No company operates in a vacuum: energy costs, the policy rate, tariffs and conflicts hang together like a spider's web. Pull on one thread and the whole web moves. Two independent maps show the situation in July 2026: the company network (how shocks travel all the way into the cost structure) and the world-finance network (euro, dollar, yen, government debt, oil, gold, and who depends on whom). Plus a section on context: the power of liquidity and historical parallels with 1914 and the 1930s. Click on nodes for the current situation, pros & cons.
Outer ring: actors · inner ring: transmission channels · centre: the company. Arrows show the direction of effect; the colour indicates whether a relationship currently strains or supports.
The strait did not reopen. On 11 August ten ships crossed the Strait of Hormuz, against roughly 130 on a normal day, and between seven and nine million barrels a day were still moving where twenty used to. Qatar described the talks as advanced; Tehran's position is that the strait stays shut until war reparations are paid and sanctions lifted. On 30 August American forces struck missile sites on Larak Island, the first attack in weeks. Iran answered with missiles and drones against the Jordanian air bases King Hussein and Al Azraq, eight of which Jordan intercepted, and on 31 August shot down an American MQ-9 over the strait. Both crude benchmarks jumped about three percent in a session. Brent stood at 97.19 dollars on 7 September, 47 percent above a year earlier.
The energy bill has reached the European inflation figure. Euro area inflation rose to 3.3 percent in August from 2.9 in July, and the driver is unambiguous: energy at 14.3 percent year on year against 10.3 in July. Austria came in at 3.2 percent, energy up ten percent, services up 4.3, food held down to 0.2 by the VAT cut. Services are decelerating. Energy is not.
Three central banks decide within nine days. The ECB meets on 10 September with its rate at 2.40 percent, the Fed on 16 September at 3.75 percent, the Bank of Japan on 18 September. At Jackson Hole on 31 August the Fed chair said underlying inflation was not slowing, and market implied odds of a September increase moved from about 40 percent into the high fifties, before a governor talked them back on 3 September. The ten year Treasury yield touched 4.818 percent on 2 September, the highest since November 2023, and gold sits near 4,400 dollars, a fifth above a year ago.
The yen is unwinding, which is the part worth watching. With a larger than usual Japanese rate rise expected on 18 September, the yen gained more than two percent in a single session on 4 September and stands near 154 to the dollar against 162 earlier in the year. That is the carry trade running backwards: positions funded cheaply in yen get closed, and the selling lands in whatever was bought with the money, which is rarely Japanese. This is the one channel on this map that transmits without any political decision anywhere.
Ukraine: three days, and not more than that. Steve Witkoff and Jared Kushner met Putin in Moscow on 5 September for three hours and were in Kyiv the next day. Russia undertook not to strike Kyiv from Saturday midnight to Monday, Ukraine not to strike Moscow. The Kremlin called the meeting constructive. The Russian position on the remainder of eastern Ukraine has not moved, and nothing in the reporting touches the energy sanctions.
And the war moved into civil aviation. On 2 September Ukraine formally notified ICAO that Russian airspace is not safe for civilian flights, after President Zelenskyy had said the day before that Ukrainian drone operations were expanding across Russian territory. Vnukovo in Moscow was disrupted, Pegasus cancelled twelve flights on 3 September and Turkish Airlines four movements, and one flight from Antalya was diverted. Whatever one makes of the tactic, the effect on this map is direct: another corridor that freight and business travel can no longer count on.
The American tariff base changed shape. The ten percent Section 122 surcharge expired on 24 July at its statutory limit of 150 days and was replaced the same day under Section 301, which has neither a rate cap nor a time limit: ten to 12.5 percent for roughly sixty economies covering 99.4 percent of American imports. EU goods sit outside that action and stay on the negotiated 15 percent all-inclusive ceiling in force since 1 July. From 1 September aircraft and parts, generic pharmaceuticals and their precursors moved to the plain MFN rate, which for most of them is zero. Steel, aluminium and copper remain at 50 percent.
China. The rare earth controls that Trump and Xi agreed away in late October 2025 were suspended for one year, so they fall due for decision again this autumn. The suspension never meant equal treatment: European magnet imports rose about 60 percent year on year in November 2025 while American imports fell 11 percent, and yttrium shipments to the United States ran at 17 tonnes over nine months against 333 tonnes in the preceding eight. Washington has set a floor price of 110 dollars a kilogram to make domestic production viable. Chinese growth for 2026 is put at 4.8 percent, carried by exports, with producer prices still negative and property still falling.
The circle of AI money got larger, not smaller. Reporting in August put Nvidia's own data centre commitments at up to 750 billion dollars, and hyperscaler capital spending for 2026 at 775 to 800 billion. The pattern the fifth view of this chapter describes is unchanged: chip makers invest in the firms that then buy their chips. Nvidia's chief executive rejects that description of it. Michael Burry is among those on the other side. Nothing that happened in August settles the question either way.
Every figure in these two cards carries a date because most of them will have moved by the time this is read. Sources in the footer.
The path is short and it is written down. The EU borrows on the capital markets against its own budget headroom and lends the money to Ukraine: the Ukraine Support Loan of 90 billion euro, finalised by the Council on 23 April 2026, of which 30 billion is macroeconomic support and 60 billion is for defence industrial capacity and procurement. For 2026 alone, 45 billion was released, 28.3 billion of it for industrial capacity. Ukraine signs contracts with arms manufacturers, submits them to the Commission, the Commission checks them against the agreed procurement plan and only then does the money move. On 24 August a further 6.1 billion was approved for air and missile defence, missiles, ammunition and radar, on top of 16 billion already cleared, of which 8.35 billion had been paid out.
The Commission's own wording is that most of these purchases are to be sourced from EU defence companies. So the money is raised in Europe, routed through Kyiv and largely spent back in Europe. Germany plans more than 11.5 billion for 2026 and states openly that part of it is industrial replenishment, so that Rheinmetall, Krauss-Maffei Wegmann and Diehl can keep delivering without emptying Bundeswehr stocks. Rheinmetall's order backlog stood at 80.4 billion euro at the end of the second quarter.
Two corrections to how this is usually described. It is a loan, not aid, and the repayment is meant to come from reparations owed by Russia to Ukraine. If those reparations are never collected, the bill sits with the EU budget, which means the member states, which means Austria among them. And the money does not have to go to Western industry at all: the Danish model pays Ukrainian manufacturers directly, more than two billion dollars by the end of 2025, up from 538 million in 2024. Ukraine's own industry is estimated to be able to build 35 to 40 billion dollars of equipment a year, and a significant part of that capacity sits uncontracted for lack of funding.
Put the pieces side by side and a suspicion follows. A European industry with an order book of that size did not have one four years ago. The state that funds the aid is also the state whose companies receive the contracts, and the mechanism that checks the contracts is the same institution that releases the money. A war that ended would end the block of demand that currently carries several industrial regions, at the very moment when their civilian business is shrinking. On this reading it is not necessary for anyone to want the war: it is enough that nobody in the chain has an interest in it stopping. This is an interpretation, and it is not an unreasonable one.
A beneficiary is not a cause. Three facts cut against the strong version. First, aid is falling, not rising: the Kiel Institute puts military assistance at just over two billion euro a month in 2026, roughly 500 million below earlier years. An actor steering towards continuation would not be reducing the flow. Second, governments cancel defence programmes when it suits them: Germany halted the F-126 frigate, which cost Rheinmetall 300 million of its own revenue guidance. Third, the backlog is mostly not Ukraine. It is NATO rearmament and national procurement, which continue whether or not this war does, and which were triggered by the invasion rather than sustained by it.
What is left. A real and documented circuit, a real concentration of interest, and no evidence of a decision to prolong. That distinction is worth holding: an incentive that exists is not the same as a motive that was acted on, and the same sentence would have to be applied to every actor on this map, including the ones the reading is aimed at.
Why rising energy costs hit Austria and Germany at the core, not at the margin.
For decades Russian pipeline gas flowed west under long-term, oil-indexed contracts at prices far below today's world market, mostly around €10 to €25 per MWh and thus close to its production cost. Chemicals, fertiliser, steel, glass, paper and ceramics in Germany and Austria built their cost structures on exactly this input; Austria still drew over 80% of its gas from Russia in 2024, and the OMV long-term contract ended only in December 2024. Cheap gas was not one factor among many. It was the quiet subsidy under the whole export model.
Since 2022 the pipeline volumes are largely gone, and since the 2026 Hormuz crisis the replacement is expensive: LNG sets the European price. The chart shows the gap that has opened between the EU and the USA. Every euro of that gap is a location handicap that a competitor in Texas or Louisiana simply does not pay, and with rising energy costs the old constellation falls apart.
Annual averages, 2026 = level of 26 August; the 2022 spike peaked around €340/MWh in August. Simplified teaching chart; sources in the footer.
A stylised energy-intensive company. Assumptions: no short-term pass-through; electricity follows gas via the merit order (→ Price & balance).
The complete network as a table, every edge with its mechanism. (Accessible alternative to the graphic.)
Outer ring: central banks, states & conflict regions · inner ring: prices and flows · centre: the global economy. Click nodes for details.
The same neighbourhood, three monetary models. Austria has transferred its interest-rate sovereignty to Frankfurt; Switzerland and Czechia decide for themselves, with very different results.
Takeaway: a currency of your own buys flexibility and costs an interest premium; the euro buys stability and costs sovereignty. Which deal is better depends on a country's trade structure, discipline and size.
The world-finance network as a table, an accessible alternative to the graphic.
The most important goods flows and their current rates. What is special about 2026: in February the Supreme Court struck down the blanket IEEPA tariffs; since then EU goods have faced a surcharge of +10% (Section 122, time-limited), while the sectoral Section 232 tariffs (steel, aluminium, copper, cars) remained in place. The legal basis is shaky, but the tariffs stay.
The US importer remits the tariff, but the burden is spread across three valves: price (paid by the US customer), margin (paid by the manufacturer) and restructuring (which turns into fixed costs). Set the pass-through:
Assumptions: 15% tariff (cars), price elasticity −1.2 (→ Costs & leverage). Valve 3, localisation (US plants instead of exports), avoids the tariff but ties up capital and raises fixed costs for years. VW currently carries a tariff burden of ≈ €5bn per year and is therefore building in North Carolina.
The US is Austria's second most important export market, and tariff policy is already hitting with full force.
Tariffs are not only a Washington phenomenon. Since 2026 the EU's carbon border adjustment has had teeth as well.
Takeaway: a tariff is rarely a victory over foreign countries. Mostly it is a redistribution at home: from consumers to the protected industry and to the state.
On 4 September 2026 the supervisory board approved unanimously the largest restructuring in the company's history: 50,000 further posts, which with the reductions agreed since late 2024 comes to about 100,000, up to a quarter of management positions, 125 billion euro of investment and 11 billion of cost savings. Roughly one employee in six is affected. Four sites received an end date for vehicle production rather than a closure: Emden and Zwickau in 2031, Hannover in 2032, Neckarsulm in 2034. The board's own wording is that closures are not inevitable. Osnabrück took a different route: the last T-Roc leaves in the summer of 2027 and the site passes to a defence venture with the Israeli manufacturer Rafael and the investor Aurelius, with Lower Saxony taking a stake, about 1,400 staff transferring and employment secured to the end of 2029.
The background sits in the half year figures of 24 July: revenue 158.1 billion, flat on the year, operating profit 5.9 billion, down 11.6 percent, margin 3.8 percent, group deliveries down 8.4 percent to four million vehicles, and China down 31.6 percent. Half a billion was written off for ending ID.4 production in the United States. This is the company at the other end of the arrow that runs from Wolfsburg to Graz.
An end date in 2031 for a plant whose product does not carry its cost today is not a decision to close, it is a decision to postpone. Until then the capital stays tied up, the cost base stays in place, and the margin that would have funded the transition is the one being consumed. On this reading German co-determination produces the same outcome every time: everyone signs, nobody closes, and the group buys peace at the price of the years in which it could still have acted. A competitor with a free hand does in eighteen months what takes eight years here. This is an interpretation, and a common one among analysts.
The figures do not describe paralysis. A hundred thousand posts, a quarter of management and eleven billion of savings is not a company refusing to act, and 125 billion of investment is not a company that only cuts. The long dates buy something real: phased exits avoid the severance and conflict cost of an abrupt closure and keep the workforce the transition needs. Osnabrück shows the option a quick closure would have destroyed, a site converted rather than shut.
What is left. A group reducing its home cost base over eight years while its most profitable market shrinks in months. China fell by almost a third in a single half year, and no closure date in Lower Saxony addresses that. Whether the slow path is prudence or delay depends on how long the China decline runs, and that is the number to watch rather than the closure dates.
Russia has turned sanctions into a new business model. Its vast oil and gas resources now flow east at a discount: China and India buy the bulk of the crude, a shadow fleet ships it, and part of the trade settles in yuan. The loop closes in plain sight: Indian refineries process Urals crude into diesel and jet fuel, and in 2024 about 13% of the EU's seaborne diesel and jet imports came from India. The Russian origin is simply refined away. Since 21 January 2026 the EU bans products made from Russian crude even in third countries, but molecules carry no passport, so the disguise continues at smaller scale.
The United States gains on three boards at once. As an energy exporter it profits from every price spike its rivals suffer. In Venezuela, a strike of barely two and a half hours in January 2026 captured Maduro and installed a friendly transition; officially a move against drug cartels, but the administration itself named access to the world's largest proven oil reserves as a core motive, and the build-up since August 2025 (including secret oil talks) supports the reading that it was long planned. Critics call it a breach of the UN Charter; the oil deals began within weeks. And on the debt side, 4.2% inflation with yields barely above it quietly devalues $36 trillion of liabilities at the creditors' expense: dollar debt in your own currency is a privilege no one else has.
China and India collect the discount. Cheap Russian energy works like an industrial subsidy, refining margins turn sanctions into arbitrage profit, and every yuan-settled barrel builds payment rails beyond the dollar. The Gulf states monetise their pipelines and spare capacity as the indispensable buffer. Switzerland earns on fear itself: safe-haven inflows, 0% rates, banking fees. Even tariff walls create local winners, from US steel mills to Chinese carmakers assembling in Graz to bypass EU duties.
And who pays? Energy-importing, export-dependent Central Europe, squeezed from three sides at once: dearer inputs, dearer money, harder markets. That is precisely why the company at the centre of this map sits in Austria. The lesson is not resignation but realism: crises redistribute rather than only destroy, and whoever understands the flows early (contracts, hedges, supplier choice, market mix) ends up on the better side of the redistribution.
Note on framing: the facts above are sourced (see footer); reading the Venezuela operation as "long planned" and US debt policy as deliberate devaluation are widely held interpretations, marked as such.
"Cash is trash, until the crash." This maxim captures the double face of liquidity: in normal (inflationary) times cash is "trash" because it loses purchasing power in real terms, but at the moment of crisis it becomes king, because then everyone else has to sell and only the liquid can choose. The problem: nobody knows when the "crash" will come. That is why smart companies hold both, a genuine reserve and productively invested capital.
The claim in detail: money is becoming "worth more" right now; whoever is liquid calls the shots. For goods prices this is measurably untrue: the euro area has 2.8% inflation, the US 4.2%. Cash there loses real purchasing power, and there is no trace of deflation in the textbook sense (broadly falling prices as in 1930–33).
But the observation has a genuine core, on the capital side. There, something like a liquidity vacuum really does prevail: real interest rates are positive, central banks have shrunk their balance sheets, banks lend restrictively, and the Fed is hiking into the weakness. Illiquid assets (property, equity stakes, growth firms without cash flow) find buyers less easily and fall in price, so measured in cash, many things are indeed getting cheaper. That is exactly what "cash is king" means.
What follows is a real shift of power towards liquidity: whoever has cash buys out of others' distress (distressed takeovers, stakes at discounts), dictates payment terms (→ Price & balance: supplier credit as an instrument of power), and holds the stronger hand in financing rounds. This reaches all the way up to states: whoever depends on the bond markets, the US with $36+ trillion of debt, Japan at ~200% of GDP, is disciplined by its creditors ("bond vigilantes"). Governments shape policy around whether "the market" will still finance them; in that sense, holders of liquidity really do steer companies and governments.
The limits of the claim, honestly stated: cash simultaneously loses 2.8–4.2% of goods purchasing power per year; the advantage exists only relative to falling asset prices and is a window of time, not a permanent state. If monetary policy turns (as in 2009 or 2020), the waiting position devalues quickly. Historical precedents for such phases: 1981/82 (Volcker), 2008/09, 2022/23. Conclusion: no deflationary vacuum in the world of goods, but a buyer's market for capital, and there the rule is: liquidity is the ability to act.
Entrepreneurial consequence (connecting all four parts): a liquidity reserve of several months' fixed costs, a flexible cost structure (→ Costs & leverage), no dependence on short-term refinancing (→ Price & balance), scenarios instead of forecasts (→ Dependencies).
Then: a highly globalised world economy, two hostile alliance blocs, an arms race, naval rivalry, and at the edges a series of "small" wars (the Balkan Wars of 1912/13) as proxy heating. The spark: the assassination in Sarajevo on 28 June 1914. It was not the trigger that was big; the system was so tense that alliance automatisms turned a regional conflict into a world war within weeks. The participants stumbled in ("sleepwalkers", Christopher Clark).
Recognisable today: bloc formation (G7/NATO vs. the Russia–Iran–North Korea axis, with China tactically in between), proxy conflicts at several edges at once (Ukraine, Middle East/Hormuz, Red Sea, Sahel), rearmament, and alliance commitments that pre-draw escalation chains. A single incident, a warship sunk in Hormuz, an episode in the Baltics or around Taiwan, would structurally play the role of Sarajevo.
The decisive difference: nuclear deterrence sets a hard threshold against the automatisms; great powers communicate directly (hotlines, back channels such as Oman/Qatar), and economic weapons (sanctions, tariffs, blockades) serve as a release valve before the guns. So far the wars have remained deliberately limited and "outsourced". But the markets' nervousness (gold, freight and insurance premiums) shows that the residual risk of miscalculation is being priced in for real.
Then: after the 1929 crash came banking crises and deflationary policy (Brüning cut into the crisis), the Smoot-Hawley tariff of 1930 triggered a worldwide tariff spiral, world trade collapsed by around two thirds, mass unemployment followed, and with it radicalisation: loss of trust in the "establishment parties", the rise of authoritarian and extremist movements, all the way to catastrophe.
The empirical finding behind this is well documented and repeats across countries and decades: the study "Going to Extremes" (Funke, Schularick, Trebesch; 140 years, 20 countries) shows that after financial crises the vote share of far-right/populist parties rises by ~30% on average, parliaments fragment and governing becomes harder. Economic insecurity (inflation, fear of decline, the feeling that burdens are shared unfairly) is the fuel, and the current environment (energy inflation, tariff nationalism, fear of war) supplies exactly that. The tailwind for right-wing parties across Europe and the US follows this historical pattern.
The differences: today, welfare states, deposit insurance and capable central banks cushion the fall. There is inflation instead of deflation, moderate rather than mass unemployment, and the EU ties its members together economically. The 1930s are not destiny but a warning with an instruction manual: price stability, social cushioning and open trade are not just economics, they are the protection of democracy. For companies this means: political risk belongs in every location and market decision.
The longer a conflict lasts, the more the fronts harden, in people's minds too. In debates, "the Russian leadership" quickly becomes "the Russians": people of Russian origin face hostility in everyday life, blanket exclusions from culture and sport hit artists and athletes regardless of their views, and a population that neither decided on the war nor can vote it out is equated with the decisions of the state. This confusion is not just unfair, it is dangerous.
The historical lesson is unambiguous: the dehumanisation of entire peoples was, throughout the 20th century, the companion and pathfinder of its catastrophes: from the propaganda of the world wars, to the blanket internment of Japanese Americans in 1942, to attributions of collective guilt after 1945 that delayed reconciliation by years. Enemy images have an economic and political function: they make harshness palatable to majorities. But they carry a high price: whoever demonises a population blocks their own path to the peace afterwards, because in the end negotiation, trade and reconstruction always happen with people, not with abstractions.
Making this distinction does not mean relativising. Responsibility for the war of aggression lies with the Russian leadership; sanctions, grounded in international law, target war-making capacity and the apparatus of power, not ethnicity. And making the distinction also does not mean naivety towards propaganda, which deliberately pursues exactly the kind of blanket generalisation of "the West" that we, in turn, should avoid. Both belong together: a clear position on the matter, no condemnation of people by origin.
For us as individuals, especially in Austria: exclude no one, separate the state from the person, do not tear down bridges. As a neutral country with Vienna as a UN seat, Austria has historically lived from being exactly this: a place of encounter between the blocs, in the Cold War as today. This role does not begin in diplomacy but in everyday life: with the neighbour, in the shop, in the club. Economically the same holds: markets and supply relationships that are indiscriminately burned in war are missed by everyone in peace.
Context, not a claim of symmetry: the comparison with historical patterns of generalisation refers to the social mechanism of enemy-image formation, not to any equation of today's politics with the crimes of National Socialism.
William Strauss and Neil Howe published Generations in 1991 and The Fourth Turning in 1997; Howe followed up alone in 2023. History, they argue, moves in a saeculum of roughly 80 to 90 years, split into four turnings of about 20: a High with strong institutions, an Awakening that attacks them, an Unraveling in which they are weak and distrusted, and a Crisis that tears the old order down and builds a new one.
By their own dating the current Fourth Turning began with the financial crisis of 2008 and should close around 2033. It is worth checking, because the claim is precise enough to be checked. Two charts below do that; the Cycles chapter works through the circle itself, the archetypes, the criticism and the question of what eighty years after 1945 means.
Every bar is one saeculum, split into the four turnings. Read the widths, not the labels.
Distance from the start of one crisis to the start of the next, using Strauss and Howe's own dates. The dashed line is the eighty years the theory is usually quoted for.
Take the theory at its word and measure it with its own dates. The gap between two crises is not eighty years. It runs from 69 years, the Civil War to the Great Depression, to 106 years, the Reformation to the New World crisis. The mean is about 88, and the spread around it is roughly a fifth in either direction. On that scale a prediction is accurate to within a human working life, which is another way of saying it does not constrain anything. The crisis turnings themselves run from 5 years to 29.
The critics are not fringe: Michael Lind, Frank Giancola, Peter Turchin, Eric Hoover, and the academic historians the New York Times asked in 2017. One more limit is easy to miss: the whole table is Anglo-American, and 1848 does not appear at all. The Cycles chapter names the objections and the falsifiable alternative.
The honest question is not whether a cycle says a crisis is due. It is how tense the system measurably is. Those numbers exist, and most of them are in the other views of this tool.
Global debt stood at about $310tn, roughly 332% of world output, at the start of 2026, against $225tn before the pandemic; governments now account for around 40% of it, up from 35% in 2019. In the 2025 fiscal year the United States paid more than $1.1tn in net interest, more than it spent on defence, for the first time. In the same year the Uppsala conflict programme counted 65 armed conflicts, 13 of them wars, and 8 conflicts between states, each the highest figure since the records begin in 1946, with about 244,600 battle deaths. Add what the other views show: gold near $4,050, a Fed turning back to hikes, the yen at its weakest since 1986, tariffs as normal policy, and a strait that was closed for months.
Every one of those is a condition, not a proof. They say the system is under strain; they do not say a clock is running. The difference matters, because a date invites waiting and a condition invites preparing.
The practical conclusion does not depend on believing in any cycle. Do not plan on a date. Plan on the conditions: a liquidity reserve of several months of fixed costs, a cost structure flexible enough to survive a bad year (→ Costs & leverage), no dependence on short-term refinancing (→ Price & balance), and scenarios instead of forecasts.
Cycles draws the circle itself, shows all seven turns of it, sets out Turchin's testable counterpart, and tests what eighty years after 1945 actually lands on.
Presented as a contested theory, not as a description of the world. The dates in the chart are Strauss and Howe's own; the arithmetic on top of them is ours and can be recomputed from the table.
The AI boom is financed in a circle. Investors put money into AI companies that are contractually obliged to buy from those same investors, and the spending comes back as revenue at the investor. Nothing about that is illegal, and some of it is ordinary vendor financing. What makes it worth a diagram is the scale and the fact that the same euro can be counted several times: once as an investment, once as an order book, once as revenue. Click a node for its position in the circle.
Click a node for its deals
The dark bar is what OpenAI has committed to pay for compute between 2025 and 2035. The other bars are the annual figures the circle actually rests on.
Around 1999 the telecom operators sold each other network capacity. Global Crossing and WorldCom booked the sales as revenue, the purchases as investment, and both sides looked as if they were growing. The fibre was real, the demand was not yet there, and when it became clear how long the capacity would stay dark, billions of market value disappeared within months. The parallel is not an equation: AI compute is being used today and Nvidia earns real cash. The mechanism to watch is the same one, though, namely revenue that has been financed by the recipient of that revenue.
Three connections. First, the price of electricity and of industrial land is being set at the margin by data centres that pay differently from a factory, which lands in the energy view of this tool and in the Price & balance chapter. Second, every AI investment is a fixed cost, so the leverage from Costs & leverage applies, and the useful life assumed for the hardware decides how heavy that block is; Costs & leverage now has a slider for exactly that. Third, when a construction of this size wobbles, it does not wobble alone: the world-finance view shows how quickly a repricing travels through liquidity and interest rates.
A barrel of Brent costs the same in Rotterdam as in Houston, give or take freight and quality. So why does European industry keep saying that energy is its problem, and why is a litre of petrol in Austria roughly twice the American price? Because three separate things are usually run together, and they have different causes and different remedies.
First, the price of the barrel, which is set on a world market and is the same for everyone. Second, the price at the pump, where the difference between Europe and the United States is mostly tax, that is domestic policy. Third, the security of supply and the direction in which the money flows, which is where the real asymmetry sits and which no tax cut can fix.
Sorting them changes the conclusion. Europe's disadvantage at the pump is largely self-inflicted and reversible by decision. Its disadvantage in gas, in industrial electricity and in exposure is structural and is not.
Petrol, euro per litre, 31 August 2026. Dollar prices converted at the August average of 1.159. Only the two bars with a split are decomposed; the others are pump prices without a breakdown.
In Austria a litre of Eurosuper cost 1.784 euro on 31 August 2026. Of that, the mineral oil tax is 48.2 cents, the CO2 component 12.5 cents, and 20 percent VAT sits on top of everything, which comes to about 90 cents of tax and leaves roughly 88 cents for the crude, the refining and the distribution.
In the United States a gallon of regular cost 4.07 dollars in the same week, which is 1.075 dollars or about 93 euro cents per litre. Federal and average state fuel taxes together come to roughly 51 cents per gallon, about 12 euro cents per litre. That leaves roughly 81 euro cents for product and distribution.
So the product side differs by less than ten percent. The entire rest of the gap, close to 80 cents a litre, is tax. Germany at 2.24 euro and Denmark at 2.52 are higher again, Malta at 1.34 is the cheapest in the union, and none of that spread has anything to do with import dependency. It is a policy choice, and it is the one part of the disadvantage that a government could change on a Tuesday.
US regular gasoline, dollars per gallon. Annual averages 2022 to 2025; for 2026 the range so far with the September value marked.
The United States is a net exporter of petroleum. It produces more than it consumes on balance, it is the largest exporter of liquefied natural gas, and the Gulf Coast alone exports enough to outweigh the net imports of every other region. On the usual reading that should insulate American consumers.
It does not, and 2026 is the proof. The pump price started the year at 2.78 dollars a gallon, the cheapest week since 2021, and peaked at 4.50 dollars in May. It stood at 4.15 dollars on 4 September, up from 4.08 a month earlier, against annual averages of 3.12 in 2025 and 3.36 in 2024. A barrel produced in Texas is sold at the world price, because the producer would otherwise export it. Self-sufficiency in volume does not buy independence in price.
That matters for the argument, because it cuts both ways. It also settles a claim that gets made in the other direction, that American consumers are back at or below their pre crisis level. They are not. They are paying about a third more than in 2025 and about half as much again as in the cheapest week of January 2026. It weakens the claim that America is simply insulated. And it strengthens the claim that the real asymmetry is somewhere else.
When the oil price rises, both economies pay more at the pump. The difference is where the money ends up. In the United States it is a redistribution inside the country: consumers lose, producers, refiners, drilling states and the federal treasury gain, and the money stays in the currency area. In Europe it is a transfer abroad. There is no domestic producer on the other side of the trade to gain what the consumer loses.
The size of that transfer is measurable. In 2025 the EU imported energy products worth 336.7 billion euro, down 11.1 percent on 2024, and the union covers 57 percent of its energy demand from imports, Germany 67 percent. That is the number that actually moves European industry and European inflation, and it is not affected by fuel duty at all.
The second half of the asymmetry is gas, and there the gap is not ten percent but a multiple. European TTF and American Henry Hub have been running roughly a factor of six apart, which is the number behind every discussion about energy-intensive production leaving. The pump price is what people feel; the gas price is what decides where a factory stands.
Share of crude and condensate transiting the Strait of Hormuz by destination, first quarter of 2025.
The Strait of Hormuz is routinely described as the world's most important oil chokepoint, which is true, and then quietly treated as a European problem, which is not. Asian buyers take about 89 percent of everything that passes through it. China alone takes 37.7 percent, India 14.7. Europe takes 3.8 percent and the United States 2.5.
So the direct exposure of both Europe and America is small, and the reading that Hormuz matters less to Washington than to Berlin is correct, but for a sharper reason than usual: it matters less to both of them directly, and it matters enormously to Beijing, Delhi, Seoul and Tokyo.
Europe is hit through two indirect channels instead. The world price rises for everyone, and Europe pays it without a domestic producer to offset it. And when Gulf barrels are disrupted, Asian buyers bid for the same Atlantic-basin crude and the same LNG cargoes that Europe needs, so Europe competes for the replacement. That is what 2026 looked like in practice.
Share of EU imports of each product by supplier, 2025.
This is the part that gets least attention and deserves most. Europe reduced its dependence on Russia and replaced it. By 2025 the United States supplied 56 percent of the EU's liquefied natural gas and 15.1 percent of its crude and refined products, the largest single share in both. Norway supplies 52.1 percent of pipeline gas. Russia still accounts for 13.9 percent of LNG and 10.4 percent of pipeline gas.
Swapping a supplier you cannot influence for a supplier you cannot influence is a real improvement in one respect, because the new one is an ally and the old one is at war on the continent. It is not an improvement in dependency. Europe now buys the majority of a critical input from a country whose energy and trade policy it does not vote on, and it pays in a currency it does not issue.
A version of the story now in circulation has Washington and Moscow back in business, with Russian oil and gas flowing to the United States. The first half of that is close to something real. The second half is not, and it is the easiest part to check.
American crude imports from Russia have been zero since April 2022 and were still zero in every reported week of 2026. There is no volume, no route and no terminal. Whatever is happening between the two governments, it is not a delivery.
On 13 March 2026, with Brent above 100 dollars after the Hormuz disruption, the Treasury lifted its own sanctions for thirty days on Russian crude that was already loaded on tankers, roughly 125 million barrels. That is five or six days of normal Hormuz throughput and a little over one day of world consumption. The buyers were refineries in China and India. The waiver was extended on 18 April and ran to 16 May. The stated purpose, from Treasury Secretary Scott Bessent, was to keep world prices down.
So the barrels moved east, not west. What America supplied was not demand but permission, and what it bought was a lower world price, which is also the price Americans pay at the pump. Read that way the move is not a favour to Moscow but a domestic price measure that happens to pass through Moscow.
Alongside it runs a private track. An American investor close to the president's circle signed a liquefied natural gas agreement with Novatek, reported in February 2026. There has been persistent reporting on American investors taking over Nord Stream 2 and on American firms acting as intermediaries who would buy Russian gas and resell it into Europe. None of it has been concluded and the Russian side has at times denied that talks were active.
In October 2025 the same administration sanctioned Rosneft and Lukoil, Russia's two largest oil companies. On 7 August 2026 the Senate passed an energy sanctions bill by 86 votes to 11, allowing tariffs of up to 500 percent on Russian imports and up to 100 percent on goods from major buyers of Russian oil such as China and India. It was still in the House in early September.
Two tracks are running at once and they point in opposite directions. Reading either one alone gives a confident answer and the wrong one. The summary that survives both is narrower: the relationship is being negotiated in public with sanctions as the currency, and no energy is crossing the Atlantic in either direction as a result.
Europe losing Russian energy is usually described as something done to it. The decisive act has a European date and a European signature. On 26 January 2026 the Council gave final approval to a stepwise ban: Russian liquefied natural gas from the start of 2027, pipeline gas from autumn 2027, with legislation on oil proposed to follow by the end of that year. Existing contracts get a transition period and the Commission may suspend the ban for up to four weeks in a declared supply emergency.
That is a choice, taken deliberately, with the cost visible in advance. It can be criticised as expensive or defended as necessary. It cannot be attributed to Washington. The part that can fairly be laid at America's door is narrower and is described two sections above: Europe replaced the supplier it banned largely with American cargoes, and thereby moved the dependency rather than ending it.
The Dutch TTF benchmark, the price European industry actually pays, began 2026 at about 29 euro per megawatt hour and traded above 66 euro on 26 August. It more than doubled. Storage went into the autumn 62.99 percent full against a five year average of 79 percent, with Germany at 51 percent, which is what removes the cushion for the winter.
The European Commission cut its 2026 growth forecast on 21 May from 1.4 to 1.1 percent for the union and to 0.9 percent for the euro area, and raised its inflation forecast by a full point to 3.1 percent, naming the disruption around Hormuz as the reason. That is the measured size of the damage, and it is large without being a collapse.
A doubling on the exchange is not a doubling on the invoice. Analysis by Oxford Economics, produced before the August spike, put euro area consumer energy prices up to 15 percent higher year on year in the fourth quarter of 2026. How fast that lands depends on the contract. In the Netherlands the adjustment is almost immediate. In France, Italy and Spain it takes a few months. In Germany and Austria it takes about a year, because household contracts are typically fixed for twelve to twenty four months.
So an Austrian household reading its current bill is mostly reading 2025. What is already decided is the grid. Network charges for 2026, set in December 2025, rose 1.3 percent for electricity on average and 18.2 percent for gas, with Carinthia at 35 percent and about 142 euro a year more for a household using 15,000 kilowatt hours. The regulator's reason is worth noting because it is not geopolitical at all: consumption is falling, so the same fixed network cost is spread over fewer kilowatt hours.
This is where a figure often quoted, that European consumers are already paying thirty percent more, runs ahead of what is measured. Wholesale gas has more than doubled. Consumer energy is forecast at around fifteen percent. The distance between those two numbers is contracts and time, and most of the 2026 wholesale shock will reach Austrian households in 2027. The number is not wrong so much as early.
Put end to end the sequence reads like this. A disruption raises the world price. Europe pays it with no domestic producer on the other side, so the money leaves the currency area. Europe has also, by its own law, given up the cheapest supplier it had and replaced it with cargoes priced on a market where it is the marginal buyer. Inflation follows, real incomes fall, and the loss is spread across every household and every energy using firm instead of appearing as a line item anyone voted on. On this reading it works as a transfer out of the European economy, and calling it a price understates it. This is an interpretation. It is a defensible one.
About half of what an Austrian household pays for fuel is domestic tax, and the grid increase for 2026 is a domestic regulatory decision driven by falling consumption. Neither is a transfer abroad. The supply ban for 2027 was voted for by European governments with the cost visible. And the same disruption raised the American pump price by more than a third, so the asymmetry sits in the balance sheet and in the exposure, not in whether anyone was spared.
What is left. A structural gap in gas and industrial electricity that no tax decision closes, a supply position in which Europe is the marginal buyer, and a cost that arrives late and is therefore attributed to whatever is in the news on the day the bill lands. Those three hold whoever is in office anywhere.
On 3 January 2026 the United States carried out a military operation in Venezuela, captured Nicolás Maduro and flew him to New York, where he was arraigned on narco-terrorism charges on 5 January. Vice President Delcy Rodríguez was sworn in as acting president and was recognised by Washington as sole head of state in March; the American embassy reopened on 30 March.
The energy side moved within weeks. Sanctions on Venezuelan oil trade were lifted, a privatisation law passed on 29 January, an initial two-billion-dollar arrangement was made with the first 300 million received on 20 January, and a fifty-million-barrel supply contract followed. By February the American energy secretary reported more than a billion dollars in sales. On 1 September 2026 the Venezuelan National Assembly approved a further agreement and it was signed the next day. As reported, the Pentagon's Office of Strategic Capital takes a 35 percent equity stake in an entity holding 17 fields with about 65 billion barrels, the State Department may buy 20 percent of output at production cost, and American citizens must hold a board majority. Chevron committed 7 billion dollars through 2031 with a target of 600,000 barrels a day, up from about 280,000. The reporting on the scope is contested and the Venezuelan side disputes parts of it.
The strategic logic is not only volume. Gulf Coast refineries are configured for heavy crude, which American shale does not produce; Venezuela does. Adding heavy barrels under American influence a few days' sailing away closes the one gap in an otherwise self-sufficient system. The size of the prize should be kept in proportion, though. Venezuela currently produces about 1.1 million barrels a day against more than three million in the late 1990s, which is under one percent of world supply even if it all returned. Rebuilding the fields is the work of a decade and hundreds of billions of dollars, and Iraq and Libya are the standing reminder that a change of government does not guarantee a recovery. Reserves in the ground are not barrels on the market, and they explain nothing about the price in 2026.
The reading. A country that is not physically dependent on imported energy can afford confrontation that a dependent country cannot. Every disruption raises the price, the price is paid disproportionately by the dependent, and the money partly flows to the independent. On that view a Washington administration can be disruptive in world politics at a low domestic cost, and Europe carries the bill without a vote on the policy. This is an interpretation, and it is a coherent one.
What argues against it. The domestic cost is not zero, as 2026 showed at 4.50 dollars a gallon; the United States is exposed through inflation, financial markets and its own allies. Calling the posture neutral does not fit an intervention in Venezuela and open pressure on European trade, which is active policy, not detachment. Causality is also not established: American energy self-sufficiency was built over fifteen years by the shale industry, not designed as a foreign-policy instrument, and interests that exist are not the same thing as motives that were acted on.
What is left. A structural asymmetry of vulnerability, which is a fact, and which translates into an asymmetry of bargaining power whoever is in office. That is the part worth planning around, and it does not depend on anyone's motives.
Fuel duty is politics and can change; the gas price is structural and will not converge soon. So an energy-intensive process in Central Europe has to be planned against a permanent cost gap to the United States, not against a temporary one. Costs & leverage has the fixed-cost leverage that decides how a bad year ends, and the energy section of the company network has the sensitivity to the gas price.
The exposure that is worth hedging is not the pump price but the electricity and gas contract, and the currency: energy is invoiced in dollars, so a weak euro raises the bill twice. And the supplier question deserves the same treatment as any other dependency in this chapter, which is to be mapped rather than assumed.
Sources in the footer. The decomposition of the Austrian pump price is our own calculation from published tax rates; the American one uses federal duty plus the average state rate. The reading in the last card is marked as an interpretation, and the counter-arguments are given with it.