasp how countries may begin to compete with each other to achieve massive health improvements, or a role in the related supply chain, that could reshape their militaries and their economies.

Economic Implications

The economic implications of GLP-1 drugs are manifold.

Dining habits

It’s easy to see how appetite restricting medicines like GLP-1s will lead to a reduction in snack food and dining out. Because the medicines impact the brain’s reward-center chemistry, they also lead users to cut back on alcohol. Even though users may switch to more protein or nutrient-dense food, there is still a reported cutback of over 5% in spending on food when using the medicine. The New York Times even suggested that those throwing business parties are avoiding seated meals to not make GLP-1-taking guests uncomfortable by having to abstain from corporate gluttony for hours around a table. To the extent the stock market has gone beyond pricing in implications for pharmaceutical companies directly, knock-on effects to food producers has been the main realm of analysis.

Economic Growth

While these medicines are financial blockbusters for the companies - and even countries - that produce them, a substantial cut back in food spend would normally be considered a major economic drag. However, obesity itself is generally considered a drag on economic growth. Thus, Goldman Sachs and Morgan Stanley estimate half to one point of GDP growth in OECD countries from GLP-1s based largely on an assumption that reducing obesity related disease will keep aging folks in the workforce longer. In an era where public health systems are under strain, the idea that subsidizing GLP-1s now could pay off in the long-run through growth may be attractive.

• Shift in trade flows: Of course, as people shift what they eat, that plays out through the entire food supply chain. As people demand less sugary foods or less processed foods or less alcohol, this changes the export pictures in some big ways. How does Brazil replace demand for sugar exports if populations turn away from sugar? Or the US for corn syrup? Or US and Asian processed food manufacturing? What about European alcohol exports? While there will no doubt be opportunities to export foods of the future - perhaps protein rich foods or high-end produce - this could wreak havoc on many economies. Beyond simple demand signals, agricultural production is both some of the politically sensitive and most subsidized outputs an economy can produce, while many countries depend on it for foreign currency and tariff revenue.

• Economies will need to reorganize: Beyond the export issue, there are major implications for economic organization as people trade off calories for longevity. It’s worth remembering this is coming at the same time that AI reshapes workforces and where the balance between physical/manual labor and knowledge work may look different than it has in the past. First, public healthcare systems will have to grapple with whether they want to spend now on GLP-1s to save later on obesity related diseases. Second, while it may be great that people can work longer, if people actually live longer, then already-stressed government pension systems will have to do a lot more work to keep people afloat over a longer time horizon if that extra work doesn’t net out to grow the pension funds. Third, many private insurance products - which are based on actuarial models that won’t match longevity gains - will also no longer be fit for purpose.

Geopolitics and National Security

Once you grapple with the fact that much of the population is a candidate for such medicine, it’s easier to see how there could be implications at the national and global scale.

• Rise in military readiness: While there are concerns that GLP-1s can reduce muscle mass or cause stomach issues for troops in the field, it should be readily apparent that healthier populations provide a much more militarily ready population than unhealthier ones. In the US, two out of three active duty troops are overweight if not obese. By increasing fitness, a country can increase military readiness and also potentially reduce long-term military healthcare costs through upfront investment in medicine, which would leave the military with additional funds for other purposes. Even food security - including rationing during war - could acquire a new dimension, though appetite suppression should not be confused with true nutritional resilience

• Globalization and Dependency: Of course we still live in a globalized world where the countries that are most efficient at producing elements of any product will get the chance to sell that on the open market. While supply chains for GLP-1s are opaque, it’s possible that countries may find themselves with dependencies on other countries in those supply chains if they choose to optimize for use of these medicines. The dependency question may not be whether today’s branded Western GLP-1 doses rely on China, which some do. It is whether the next phase of the market — generics, oral formulations, equipment etc— creates new pharmaceutical chokepoints.

• Conservatism, Body Image and Social Equality: There’s a healthy literature that as populations live longer, they tend to become more conservative as older voters cast votes against change. This is neither a good nor bad thing, but is an implication for political culture we can anticipate as countries go down this path. At the same time, there are many historical periods where body image obsession often co-exists with racism, colonialism and even fascist tendencies. Countries that are able to afford to body hack their populations into physical fitness may also use that as a psychological weapon to justify looking down on those populations that cannot or do not. This may be true within a country, where the cost of medicine may be out of reach of those who would otherwise take it, or across countries. Of course, it’s important to emphasize that there are many cultures that value fuller bodies as a sign of wealth or simply for other cultural reasons. My broader point is that we could see greater fissures in domestic and international politics if GLP-1s allow populations to live longer and present fitness or thinness as the dominant physical form.

Black swans rarely announce themselves in an unruly world. But metabolic competition has been building in plain sight — in obesity data, pharmaceutical advancements, supply chains, public health debates and military readiness reports. The signals are there. The question is whether we are all connecting the dots.

-SW

PS Thanks to and for their contributions.

By collapsing the time frame of any analysis, you get an edge in your planning, reactions, or trading depending on your vantage point. But for the analysis to be valuable, it must also be excellent. We submit our expansive snap reaction as your definitive guide to the summit outcomes and future trajectory. You may not agree with all the conclusions but the analysis is here and now, ready to challenge your thinking and prompt you to dive deeper on the points you find most compelling.

Here you can find our full report that covers what happened at the summit, how the US and Chinese statements differed, what topics were raised and which were avoided, and what uncertainties were left. Then we go further to assess the implications across key global regions ranging from East Asia and Europe to Latin America and Africa as well as across global macro themes. We do this purposefully as the myopia of the media typically only focuses on headlines and never gets around to the fact that summits between two global powers necessarily affect every region.

That’s because a report of this kind has historically required a large analytical team of experts: a China analyst reading Chinese news in Mandarin, an Iran analyst tracking local Iranian reporting, a markets desk pulling commodity prints, an aerospace specialist for Boeing, a Fed-watcher for the Warsh Federal Reserve connections, a Brussels correspondent for ECFR’s reaction, an Africa or Latin America analyst for local implications–each writing on their own timeline, edited up a chain, harmonized by someone who has read maybe a tenth of the underlying material. Unruly collapses that process. It pulls more than fifty sources across eight categories and seven languages and holds them in a single working memory while it writes — no chain of edits to dull the signal, no (unintentionally) competing theses quietly muddying the message.

Given domestic audiences in China and the US, as well as global markets and countries around the world, we decided to look at US and Chinese readouts as parallel artifacts of domestic political theater rather than competing accounts of reality. It surfaces eight named contrarian theses — the Hormuz transit was pre-arranged between Beijing and Tehran; the Nvidia H200 approval may produce no actual deliveries because Beijing is telling firms to pause; the Boeing order may quietly evaporate the way the 2017 announcement did.

That is how careful analysts actually think but rarely have time to write down. The report marks its own confidence levels, lists what it doesn’t know, and tags every claim by perspective (us_view, china_view, markets, local_regional, third_party), so a reader can see where the framing is coming from rather than absorbing a single house view.

For a CEO routing supply chains, a PM sizing a position before Asia opens, or an EU desk officer trying to figure out what just changed, the value is having the synthesis in hand while the decision window is still open. By the time a comparable human-authored product lands in your inbox three days later, the rare earths print has moved, Boeing has already made its move and the question you needed answered has been answered for you by the market.

You should demand more from political risk analysis as the frontier of what’s possible moves outwards. Be in touch with your hardest questions on this or or your other key issues so we can show you more of what’s possible.

P.S. The post-summit app is completely free with no registration required so do check it out.

-SW & DB

Salespeople are usually right about who will buy from them. They are usually right about how much that buyer will spend. But they are almost always completely wrong about timing.

For anyone who has managed sales teams in the professional services world, this no doubt rings true. Deals that seem within reach stretch on perpetually. Deals that were verbally confirmed are suddenly upended by a firing or a new hire. Even deals that are signed can become unsigned at the plea of a client for whom external conditions have changed.

Salespeople often bear the brunt of these outcomes, being fired for not meeting goals, only for an executive or a new hire to claim victory by capturing the same deal a year later. In reality, the original sales person was right about the client, they were right about the amount, but they could not control timing.

In an unruly world, this is becomes an even bigger challenge.

Sales in an Unruly World

There are two key factors to appreciate when thinking about selling things in the new world - particularly if you sell knowledge goods.

First, your customers are dealing with far tougher “external conditions” than in recent decades absent acute crisis moments. Some of this is geopolitical. Clients are typically not in buying mode when the cost of energy spikes, their supply chains get reorganized, and their goods are facing an uncertain soup of tariff codes.

As humans, they are still genuinely interested in your service. And when things calm down, they may still buy from you. But will things calm down? If you’re familiar with my view of the unruly world, I think you’ll be waiting a long time for peace.

Second, every single day that your client is not cementing a deal with you, they are facing a raft of substitutes for what it is you do. This is is very different than before. If you are trying to sell a research suite to a client who asks you to re-engage in three months, are they more or less likely to see AI-powered versions of research that render your old-school approach obsolete? Are they more or less likely to experiment with their own tooling? Are they more or less likely to be hit by AI-powered sales techniques of all your competitors touting their superior offers?

Law firms and strategic consultancies have long relied on soft selling approaches, wining and dining clients for months or years hoping that when a bet the company issue arises they get the first phone call. Higher-end advisory firms often rely on hosting banner client events to build brand and regularly following up with clients to try to discover a hook for work they can do together

In today’s world, if you’re unable to close the deal quickly, then you aren’t just wrong about timing - you’re more likely to be wrong about the deal overall.

Keep the platform burning

If you’re not proactive, you’re dead. But proactivity does not mean frantically calling clients to pressure them to do a deal.

An AI-native accounting firm was trying to sell me their solution recently, and I made clear I wanted it but I didn’t think their pricing was reasonable. Instead of addressing the pricing issue or adding further value to my business, they offered me a $250 Lululemon gift card if I signed up within 48 hours. Now, corrupting consumer psychological techniques (like offering the buyer a benefit that accrues to them personally to take a bad deal for the company) can be effective - though, for me, that would be something like coffee rather than yoga pants.

But the bigger issue is once I went back into the market, I found there are a dozen other companies doing exactly the same thing, all willing to undercut the first. That’s because instead of there being one AI-native accounting firm there are now dozens - and the same goes for whatever it is you are trying to sell. If the first company had a lower reservation price than they articulated, they won’t get the chance to articulate it in the future because I quickly discovered I had a lot of choices and I’m out of the market.

Proactivity would have been issue spotting rather than pressure sales. If that first company had told me that the issues I wanted to solve were only part of my challenge and by looking at public information about my company, there were three things I wasn’t thinking about that they could help me with, I probably have been disarmed about pricing. I probably would have felt like I needed to get to work on the balance of issues quickly. But most importantly, I would have felt like the company understood me. Instead, the company treated me like an undifferentiated opportunity.

In the new world of sales, it’s not enough to develop a pre-meeting briefing about a client before you walk in the door to try to wow them. You need to be 24/7 on in monitoring what’s important to these clients and you need to be able to add value to those clients over the course of their consideration cycle at every turn. It is literally costless to know as much as you can about your client - and it will cost you a lot if you don’t.

At Unruly, we are building software to increase client awareness for law firms and consulting firms. We track all of a firm’s top clients and help them see every concrete issue that their client is dealing with in the external world today. This helps the firm see scenarios that will affect that client tomorrow and to be the first one to reach out to their client every day with an understanding of what that client needs.

Sales people are often taught to create a “burning platform” so that a client feels like they have no choice to jump off into the safe waters of whatever is being sold to them.

The goal now is to figure out how to keep the platform burning by throwing on the accelerant of real-time evidence of all of the issues that will imperil the client’s business if they don’t make the decision to work with you now.

Every day, there are more and more companies in the AI-powered lifeboat business. In an unruly world, if the client doesn’t jump soon, someone else will now rescue them before they catch fire.

-SW

Geopolitics bedevils corporate decision-making. Whether it’s long-standing trade tensions with China, the weaponization of tariffs, or the spontaneous reshaping of the geopolitical map through force, corporations are faced with an evolving list of risks to manage. Today these risks have moved from peripheral to squarely material, so we wanted to understand how companies are talking about them – particularly when they are legally obliged to disclose and tell the truth. Since we can’t access their private strategy meetings, we turn to their filings and earnings transcripts as a proxy.

Our colleague Alexander created a tool to ingest and analyze the last five years of earnings call transcripts and SEC filings from Fortune 50 companies. Then, an using our Unruly AI toolkit we gave each document a salience score from 1 (boilerplate risk disclosure) to 5 (all about geopolitics). Salience measures how explicitly management named and quantified specific policies, regimes, and cross-border events. We then used these salience scores to look for patterns and interesting outliers. It’s probably no surprise that managed care ranked the lowest while energy and information technology were the highest.

Across the dataset, while several Fortune 50 companies scored highly on salience, NVIDIA stood apart in that its salience remained almost consistently rising with only two downward swings after 2Q2022. Digging deeper, we found that while export control to China was the catalyst in raising its geopolitical salience score, beginning in Q4 2023, NVIDIA executed a deliberate narrative strategy to leverage the concept of “sovereign AI” as an explicit way to reframe its geopolitical exposure as a commercial opportunity. Sovereign AI refers to the idea that governments should develop domestic AI capabilities, including compute infrastructure, data systems, and models, as a concern of national economic and strategic autonomy (NVIDIA Blogs).

NVIDIA’s recent disclosures show how companies can turn geopolitical exposure into a narrative about geopolitical opportunity. Between Q2 2020 and Q1 2026, the company shifted from acknowledging geopolitical disruptions to making geopolitics central to the story it told to its investors. As U.S. export controls restricted NVIDIA’s data-center business in China, its management reframed the geopolitical environment it found itself in as an opportunity for new demand in sovereign AI.

Notably, other companies sitting at a similar geopolitical intersection defaulted instead to conventional risk management narratives.

The rest of this blog examines how NVIDIA turns China exposure into sovereign AI opportunity, why that story appears most clearly in earnings calls while SEC filings retain the language of legal and compliance risk, and how this makes NVIDIA unusual among the Fortune 50.

NVIDIA vs the Fortune 50

Across the Fortune 50, NVIDIA is an outlier not only because geopolitics appears frequently, but because management makes it central to the company’s growth story.

For most companies, geopolitics appears as an external disruption: tariffs, sanctions, war, supply-chain instability, commodity volatility, foreign exchange, or country-specific operating risk. These issues are material, but they are usually framed as shocks to be managed. The company absorbs the disruption, adjusts the supply chain, manages compliance, hedges exposure, or passes through cost.

NVIDIA is also unique among much of the technology sector. Apple, Microsoft, Dell, and others all disclose or discuss geopolitical exposure, but in a more fragmented pattern. Apple’s geopolitics clusters around China manufacturing, EU regulation, App Store rules, and supply-chain adjustment. Microsoft’s clusters around sanctions, cyber risk, public-sector demand, and export restrictions. Dell’s centers more on supply chains, AI servers, and China-related controls.

Other companies also sometimes turn geopolitics into a positive story. GM and Ford frame U.S. industrial policy as support for domestic EV and battery manufacturing. UPS treats nearshoring and the shift of manufacturing closer to the United States as an opportunity for its Mexico and cross-border network. Energy companies make a parallel argument that geopolitical instability increases the value of diversified LNG, upstream production, and reliable supply.

But these are narrower reframes. For GM, Ford, UPS, Exxon, or Chevron, the opportunity usually attaches to a policy, asset, or supply-chain adjustment. NVIDIA’s approach is more concentrated and more ambitious. Its management not only tells investors that geopolitics affects the business, but also actively positions geopolitical fragmentation as an opportunity to expand demand for its services – if export controls restrict one market, then sovereign AI, allied infrastructure build-outs, and national compute strategies attaches it to others with an explicitly geopolitical agenda.

NVIDIA’s Geopolitical Narrative Strategy

NVIDIA entered this decade carrying an unusually concentrated set of geopolitical exposures. Its high-performance GPUs, central to AI training and data-center acceleration, placed the company at the heart of the U.S.-China technology rivalry, which was strategically valuable because its chips enabled frontier AI development, but also strategically vulnerable because that same importance made them a target for export controls.

By late 2023, NVIDIA disclosed that U.S. licensing requirements covered products including A100, A800, H100, H800, L40S and other high-performance chips, and that sales to China and other affected destinations had contributed roughly 20–25% of Data Center revenue over the prior few quarters. Its reliance on Taiwan compounded this exposure as NVIDIA’s supply chain used TSMC to produce semiconductor wafers, and NVIDIA itself warned that tensions involving Taiwan and China could materially affect suppliers, contract manufacturers, and assembly partners critical to supply continuity. By 2023, then, NVIDIA was not simply another multinational exposed to geopolitical noise. It occupied the rare position of being both highly exposed to geopolitical risk and capable of shaping that risk through its own product design, customer choices, supply-chain decisions, and role in the global AI infrastructure build-out.

To understand how management communicated about these geopolitical risks, we used an LLM to label and analyze every NVIDIA earnings call from Q2 2020 through Q2 2025. For each call, we extracted distinct geopolitical event mentions – named policies, countries, regulatory regimes, wars, sanctions, export controls, and cross-border dynamics – and classified them by theme. We also tracked whether the subject was raised by analysts or introduced by management.

Across the first fourteen quarters analysed (Q2 2020 to Q3 2022) NVIDIA’s geopolitical narrative is mostly defensive. Export controls recur throughout the period, Russia-related mentions rise after the February 2022 invasion of Ukraine, and China/Taiwan supply-chain risk becomes more prominent in May and August 2022 as advanced-semiconductor restrictions moved to the center of U.S.-China technology competition. The 2022 BIS restrictions marked the turning point, because once the export of its A100 and H100 chips to China and other restricted destinations required U.S. government licenses, geopolitics became a direct constraint on NVIDIA’s Data Center business.

NVIDIA’s language in this phase is managerial rather than opportunistic. It identifies sales at risk, records inventory charges, warns of disruption to product development and China-based operations, seeks regulatory exemptions, and introduces compliant substitutes such as the A800 chip. Geopolitics, therefore, is presented as a challenge that it is absorbing and finding ways to route around.

Then its earnings call in November 2023 changes the narrative around its geopolitical exposure.

That earnings call contains the highest number of distinct geopolitical event mentions in NVIDIA’s earnings-call record to that point, but the important feature to look at is the topic composition of that call. Earlier geopolitical mentions had mostly clustered around constraints, such as export controls, China/Taiwan risk, and supply-chain exposure. In this call, the focus shifts towards “Sovereign AI”. India, France, Japan, Sweden, Germany, and the UK are named individually, alongside a broader global claim that national investment in computer capacity had become a “new economic imperative” and a “multi-billion dollar opportunity.” NVIDIA was, by then, no longer telling investors where risk was emerging but was pointing them towards where new demand might come from. In other words, NVIDIA recast the same geopolitical fragmentation that was underpinning access restriction to China as the driving force behind U.S.-allied governments elsewhere wanting to build domestic AI infrastructure.

The timing in November 2023 is significant because this was also the same reporting cycle in which NVIDIA had to tell investors that new U.S. export restrictions would hit products previously sold into China and other affected destinations, which had contributed roughly 20–25% of data center revenue over the prior few quarters. What follows is a consolidation of the narrative. Sovereign AI continues to appear in later calls, but the event count normalizes and management shifts from naming countries to sizing the opportunity: first as a multi-billion-dollar market, then as high single-digit billions, then low double-digit billions. The chart captures the launch. The revenue commentary explains what the launch was for.

Proactive vs Reactive

One of the most important learnings in our analysis of Fortune 50 earnings calls was that some companies proactively shape their geopolitical narrative through their SEC disclosures and mentions on calls by their executives while others only engage on the topic when asked about it by analysts. Interestingly, companies overall are growing more proactive, with forward-looking disclosure rates doubling from 12.8% in 2024 to 27.7% in 2025.

The focus on “sovereign AI” in NVIDIA’s earnings calls appears to be initiated by management rather than by analysts. Analysts mainly asked about China restrictions, supply constraints, and near-term revenue exposure. Management, in turn, answered those questions but then widened the frame. Colette Kress, NVIDIA’s CFO translated sovereign AI into a revenue category. Jensen Huang, CEO, then gave it a more ideological and geopolitical texture, arguing that countries increasingly see their data and AI capacity as national resources, and therefore want to build domestic AI infrastructure.

The division of roles within NVIDIA’s management is also important. The timeline shows that sovereign AI first appears as a response to the export-control shock of late 2023, when China risk had become harder to avoid and management needed a credible alternative growth story. On the November 2023 call, Kress and Huang begin that pivot together: Kress frames sovereign AI as a “multi-billion dollar” revenue opportunity, while Huang gives the idea its broader strategic logic by pointing to governments that want to build their own AI infrastructure. That split becomes clearer through 2024. Kress increasingly translates sovereign AI into a measurable revenue category, moving from general opportunity language to claims that it could grow from virtually nothing into billions of dollars in annual revenue. Huang, by contrast, supplies the geopolitical and ideological texture, arguing that countries now see their data as a national resource and want to develop their own “digital intelligence.” The result, perhaps somewhat unsurprisingly, is that Kress takes on the work of articulating “Sovereign AI” as a market, while Huang makes it legible as a theory of national technological power.

National Champions

NVIDIA’s sovereign AI narrative is interesting because it helps square what might otherwise look like a contradiction. On one level, NVIDIA increasingly looks like a U.S. national champion: a company whose chips sit at the center of American AI power, export-control policy, and technological competition with China. At the same time, its sovereign AI strategy is about helping other countries build their own domestic AI capabilities. But those two things are not really at odds, rather they are part of the same strategy.

The point is not that NVIDIA is helping other countries become independent of U.S. technology. It is that NVIDIA is offering them a version of AI sovereignty that still depends on U.S.-designed infrastructure. For governments, this is attractive because it promises national compute capacity, domestic control over data, and a stronger claim to participation in the AI economy. For the U.S., it is attractive because those capabilities are being built through an American company. NVIDIA’s role, then, is not just to sell chips. It is to extend U.S. influence through the infrastructure other countries use to build their own AI systems.

This also explains why the sovereign AI story works so well as a business strategy. If China is increasingly closed off, and NVIDIA now says it has 0% market share there, the company needs a growth story that does not depend on recovering that market. Sovereign AI provides one. Instead of a world divided between the U.S. and China, NVIDIA presents a world of many national AI markets, each with its own demand for compute, data infrastructure, and model-building capacity. Geopolitical fragmentation, in this telling, does not only close off one market. It creates the logic for building many others.

That does not mean NVIDIA escapes geopolitical risk. Export controls still bite. China still matters because Taiwan remains a major structural vulnerability. But completely escaping risk can never really be the end goal. The point is to be better positioned when those risks inevitably arise. NVIDIA’s achievement is that it has stopped geopolitics from being only a story about constraint. Instead, it has made geopolitical exposure part of the reason the company matters.

That is why the national champion frame is useful here. As Sean West argued in GeoLegal Notes in March 2025, companies under political pressure increasingly need to make their case in the language of both economic logic and patriotism. NVIDIA’s sovereign AI narrative does exactly that. Governments get a story about national capability. Investors get a story about durable demand. Washington gets a story about allied technology leadership. NVIDIA gets a way to turn the same geopolitical fragmentation that threatens one market into the rationale for building many others. In an unruly geopolitical environment, the companies best positioned will not simply be those that manage risk well. They will be those that can explain why their growth also serves the strategic priorities of states.

About the Authors

Leo is an intern at Unruly and a History and Political Science Major on the Columbia – Sciences Po Dual BA. Alex is an intern at Unruly, where he works on applied AI tools and research, and a junior in high school planning to major in data science.

It is against this backdrop that I note I have been sent Anthropic’s job posting for a geopolitical intelligence analyst by nearly a dozen people. I’ve been sent it by my wife. I’ve been sent it by other people’s spouses. I’ve been sent it by employees who I hope didn’t apply. I’ve been sent it by friends of friends who think it’s something I should know about. I’ve been sent it by enemies who want me to fold up shop at Unruly and stop trying to take their market, retreating instead to the internal affairs of Anthropic.

The job posting itself has even been reported as news. Never have I seen the geopolitical risk internet (a small corner of the universe to be sure) in such a tizzy over a role.

Here are four reasons why this is important.

First, it underscores that even Anthropic won’t simply use Claude to do geopolitical work! The geopolitical world has begun to dabble in the art of AI, using LLMs to support their work. But when arguably the hottest company in AI and vibe coding still needs humans to “conduct horizon scanning” and “provide geopolitical context for international conferences”, we all must realize that the general purpose models, given where they stand today, do not yet automate geopolitics. (I would argue that purpose built solutions are much better at this).

Second, geopolitical risk work is now a bedrock of corporate activity. Big energy companies have hired geopolitical analysts for a long-time but tech companies typically didn’t. Today, only a small percentage of companies overall have geopolitical teams and those teams are often left to wrestle with strategy, government affairs and the like for influence. Few companies place geopolitical analysis at the core of their of strategic decision-making.

But this job is a bit different. Anthropic itself is a geopolitical football being kicked around by Washington. The job explicitly works on “international expansion and facility siting.” As data centers become flag-bearing national outposts of American companies, choosing where to be and where not to be is much more than risk management. It represents some of the most strategic choices a company will make.

Whether this actually elevates the role of the in-house analyst or not remains to be seen but it underscores that you can no longer do corporate or external affairs without geopolitical affairs.

Third, the opportunity solidifies that the most exciting jobs in the geopolitical risk space may now be inside corporate, if not inside tech and AI specifically. For decades, the dream of many political science students was to land a seat at a major geopolitical advisory firm and work up to being a Managing Director or Partner. I followed this path. Yet, most consulting firms in the space are looking to hire ex-government movers-and-shakers now, not analysts, because they believe that access cannot be replicated by AI. Thus, the pathways to train up to become excellent are closing as AI eats more of the entry level work.

In fact, the number of geopolitical providers I’ve talked to who are running scared about what AI will do to their business model underscores even those sought-after political risk industry roles may only be a temporary safe-haven. At Unruly, we are finding we can serve more and more companies with AI generated geopolitical risk research. That means the role of the human in the model must evolve considerably which offers a different value-proposition to would-be analysts.

Of course, a key selling point of such a role would be building this at Anthropic itself. If you suddenly had access to all the tools and all the compute, you could probably develop some pretty novel ways to model and understand the world. While Anthropic may be an outlier, corporations are going to offer those types of compute packages far more than the consulting industry will in the future (unless, you work for an actual geopolitical tech company, like ours). Thus, if you want to solve big problems, you need big compute, and big compute means big budget which means big corporate.

Finally, the excitement around the job posting implies that even the most risk averse industry - the risk industry - gets excited by a one-in-a-lifetime bet. The job advertises top compensation at $220,000 which is less than what a top analyst would earn in the consulting world, even a lower salary than government Senior Executive Service roles. But presumably the job comes with options over Anthropic shares, which is like buying a lottery ticket where five your six numbers have already been drawn and matched. You’re just waiting for the sixth number so you know how much you’ve won.

You don’t need to be an intelligence analyst to realize that would be an intelligent move. Especially when there are far fewer lottery tickets available to geopolitical analysts than, say, AI engineers.

Based on the all the hubbub, I would guess that roughly half the people in my professional universe applied for this job. To the winner: I hope you have a good few years training Claude before it takes your job too.

-SW

Executive Summary

Viktor Orbán’s 16-year reign ended on 12 April 2026 in a landslide. With 95.9% of list votes counted, Péter Magyar’s Tisza party leads with 53.7% against Fidesz-KDNP’s 37.8% (Europe Elects). Seat projections based on 90.9% of votes counted show Tisza on track for approximately 138 of 199 parliamentary seats — a two-thirds supermajority — while Fidesz-KDNP is projected at roughly 54 seats (Europe Elects; TVP World). Turnout reached 77.8% by 6:30 PM — before polls even closed — surpassing the previous all-time record of 70.5% set in 2002 (Unruly / AFP). Orbán conceded, calling the result “painful but unambiguous” (Unruly / AFP).

The core judgment: This is not a clean democratic restoration. It is the beginning of a protracted institutional reckoning in which a former Fidesz insider — socialised in the very system he promises to dismantle — will attempt to reverse 16 years of state capture using tools that Orbán deliberately booby-trapped before leaving office (Politico). The supermajority, if confirmed in final results, is the single most important variable: it gives Magyar the constitutional authority to rewrite the rules Orbán embedded in cardinal laws. The threshold is 133 seats (Bloomberg). Current projections of 138 seats suggest Tisza will clear it, but final results — including individual constituency races — are not expected until later this week.

Biggest geopolitical winners: The European Union, Ukraine, NATO cohesion, and Hungarian civil society. Biggest losers: Russia’s leverage in the EU Council, China’s privileged access to Central Europe, and the transnational illiberal-populist network that treated Budapest as its ideological capital. The defeat also lands as a symbolic blow to Trump-world at a moment when the White House invested direct political capital — JD Vance rallied with Orbán in Budapest five days before the vote (BBC) — and lost.

What Happened and Why It Matters

Orbán lost because corruption became visible enough to override culture-war politics. The battery-plant environmental scandal — involving Chinese and South Korean factories accused of toxic discharges with government complicity — crystallised public anger over a regime that had become more interested in enriching its network than delivering services (Reuters; Le Monde). Healthcare deterioration, economic stagnation, and the 2024 presidential pardon scandal involving a child abuser’s accomplice created the opening (Unruly / AFP). Magyar, a former government insider who knew where the bodies were buried, filled it.

The result is historically significant for three reasons. First, it is the most decisive defeat of an entrenched illiberal leader through the ballot box in 21st-century Europe — Hungarians turned out in record numbers to end Orbán’s rule (AP News). Second, it occurred despite an electoral system Orbán had gerrymandered over four election cycles, which analysts estimated could have allowed Fidesz to retain a majority even while losing the popular vote by three or four percentage points (Unruly / AFP; New York Times). Third, it happened after the sitting US Vice President flew to Budapest to campaign for the incumbent — and the electorate rejected the intervention (Salon).

Magyar declared the election win had “liberated Hungary,” telling cheering supporters in Budapest: “Together, we brought down the Orbán regime” (Unruly / AFP). Orbán, for his part, stated: “The election results, though not yet final, are clear and understandable; for us, they are painful but unambiguous. We have not been entrusted with the responsibility and opportunity to govern. I congratulated the winning party” (Unruly / AFP).

What remains uncertain: the final seat count (official results expected later this week), whether Tisza secures the full two-thirds needed to amend the constitution (133 seats; current projections suggest 138), and how quickly Magyar can form a government given the institutional traps Orbán has laid.

Is China a Loser?

Yes — but not catastrophically, and not yet irreversibly.

Orbán was Beijing’s most reliable partner inside the EU. Hungary served as China’s gateway for EV battery supply chains — CATL’s €7.3bn plant in Debrecen being the centrepiece (Reuters) — along with BYD’s approximately €4bn factory in Szeged, which has already begun trial production (Reuters), the aborted Fudan University campus in Budapest, and Huawei’s logistics hub. More critically, Orbán reliably blocked or diluted EU consensus positions on China, from Taiwan statements to investment screening (South China Morning Post).

Chinese firms in Hungary now face a “post-election reckoning” regardless of who won, as the South China Morning Post reported (SCMP). Magyar has signalled a rebalancing, not a rupture. He is unlikely to tear up existing investment contracts — the battery factories employ thousands and represent billions in sunk costs. But the political environment will shift materially:

• EU veto cover disappears. Hungary will no longer reflexively shield Beijing in Council votes on trade defence, investment screening, or diplomatic statements. This is the single largest loss for China.

• Regulatory scrutiny increases. The battery-plant scandal was a campaign issue. Revelations about foreign battery makers ignoring safety and environmental regulations — and the government helping cover it up — were a major headache for Fidesz (Balkan Insight; Green European Journal). A Magyar government will face domestic pressure to enforce standards that Orbán waived as sweeteners for Chinese investors.

• Fudan and telecom access face review. The Fudan campus project, already politically toxic, is likely dead. Huawei’s position will face new scrutiny aligned with broader EU security concerns.

• Supply chain positioning survives — for now. Chinese firms already operational in Hungary will adapt. Beijing’s commercial footprint is too large to unwind quickly, and Magyar needs the jobs and tax revenue. CATL’s Debrecen plant alone is expected to create around 9,000 jobs (Bloomberg).

Net assessment: China is a significant loser on diplomatic leverage and EU veto politics, a partial loser on investment access and regulatory environment, but retains its commercial footprint in the medium term. The real question is whether a Magyar government aligns Hungary with the EU’s emerging “de-risking” consensus on China — which would represent a structural, not merely tactical, shift.

What Orbán’s Defeat Means for Trump and the Global Right

The defeat is a genuine blow, and the White House made it worse by going all-in. JD Vance’s Budapest rally on 7 April — where he attacked “disgraceful” European Union interference in Hungary’s election (Reuters) and Trump promised to bring US “economic might” to Hungary if Fidesz won (Unruly / AFP) — turned the election into a referendum on American interference. Hungarian voters answered clearly. As Fox News reported, Hungarians turned out in record numbers despite Orbán being “a strong ally of Donald Trump and JD Vance” (Fox News).

Three implications for Trump-world:

First, the symbolic loss is real. Orbán was not merely an ally; he was the proof of concept — the leader who demonstrated that illiberal democracy could be built inside the EU, sustained through multiple elections, and exported as a model. CPAC Budapest, an overseas version of the most prominent right-wing political event in the US, was the movement’s ideological showcase (euronews). That showcase is now shuttered. As Newsweek asked: “Can MAGA Go Global If It Can’t Hold Hungary?” (Newsweek). Rahm Emanuel argued this marks “three straight losses for Trump-style politics in Europe: Slovenia, French local elections, and now Hungary” (Rahm Emanuel on X) — though this framing overstates the case. In Slovenia, the liberal incumbent beat Trump-allied Janez Janša by barely 1% in the closest election in the country’s history (Bloomberg), and in France’s municipal elections, the far-right National Rally fell short of key targets in Marseille, Toulon, and Nîmes but still made unprecedented gains (Le Monde). Hungary is the only unambiguous rout.

Second, the mechanism of defeat matters. Orbán did not lose on culture-war terrain. He lost because voters prioritised corruption, healthcare, cost of living, and institutional rot over migration fear and anti-woke messaging. As sociologist Andrea Szabo of ELTE University told AFP: “Fidesz decided to run a purely negative campaign... What they talked about was war, war, war” (Unruly / AFP). This is the vulnerability that Trump-aligned movements globally prefer not to discuss: culture wars work until the state visibly fails to deliver basic services.

Third, the exportable parts of the Orbán model survive. The playbook — media capture, judicial packing, electoral gerrymandering, oligarchic patronage networks — has already been studied and partially adopted by actors from Serbia to the US. Orbán’s shrewd tweaking of his country’s political system over 16 years created a model that “illiberal political leaders look up to... as a role model, who has made it, managed to take power,” as Emilia Palonen of the University of Helsinki told AFP (Unruly / AFP). The defeat of the practitioner does not erase the practice.

How Hard Will It Be to Fix the Institutions Orbán Remade?

This is the hardest question, and the answer is: harder than the celebratory mood in Brussels suggests. Orbán spent 16 years embedding loyalists, rewriting laws, and — as Politico documented in detail — laying deliberate “traps” for his successor. He appointed supporters to key state institutions to block budgets and laws (Politico). In December 2025, Hungary’s Fidesz-dominated parliament even passed a bill making it harder to unseat the head of state (DW).

• Constitution & cardinal laws

  • Medium (if supermajority holds) / Extreme (if it doesn’t) difficulty to reverse

  • Orbán’s 2011 Fundamental Law and dozens of cardinal laws (requiring two-thirds to amend) locked in everything from the structure of the judiciary to media regulation. Armed with a two-thirds majority in 2010, Orbán “implemented a root-and-branch reform of Hungarian state institutions and introduced a new constitution steeped in conservative values” (Unruly / AFP). A supermajority (133 of 199 seats) is the skeleton key. Without it, Magyar governs inside Orbán’s constitutional cage.

• Judiciary & Constitutional Court

  • High difficulty to reverse

  • Orbán packed the Constitutional Court with loyalists, including a former Fidesz defence minister (Politico). Judges serve long terms. Even with a supermajority, replacing sitting judges raises rule-of-law concerns that Brussels will scrutinise. Magyar has said courts must decide Orbán’s fate and that his government would focus on restoring judicial independence (TVP World).

• Media ecosystem

  • Extreme difficulty to reverse

  • The Central European Press and Media Foundation (KESMA) consolidated over 470 pro-Fidesz outlets into a single entity in 2018, exempted from antitrust review by government decree (MAPMF; New York Times). Reporters Without Borders (RSF) has classified Orbán as a “press freedom predator,” stating he has “nearly wiped out independent journalism” in Hungary (RSF). Ownership is opaque, legally entrenched, and intertwined with oligarchic networks. You cannot legislate an independent media ecosystem into existence.

• Procurement & oligarchic capture

  • High difficulty to reverse

  • Hungary is ranked the most corrupt country in the EU, together with Bulgaria, according to Transparency International’s 2025 Corruption Perceptions Index (Transparency International Hungary). Billions in EU and state contracts flow through Orbán-linked oligarchs, most notably Lőrinc Mészáros, a childhood friend of Orbán who became Hungary’s richest man (Forbes; Bloomberg). Unwinding these networks requires forensic auditing, legal proceedings, and political will.

• President of the Republic

  • High difficulty to reverse

  • Hungarian President Tamás Sulyok, who is close to Fidesz and will be in office until 2029, retains the power to call a snap election if the government can’t form (Politico). Parliament also recently reinforced the president’s position legislatively (DW). This is one of Orbán’s most potent institutional traps.

• Security services

  • High difficulty to reverse

  • Several whistleblowers from police, military and other state authorities came forward during the campaign to publicly accuse Orbán’s government of incompetence and influencing state institutions for political gain (Unruly / AFP). A growing list of Orbán loyalists began defecting before the election as his vulnerability became apparent (New York Times). Reforming these services without destabilising them is a delicate operation.

The European Policy Centre has argued that the EU should agree with a new Hungarian government on a “binding rule-of-law” framework, with institutional reforms phased rather than frontloaded (EPC). This is the right instinct. Magyar has promised to carry out reforms required to unfreeze billions of euros in EU funds earmarked for Hungary (Unruly / AFP), but the Orbán system was designed to survive its creator. As political scientist Attila Gyulai of ELTE University put it: Orbán “acted as a battering ram” so he could be the one who “wears out first” (Unruly / AFP).

6. Global Implications

EU decision-making transforms overnight. Hungary’s veto — wielded by Orbán to block sanctions on Russia, delay Ukraine aid, obstruct rule-of-law mechanisms, and shield China — is gone. European Commission President Ursula von der Leyen declared “Hungary has chosen Europe,” while European Parliament President Roberta Metsola said “Hungary’s place is at the heart of Europe” (Unruly / AFP). One European diplomat had told AFP before the vote, speaking on condition of anonymity: “Most member states would be quite happy to be rid of Orbán. The patience has worn very thin” (Unruly / AFP).

Ukraine gains a neighbour, not an adversary. Magyar has pledged to make Hungary a “reliable NATO ally” and dropped Orbán’s hostile rhetoric toward Kyiv (Unruly / AFP). He still opposes arms transfers and rapid Ukrainian EU accession — continuity that reflects genuine Hungarian public opinion — but Ukrainian President Volodymyr Zelensky congratulated Magyar, writing: “We are ready for meetings and constructive joint work in the interest of both nations, as well as for peace, security and stability in Europe” (Unruly / AFP). Ukrainian Prime Minister Ioulia Svyrydenko added that “the Hungarian people have said a clear and categorical ‘no’ to any attempt to bring their country back into Moscow’s orbit” (Unruly / AFP).

Russia loses its inside man. Orbán was Moscow’s most valuable asset in the EU — not as an agent, but as a veto-wielding disruptor who reliably slowed Western consensus. Leaked recorded phone conversations caused EU-wide alarm about Orbán’s and his foreign minister’s close relations with Moscow during the campaign, with Polish and Irish leaders condemning the links as “sinister” and “repulsive” (Unruly / AFP). The EU’s top diplomat Kaja Kallas was even more scathing, saying “European ministers should work for Europe not for Russia” (Unruly / AFP). Reports also claimed an ongoing covert Russian social media campaign to boost Orbán and weaken the opposition (Unruly / AFP). An Orbán loss was described by France 24 as “the turning point Putin fears” (France 24). Polish Prime Minister Donald Tusk’s victory message — “Ruszkik haza!” (Russians go home!) — captured the mood across Central Europe (Unruly / AFP).

The democratic-backsliding narrative shifts. For a decade, Hungary was Exhibit A in the global story of democratic erosion. The European Parliament had denounced Orbán’s Hungary as a “hybrid regime of electoral autocracy” (BBC). Its reversal — through elections, with record turnout, despite a rigged system — provides a counter-narrative. As Andrea Szabo, a senior research fellow at ELTE University, told AFP before the vote: “This is the last moment in which this process can be halted, and the pendulum can swing back in a democratic direction” (Unruly / AFP).

7. Winners and Losers

Winners

• The European Union — regains Council cohesion, loses its most persistent veto blocker. German Chancellor Friedrich Merz wrote: “Let us join forces for a strong, safe and above all united Europe” (Unruly / AFP). French President Emmanuel Macron hailed “a victory for democratic turnout, the Hungarian people’s attachment to European Union values, and for Hungary in Europe” (Unruly / AFP). Spanish PM Pedro Sánchez declared: “Today, Europe and European values won” (Unruly / AFP).

• Ukraine — loses an obstructionist neighbour, gains a non-hostile one. Zelensky congratulated Magyar and signalled readiness for “constructive joint work” (Unruly / AFP).

• Poland’s Tusk government — gains a regional ally in the democratic-restoration camp. Tusk hailed “a glorious victory” (Unruly / AFP).

• Hungarian civil society and independent media — the groups Orbán vowed to crush (”fake civil society organisations, bought journalists, judges and politicians”) gain political space (Unruly / AFP).

• UK Prime Minister Keir Starmer — called it “an historic moment, not only for Hungary, but for European democracy” (Unruly / AFP).

• Giorgia Meloni — paradoxically, Meloni emerges repositioned. She congratulated Magyar for his “clear election victory” while thanking “my friend Viktor Orbán for the intense collaboration of these last years,” adding she was “certain that we will continue to collaborate constructively” (Unruly / AFP). She is now the undisputed leader of the European right’s illiberal-adjacent wing, without the baggage of Orbán’s defeat.

Losers

• Russia — loses its most effective EU disruptor. Evidence of Russian interference in Hungary’s election had been growing throughout the campaign, including reports of a covert Russian social media campaign to boost Orbán (Unruly / AFP), and it failed.

• China — loses privileged access and EU veto cover (SCMP).

• Trump-world — loses its proof-of-concept and ideological capital. The question of whether the Vance rally helped or hindered Orbán was raised even before the vote (Irish Times).

• Fidesz-linked oligarchs — face potential audits, procurement reform, and loss of political protection. Lőrinc Mészáros, Orbán’s childhood friend who became Hungary’s richest man through state contracts (Forbes), is among those most exposed.

• Orbán personally — moves from the most powerful figure in Central European politics to opposition leader. He had “transformed his small central European country from a burgeoning Western democracy into an illiberal one” (CNN) — and voters reversed it. Born in 1963, he first became prime minister in 1998 at just 35 — the second youngest in Hungarian history (Wikipedia). He returned to power in 2010 and held it for 16 years.

• The “illiberal democracy” brand — the term Orbán used in his infamous 2014 speech at Băile Tușnad, where he declared his government was building an “illiberal state” within the EU (Heinrich Böll Foundation), loses its only successful long-term practitioner inside the EU.

8. The Magyar Question: Can the Insider Deliver System Change?

The biggest risk to the democratic-restoration narrative is Magyar himself. As analyst Andrzej Sadecki of the Warsaw-based Centre for Eastern Studies told AFP: “In a way, Magyar is like Orbán 20 years ago without all the baggage, the corruption and the mistakes made in power” (Unruly / AFP). But that cuts both ways.

Magyar was born on 16 March 1981 (Wikipedia), making him 45 at the time of the election. He was born into a family of prominent conservatives, befriended Orbán’s current chief of staff Gergely Gulyás during university, married Judit Varga in 2006 (she later became Orbán’s justice minister), headed the state’s student loan provider, and sat on the board of multiple state companies (Unruly / AFP). He has even stricter anti-immigration views than Orbán, pledging to end the government’s guest worker programme. His stance on LGBTQ rights is vague, though he emphasises equality before the law. He rejects sending arms to Ukraine and opposes the country’s quick EU integration (Unruly / AFP).

“As he was socialised in Fidesz, there are also doubts whether he can provide a genuine rupture with Orbán’s rule,” Sadecki cautioned. “Left-wing voters might not be fully happy with his agenda, but they still support him, because he represents the biggest chance for change” (Unruly / AFP).

9. Scenarios for the Next 12 Months

Scenario 1: Democratic Restoration Accelerates

Probability: 25%

Magyar secures a confirmed supermajority (133+ seats), moves swiftly to amend cardinal laws, begins judicial and media reform, unfreezes EU funds, and aligns Hungary with mainstream EU positions. Orbán’s institutional traps are dismantled within 12 months.

Triggers: Final seat count confirms supermajority (current projections suggest ~138 seats) (TVP World); EU provides structured support and conditionality framework (EPC); Fidesz fragments in opposition, as loyalist defections accelerate (New York Times).

Scenario 2: Partial Reform, Partial Continuity (Base Case)

Probability: 55%

Magyar governs with a supermajority but faces fierce resistance from Orbán-embedded institutions — the Constitutional Court, the Media Council, the State Audit Office, the presidency. Reform proceeds unevenly: EU funds are partially unfrozen, foreign policy realigns, but media and judicial reform stall. Magyar’s own Fidesz socialisation produces policy continuity on migration and some social issues. The oligarchic networks adapt rather than collapse.

Triggers: Institutional resistance from Orbán loyalists in fixed-term positions (Politico); legal challenges to reform legislation; Magyar’s own conservative instincts limit the scope of change.

Scenario 3: Institutional Sabotage and Backlash

Probability: 20%

Orbán-aligned institutions — particularly the president (in office until 2029), the Constitutional Court, and the State Audit Office — actively obstruct Magyar’s government. Budget vetoes, constitutional challenges, and potential snap-election triggers create governance paralysis. Fidesz regroups around a narrative of illegitimate foreign-backed regime change.

Triggers: President Sulyok blocks key legislation or triggers constitutional crisis (Politico); Constitutional Court strikes down reform laws; economic downturn blamed on new government.

10. Key Judgments and Confidence Levels

1. Orbán’s defeat is genuine and irreversible as an electoral outcome. He conceded and congratulated the winning party (Unruly / AFP). He will not return to power in this parliamentary term. — High confidence

2. Tisza will secure a two-thirds supermajority, based on projections of ~138 seats with 90.9% of votes counted (Europe Elects; TVP World). — Moderate-High confidence (final count pending; individual constituency results could shift the total)

3. The supermajority is necessary but not sufficient for institutional reform. Orbán’s traps — loyalist appointments, the presidency, the Constitutional Court — will slow and complicate reform regardless (Politico). — High confidence

4. China’s diplomatic leverage in the EU Council is materially reduced. Beijing loses its most reliable veto partner on trade, investment screening, and foreign policy (SCMP). — High confidence

5. China’s commercial footprint in Hungary will persist in the medium term, even as the regulatory and political environment tightens. CATL’s €7.3bn Debrecen plant (Reuters) and BYD’s ~€4bn Szeged factory (Reuters) represent too much sunk investment to unwind quickly. — Moderate confidence

6. The Vance rally was a net negative for both Orbán and Trump. It allowed Magyar to frame the election as a choice between Hungarian sovereignty and foreign interference — ironically inverting Orbán’s own playbook (Irish Times). — Moderate confidence

7. Magyar will align Hungary with mainstream EU foreign policy positions on Russia and Ukraine within 90 days, but will maintain opposition to arms transfers to Ukraine (Unruly / AFP). — Moderate confidence

8. Media reform will be the hardest institutional domain to address, given the opacity of ownership structures and the legal entrenchment of KESMA, which consolidated over 470 pro-Fidesz outlets (MAPMF; New York Times). RSF has described Orbán as having “nearly wiped out independent journalism” in Hungary (RSF). — High confidence

9. The Orbán system was designed to survive its creator. As political scientist Gyulai noted, Orbán “managed to build up the political system around himself” so that “all policy issues, ideological preferences, socio-cultural perceptions culminate in one referendum-like question: do you want Viktor Orbán? Yes or no?” (Unruly / AFP). The next 12 months will test whether that design holds against a supermajority and EU pressure. — High confidence

10. Orbán’s defeat does not signal the death of right-wing populism in Europe. Meloni remains powerful in Italy; the far right made gains (though fell short of key targets) in French municipal elections (Le Monde); and Janša nearly won in Slovenia (Bloomberg). The lesson other populists will draw is operational — avoid visible corruption, maintain service delivery — not ideological. — High confidence

This analysis reflects information available as of 13 April 2026, the day after the Hungarian parliamentary election. Final official results are expected later this week. All seat projections are based on partial counts (90.9% of list votes) and Europe Elects projections.

Forecast Trump in 10 days

I was given a hard deadline by my publisher of no further edits come November 15, 2024. That meant that I would have 10 days to update the book, including a chapter about America, following the 2024 presidential election. When Trump won on November 5th, I faced the prospect of having to commit my gut instincts into the permanence of a book that I hoped would be read for years.

Many of those gut reactions were correct. I wrote of eliminating federal oversight and regulation on nearly everything. I wrote of tariffs moving from the technical and legal channels into a political football that would result in extra high levels across the board. I wrote about public support for Trump’s deportation plans and how this would permit enforcement tactics like what Trump ultimately did with ICE. I wrote of politicized courts that would do Trump’s bidding.

But one of my other contentions was terribly wrong.

My big miss

After it was clear Trump had won, I went back through the book and started to add Trump as a punctuation to elements which his election validated. And I leaned on what I knew about Trump to predict the future.

I noted that “Trump was elected again in 2024 on similar promises of even bigger, unapologetic tariffs and the implementation of a new isolationism.” I noted that the US was losing its desire to act as the world’s policeman, “a development that is likely to be reinforced by the isolationist instincts of the Trump administration.”

But, I also wrote “The second Trump administration will revert to a more isolationist foreign policy, as it did in his first term. For all of Trump’s tough talk, he engaged in the fewest military actions of any U.S. president in decades.”

Now, there’s no doubt Trump has been isolationist. He has defunded pretty much every international organization. He has eliminated foreign aid. He keeps the sheer existence of NATO an open question that is raised nearly every week.

But, the implication of the point about few military actions in his first term was that this would be the norm in his second term. Clearly, that has not been the case.

Why I got this wrong

Trump had won in 2016 after building an image of being ardently against the 2003 Iraq war. As I noted, his first term saw virtually no international intervention and instead a focus on holding the line against international commitments. In 2024, Trump campaigned on ending the “stupid” war in Ukraine “in 24 hours”. He picked a Vice President who famously didn’t care what happened to Ukraine. And he focused all of his campaign energy on signature domestic issues.

Thus, I made a few mistakes. First, I conflated isolationism with non-intervention. Yes, the US was pulling back from its role as globalization police force. But not intervening militarily is very different.

I probably did this because both had aligned in the politics of Trump’s first term. The population was sick of foreign wars and of foreign aid / support. Foreign countries feared Trump and decided not to antagonize him and instead wait him out. Thus, his lack of military intervention was a function of external conditions too, not necessarily a world view.

Which brings me to the second analytical error: Path dependency. It didn’t seem that foolish to expect Trump to act in his second term as he did in his first term, except that expecting consistency from Trump is probably foolish to some extent. Or, less charitably, I evaluated him within the typical parameters of historical presidents, not someone that would wholesale rethink a foreign policy posture because a brand new approach would suit him politically and maybe financially.

Thus, instead of understanding Trump as a transactionalist who played power politics, I was looking at his actions in the first term as representing a world view. In fact, it was more about the politics of the moment than about his own red lines.

Finally, I saw dots when I should have seen a line related to the Middle East. In his first term, Trump assassinated Iran’s most powerful military commander. He struck Syria more than once. He conducted large scale drone strikes. He did little else militarily, but taken together this showed a willingness to break precedent to counter threats in the Middle East. While the narrative about Trump’s first term was about avoiding foreign entanglements, there was some evidence to the contrary.

Could AI have done better?

AI is not better at predicting the future but it is better at abandoning the past. And if you reframe the game from “being right” to “quickly realizing you are wrong” then I would expect AI to outperform humans.

AI that focuses on news narrative might have bought the “Trump is isolationist” narrative; indeed, Claude comforts me by telling me:

In practice, most LLMs (including me, honestly) would have made the same mistake because we’re trained on text that predominantly labeled Trump as isolationist, which reflects how journalists and analysts talked about him rather than what he actually did...LLMs in late 2024 were actively reinforcing the isolationist frame because that’s how most of their training data described Trump. So on this specific failure mode, AI probably would have been as blind as you were.

ChatGPT tells me it would have made the same mistake because “conflating isolationism with non-intervention was not a human error - it’s a data problem.” GPT notes that its training data labeled Trump as “isolationist”, “anti-war” and “against foreign entanglements”, which becomes compressed to “less likely to use force.”

AI also has trouble with out of the ordinary scenarios. When we were publishing our flagship Iran report a couple weeks ago, we uploaded it to Claude Code which replied:

The data covers a (fictional/scenario) world where a US-Israel war with Iran has broken out, the Strait of Hormuz is closed, and various cascading geopolitical effects are unfolding.

When we asked why Claude thought it was hypothetical, Claude explained:

A few things triggered it:

1. The dates are in the future — March 2026 is still in the future from my training data perspective, and today is March 17, 2026. So I was reading detailed reports about events supposedly happening right now or very recently, which pattern-matched to “scenario” rather than “reporting.”

2. The escalation speed — Khamenei killed on Feb 28, Strait of Hormuz closed, 10+ tankers hit, Gulf states under missile attack, IEA releasing 400M barrels, oil threatening $200 — all within ~12 days. That’s an extremely compressed escalation timeline that felt more like a stress-test scenario than how events typically unfold.

So AI would have started 2025 from the same position as I did. And we can get a sense that it would have trouble developing scenarios about what is underway in such an unruly world because it is simply hard to believe. However, AI wouldn’t have to commit that in the form of a book.

As a result, AI could then be asked to monitor for signals of divergence from the core contention. By March 2025, Trump discussed militarily invading Greenland. By June 2025, Trump had struck underground nuclear facilities in Iran, and saw that Iran’s military response was limited. By mid-2025, Trump was shooting missiles at Venezuelan boats, and moved an aircraft carrier nearby in October. As a result, it would be clear to human analysts and AI that Trump was not a non-interventionist.

And, yet, even with all the evidence, human analysts still suffer confirmation bias. Few warned that a Venezuelan operation was likely or that a war on the scale of what we see in the Gulf was potentially going to occur. AI would likely have done better at helping us understand that probabilities were recalibrating even if it didn’t make an accurate point forecast.

Where AI wins

AI has a few advantages over humans:

• AI holds probabilities, not positions: Humans defend positions while AI tracks probability distributions. If you want to get a sense of what’s possible, probabilistic thinking is much more helpful.

• AI is better at separating labels from behaviors: Both humans and AI might shortcut to the heuristic of Trump as “isolationist” but AI can quickly update for changed behaviors while humans ask if the behavior has moved enough to change the label.

• AI treats every day like Groundhog Day: Humans commit to views and wait to be proven wrong. AI is can be told to think “if I had no prior view, what would I conclude today?”

• AI cares about adaptation, not consistency: Humans want to be coherent and story-tell. AI can hold contradicting evidence in its system and continuously update to make conclusions. It also doesn’t care if it was wrong yesterday, which can often stop a human from forecasting again.

Warning vs Prediction

Thus, if you think about AI less as a predictive tool and more as a warning tool, you can harness it to get leverage over what’s actually happening in the world and what it means.

Any horizon scanning software, like that which we’ve built at Unruly, can pick up when the game is changing and help you prepare.

Human authors are left explaining why they published a wrong sentence a year ago.

-SW

Yet, in the hours following the first bombs, most companies were focused on what it actually meant to their businesses today. Companies have sprawling global supply chains. They have employees sitting in the Emirates airport lounge. They are enmeshed in the middle of churning bilateral relationships. Whether the war ends sooner or later, they needed to figure out their existing exposure. Quickly.

Companies without a dedicated geopolitical risk budget turned to Google News, GPT, or maybe an Economist or FT subscription to get a sense of ramifications. These are affordable, but not able to provide true insights related to their specific business.

Companies with larger budgets turned to a growing number of boutique geopolitical risk research firms. While these firms produce high quality analytical research, it is delivered slowly and rarely iterates at the speed of business decisions or events unfolding. While custom insights can be had with expert phone calls, those insights are out of date with the next headline.

Could AI, instead, give these companies swift, dynamic insights?

Understanding a war - quickly and dynamically

It is into this space that Unruly Corp’s Global Impact of the Iran Conflict Report fits. We have spent a number of months building a system that can size up global events from every perspective and write like an analyst would. So, we unleashed our AI agents to draft a 150 page report that analyzed the impact of the Iran conflict on 17 countries, sectors, and themes across different areas like economics, security, and politics, as well as scenarios and specific implications. The report created specific scenarios for each region, sector or theme as well as detailed signposts and implications.

Suddenly, even the most minor implications of the war were visible. While analysts might focus on a spike in oil prices due to closure of the Strait of Hormuz, more than 40% of the world’s seaborne sulphur and urea would also be at risk. This wouldn’t immediately be obvious to a regional analyst, but AI seamlessly highlighted how this would be a catastrophe for agricultural firms and those upstream from them. There are many other examples.

The first turn of the report took less time than a research house would spend in a planning meeting to scope out how to produce a comparable report and cost less in compute than the coffee and donuts for that meeting.

A sample of the TOC for our global Iran report, which you can access at iran.unrulycorp.com

We got quite positive feedback on the report’s breadth, analytical insights, and ability to spin up unique scenarios from a variety of perspectives. Fortune 500 executives liked having everything in a single PDF that was indexed so they could navigate to what they wanted. Others were happy to put the PDF through their own AI agents to digest implications for themselves.

However, we also got answers like:

“There is zero chance I can read a PDF of 150 pages when the world is at war! 😆”

So the following week we created (and expanded) the report while also deploying it as an app for easy consumption. We added entity linking (i.e. searching for a mentioned company or group) to let people get directly to the insights they need. We made the app free to try to help those making sense of the war; you can access in full here here.

The research was an analysis not of the actual conflict itself–there is plenty of high quality daily reporting on that–but rather on the widespread global impact of the conflict that much analysis doesn’t have the bandwidth or time to cover. Perhaps there is coverage of the impact on energy markets and major countries like India, China, the Gulf states, and Europe, but what about the impact on Latin America or sub-Saharan Africa? Or the defense industry? Or cyber? Perhaps in ad hoc ways, but never holistically.

Full Visibility

Why aren’t there more bespoke reports like this? The answer is largely due to cost and the sheer time it would take human analysts to do it. Whatever the real analytical shortcomings of AI-generated geopolitical analysis, its benefits include not only vastly reduced cost but also unconstrained bandwidth to produce whatever analysis is needed at whatever moment for whatever use case.

If everyone’s money-is-no-object first choice is to commission a team of the world’s greatest geopolitical experts to drop everything and write them a custom report on Iran and its impact on an unlimited array of countries, regions, industries and themes, a really good second choice is to have an AI system produce a report that is nearly as good for a fraction of the cost. And then produce an update to that report the next hour, day or the next week. And then customize it for a particular client’s exposure. And then do it all again for the next global development.

Indeed, the fact that there is very little marginal cost to successive insights allows us to explore angles in the first instance that might not be explored for weeks by human analysts focusing first on where all the money is. For instance, if you are focused on Southeast Asia, you know that the closure of the Strait of Hormuz will cause an energy supply crunch, but are there firms putting out research about the Iran conflict from the perspective of Southeast Asia? Or on the metals and mining industry? Or food prices in East Africa? Or more specifically, all the ways that the crisis is affecting shipping giant Maersk or the implications for cybersecurity firm Crowdstrike?

Such analysis is unlikely to be proactively created and constantly updated by humans because the time and cost involved isn’t worth the narrow demand. AI geopolitical analysis fundamentally changes that equation around what is possible and economical.

Share

Cost+Speed

The obvious savings here is cost, but the far more valuable savings here is time. Cost reduction is always appealing, but given the unruliness of today’s world, cost reduction essentially only reduces cost, whereas speed increases upside opportunity.

Across industries, speed can only be increased to a point: freight can’t go faster than airplanes fly, legislation moves at the pace that it does, and analysts and editors produce analysis at the speed they do.

So, can a human-powered firm provide a report about the full global impact of the Iran conflict for me tomorrow? For most firms there is no price that can make that possible. But leveraging AI-generated analysis not only makes that possible, it then also opens up the possibilities of having those insights faster. The marginal cost of another report about another angle on another issue is obviously quite low; but the ability to fast forward in time to get those insights faster is the thing that will transform how businesses use geopolitical analysis to defend against fast moving risks and create opportunities among chaos.

In a quickly changing world, speed is a force multiplier. Analyses, topics, and insights need to evolve as quickly as events morph and as quickly as business opportunities manifest. Each morning a business leader can now have an in-depth report on the global implications of the Iran war on key geographic locations and industries of critical interest. But then the real trick is that there could be another updated report that is focused on a different narrow issue and customized for a particular executive by lunch as circumstances change.

Whether done with our software or constructed in a foundational LLM, this opens up a fundamentally new way to integrate geopolitical analysis, more akin to instant market data to make financial decisions than a weekly research push. It allows rigorous analysis to be disseminated quickly and at a low cost, which in turn opens the possibility of interactive streaming of geopolitical insight tuned to particular exposures and opportunities.

Speed also shifts expectations with regards to prediction. Yes, the holy grail of geopolitical analysis is to know what’s coming next, but explaining what happened yesterday is actually pretty hard too. With AI, we’re able to solve the yesterday problem at scale, which allows us to understand more of what’s happening today - thus, we can respond in real-time much better.

While we will keep innovating new and different ways to cover global crises, the real question is how your business will change when fast, flexible geopolitical insight is available through any lens, on any timeline, tuned to your specific exposures. More on that in future posts.

-SW & DB

After 68 issues of GeoLegal Notes, the writing and launching of my book Unruly: Fighting Back when Politics, AI and Law Upend the Rules of Business, and the spin-out of our geopolitical AI company into The Unruly Corporation, I took a haitus from writing.

Some of it was motivated by intellectual exhaustion: I felt like I had covered the heck out of GeoLegal risk, warning at every twist and turn what could be ahead.

But a bigger part of it was that we woke up in an Unruly nightmare of ever grander proportions than I could have imagined. Yes, my book and Substacks warned about an erosion of rule of law, about the weaponization of technology, and about politics that would manipulated by AI. But suddenly we were living in a world where it was all really happening. And at scale.

As a result, mainstream media had picked up the torch to write about geolegal risk and other forms of unruliness in real-time because it was our everyday life. Piling on might get clicks, likes, and quotes in the press but that’s not what people need today.

What business leaders really need today are tools to actually manage the storm around them. One day that storm is in Venezuela. The next it is the entire Gulf region. Tomorrow it may be in Cuba. And then Taiwan.

Subscribe now

We are lucky because at the precise moment the world has spun out of control, AI technology has armed us with ways to track, act and potentially anticipate. That’s the really interesting story.

So, what I am going to do here is catalog the journey of those building the future of AI to understand and manage geopolitical risk. You will learn a huge amount about the substantive risks the world is facing and about the challenges we face in using technology to wrangle it.

If you are a CEO, head of strategy, head of corporate affairs, head of government affairs, head of threat intelligence, general counsel or many other roles you need to be aware right now of the advances that AI is making to support your business. Or else you will remain exposed and vulnerable while your competitors gain advantage.

Let me be clear: I am building in the space and I am proud of what we are building at The Unruly Corporation. So, yes, I will track our own experimentation and output, giving a platform to my colleagues to talk about how they solve extremely cool geopolitical challenges in real-time.

BUT, I have a very big tent philosophy when it comes to preparing companies for political risk. The more builders in the industry working on solutions, the better that is for client readiness. And, if more clients choose to transform, that means an even more fertile ground for everyone building tools to help them.

So, I am making two commitments here.

First, what you read here will not be bland content marketing. It will continue to be thought provoking approaches to how we manage the world around us, experimenting in plain view.

Second, I want to interview and shine a light on founders in the space who are doing amazing things that you may never have heard of. In the process, I’ll build out the market map for geopolitical AI which can help you with choosing who you want to work with.

Share

If you were a subscriber to GeoLegal Notes, I added you here. If you want off this bus, just pull the cord. If you’re glad to be here, tell a friend.

-SW