Paper, Silicon, and Crude

How the Iran Conflict Connects to the AI Supply Chain

I. EXECUTIVE SUMMARY

On March 1, 2026, the United States and Israel launched a full-spectrum military campaign against Iran — airstrikes on nuclear facilities, missile production sites, and naval assets, concurrent with the confirmed killing of Supreme Leader Khamenei and approximately 40 senior officials. This is not a Soleimani-style targeted strike. It is a stated regime change operation, building on the June 2025 nuclear facility strikes and weeks of visible military buildup.

This report is not a geopolitical forecast. We are not predicting how the war ends. We are mapping two financial transmission mechanisms that connect a Middle Eastern conflict to the AI infrastructure supply chain — and specifically to CoreWeave, the most leveraged link in that chain, where we hold a short position via a September 2026 bear put spread.

Transmission 1: Hormuz → Oil → Inflation → Rates → Interest Coverage. The Strait of Hormuz has effectively shut down. Roughly 20% of global oil supply — ~21 million barrels per day — transits the Strait. Every day of closure removes $2-3B in crude from global markets. Oil was already up 17% YTD before the attack. If the disruption persists beyond days, the inflationary impulse feeds directly into the Federal Reserve’s rate calculus, which feeds directly into CoreWeave’s blended cost of debt, which feeds directly into the interest coverage squeeze we documented in our CoreWeave report.¹ At 1.5x EBITDA-to-interest coverage in Q1 2026, CoreWeave cannot afford rates staying higher for longer. Hormuz staying closed is a direct threat to the rate relief the bull case requires.

Transmission 2: Pacific Instability → TSMC → NVIDIA → CoreWeave. Within hours of the first strikes, the PLA Daily — the official newspaper of the Chinese military — published a front-page editorial arguing that the US cannot cover all its commitments simultaneously, using the idiom 十口锅九个盖 — “ten pots, nine lids.” On the same front page: PLA Southern Theater Command combat readiness patrols at Scarborough Shoal. China is not going to invade Taiwan this month. But the supply chain we mapped in our CoreWeave report — MSFT → CRWV → NVDA → TSMC — terminates at a foundry on an island that sits at the center of China’s stated territorial claims.² If China tests shipping lanes in the western Pacific while the US is committed in the Persian Gulf, the TSMC-to-NVIDIA link in that chain comes under pressure. A single quarter of GPU delivery delay costs CoreWeave approximately $750M in incremental carrying costs.

These are not speculative scenarios. The Strait is closed now. The PLA Daily editorial ran today. The transmission mechanisms are structural. What follows maps each one.

II. THE STRAIT

The Strait of Hormuz is a 21-mile-wide chokepoint between Iran and Oman through which roughly 21 million barrels of crude oil transit daily — approximately one-fifth of global supply. As of the first hours of the US-Israel strike, multiple independent sources confirm the Strait has effectively shut down. An Iranian tanker has been struck near Oman. Commercial shipping has halted.³

Iran does not need to hold the Strait militarily. It needs to make transit too risky for commercial shipping insurers. The Congressional Research Service documented Iran’s capabilities in detail: anti-ship cruise missiles (Noor and Ghader systems), fast-attack craft swarms, naval mines, and shore-based artillery positioned along the northern coast.⁴ The moment Lloyd’s of London refuses to underwrite tanker passages, the Strait is functionally closed regardless of who controls the water.

The Abqaiq Precedent

The closest market-impact analogue is the September 2019 attack on Saudi Arabia’s Abqaiq oil processing facility, which disrupted 5.7 million barrels per day — roughly 5% of global supply. The market response: a 14% spike in both WTI and Brent crude in a single trading session.⁵

That disruption resolved within weeks. Prices retreated almost as quickly as they spiked. The current situation is categorically different:

The Abqaiq precedent suggests that even a small Middle Eastern oil disruption causes outsized price spikes. The current disruption is four times larger and has no clear resolution timeline.

Duration: The Air Campaign Is Not the Timeline That Matters

The air campaign itself has a ceiling. The US has been drawing down precision munitions stockpiles through years of Ukraine aid, Israel resupply, and the June 2025 strike. Reporting from multiple sources indicates the US is racing to accomplish its objectives before stocks run thin — a 7-14 day window for active bombardment.⁶

The bombardment phase will likely be fast and decisive. The US has overwhelming air superiority, has clearly pre-positioned assets after weeks of buildup, and has no peer adversary contesting the airspace. The military outcome is not in question. Expect a press conference within two weeks declaring the operation a success — nuclear sites destroyed, missile production eliminated, naval assets sunk, regime leadership decapitated.

The market should not confuse that press conference with the end of the disruption. There are four distinct timelines at work, and only the shortest one is controlled by American air power:

The critical distinction: the air campaign is the only phase where the US sets the timeline. Every subsequent phase is controlled by actors and processes that do not respond to press conferences.

Hormuz does not reopen when the bombing stops. It reopens when mines are cleared from the shipping lanes and commercial insurers resume underwriting tanker passages. Iran’s mine-laying capability is well-documented — CRS identified naval mines as one of Iran’s most effective asymmetric tools, precisely because they are cheap to deploy and expensive to remove.⁴ The US Navy has not conducted a major mine clearance operation since the Persian Gulf in 1991. Modern mine countermeasures require weeks of painstaking, ship-by-ship sweeping. Even after the last mine is cleared, Lloyd’s will not normalize premiums overnight. Insurers price risk backward — a strait that was mined once can be mined again. The insurance premium overhang alone can suppress tanker traffic for weeks after military operations conclude.

The proxy network does not disband when the regime falls. The IRGC has already announced “the most ferocious offensive in history.” This is not an empty threat — it is an organization with operational cells embedded across four countries (Lebanon, Yemen, Iraq, Syria) and a command structure designed to function independently of Tehran. Killing the Ayatollah and 40 senior officials removes the political authority above the IRGC. It does not remove the IRGC. It may remove the restraining hand. Hezbollah alone has an estimated 150,000 rockets pointed at Israel. The Houthis have already demonstrated the ability to disrupt Red Sea shipping for over a year. Iraqi Shia militias have struck US bases repeatedly since October 2023. These forces activate on their own timelines, in their own theaters, against their own target sets. No amount of air power over Tehran changes that.

The power vacuum is the longest and least predictable timeline. Iran is a country of 85 million people with a sophisticated military-industrial complex, a clerical governance structure with no clear succession mechanism, and an IRGC that functions as a parallel state. Decapitating the leadership creates a succession crisis that could resolve in months — or fracture into years of internal conflict. Iraq in 2003 is the direct precedent: the air campaign lasted three weeks, the occupation lasted eight years, and the insurgency it spawned eventually produced ISIS. We are not predicting a repeat. We are noting that every US regime change operation in the Middle East has followed the same pattern — swift military victory, protracted and costly aftermath — and that the aftermath is where the economic damage concentrates.

Why This Time Is Not Sabre-Rattling

The obvious counterargument: Iran threatens apocalyptic retaliation every other month. After Soleimani, the IRGC promised “severe revenge” and delivered a missile barrage on Al-Asad air base — with advance warning, designed to demonstrate capability without triggering escalation. After the June 2025 nuclear strikes, Iran condemned and calibrated. The pattern is familiar: threaten fury, deliver a measured response, preserve the regime. Why should this time be different?

Because every previous Iranian response was a rational calculation by a regime that had something to preserve. The Supreme Leader was alive. The succession structure was intact. The nuclear program was damaged but recoverable. The IRGC’s economic empire — construction firms, banks, telecom companies, an estimated 20-30% of Iran’s GDP — was unthreatened. Restraint was rational because the status quo, even a diminished one, was preferable to escalation. The regime could absorb a targeted strike and survive. It could not absorb regime change.

That calculus has inverted. The Supreme Leader is dead. Forty senior officials are dead. The nuclear program has been struck for the second time in nine months. The navy is being sunk. The missile production infrastructure — the IRGC’s primary deterrent — is being dismantled. The conditions that produced past restraint — a functioning regime with something to preserve — no longer appear to hold. That does not guarantee a maximalist response. But it removes the structural incentive for a calibrated one.

The proxy network makes large-scale retaliation operationally feasible, not just rhetorically promised. Iran spent decades and billions of dollars building Hezbollah, arming the Houthis, embedding Iraqi Shia militias, and funding Hamas for exactly this scenario — a direct US-Israeli assault on the homeland. The entire strategic logic of the proxy network is to impose costs on the US and its allies that make an attack on Iran prohibitively expensive. If Iran does not activate these assets now — when the Supreme Leader has been assassinated and the state is under existential threat — then the deterrent was never real. Every proxy commander in the network understands this incentive structure. That does not mean full activation is certain — command-and-control disruption, internal power struggles, and individual proxy calculations could fragment the response. But the institutional incentives all point toward escalation rather than restraint.

The Houthis have already provided the proof of concept for what a motivated asymmetric actor can achieve. For over a year, a militia with no air force, no navy, and no industrial base has disrupted Red Sea shipping — one of the world’s busiest commercial corridors — despite sustained US and UK airstrikes. They have demonstrated that sustained economic costs on global trade are achievable without anything resembling conventional military capability. That is the Hormuz playbook. And the forces available to execute it at Hormuz are better armed and better trained than the Houthis.

For this report, only the first two timelines matter directly. The air campaign determines headlines. The Hormuz disruption determines oil prices. And oil prices determine whether the inflation transmission mechanism described in Section III activates. Everything beyond that — the proxy wars, the power vacuum, the regional realignment — is structural damage that compounds over months and years, reshaping the operating environment in ways that are harder to price but no less real.

III. THE INFLATION TRANSMISSION

Here is where the geopolitics connects to a balance sheet.

Every $10 per barrel increase in crude oil prices shaves approximately 0.2-0.3% off US GDP growth and adds measurable inflationary pressure to the CPI basket. Energy costs feed through to transportation, manufacturing, petrochemicals, food production, and consumer sentiment. The Fed knows this. The market knows this. What matters is the duration.

A one-week oil spike is a headline. A two-month oil spike is an inflation print. And an inflation print changes the rate calculus.

The Rate Path Before Hormuz

The consensus rate path entering March 2026 assumed continued disinflation — the Fed holding steady or beginning to ease, with the long end of the curve gradually compressing. CoreWeave’s bull case depends on this. The company carries $21.4B in total debt at a blended interest rate of approximately 8%. Every basis point of rate relief flows directly to the bottom line. At 1.5x EBITDA-to-interest coverage in Q1 2026, CoreWeave is operating at the thinnest margin of any company in the AI supply chain.¹

The Rate Path After Hormuz

If oil sustains above $100 for more than a few weeks — let alone approaches the $150 some analysts are projecting — the Fed cannot cut. It may not need to hike, but the “higher for longer” narrative, which the market spent all of 2025 trying to price out, reasserts itself immediately.

For CoreWeave, this means:

The critical insight is not that higher rates break CoreWeave immediately. It is that the bull case requires rate relief to work. CoreWeave’s interest expense was guided at $510-590M for Q1 2026 alone.¹ If the blended rate does not decline through 2026 — if the refinancing environment stays hostile because oil is keeping inflation elevated — then the margin trough CoreWeave projected for Q1 does not remain a trough. It becomes a plateau. And at 1.5x coverage on a plateau, a single execution miss pushes coverage below 1x.

This is not a prediction about oil prices. It is a mapping of the transmission mechanism. Hormuz stays closed → oil stays elevated → inflation stays sticky → rates stay high → CoreWeave’s interest coverage does not improve → the timeline to profitability extends → the equity reprices.

IV. CHINA’S CALCULATION

Beijing’s response to the Iran strikes landed at Level 3 out of 5 on the escalation scale we track — “firmly opposes” (坚决反对), “strongly condemns” (强烈谴责) — matching the language used after the Soleimani killing in 2020.⁷

What Beijing is not saying matters more. None of the six tripwire terms that have historically preceded Chinese military action appeared in the first 24 hours of coverage:

• No 核心利益 (core interest)
• No 玩火自焚 (playing with fire)
• No 不惜一切代价 (at all costs)
• No 勿谓言之不预 (don’t say we didn’t warn you)

The last of these — 勿谓言之不预 — has appeared in People’s Daily before every major Chinese military action since 1962. Its absence is itself data. Beijing does not view Iran as worth a military confrontation with the United States. Iran is a trading partner and oil supplier, not a treaty ally.⁸

What Beijing Is Doing

But the PLA Daily told a different story from the Foreign Ministry — and in China, when the military newspaper and the diplomatic apparatus send different signals on the same day, both signals are intentional.

Three items appeared on the PLA Daily’s front page on March 1:

1. A “ten pots, nine lids” (十口锅九个盖) editorial — an explicit argument that the US cannot cover all its strategic commitments simultaneously

2. PLA Southern Theater Command combat readiness patrols at Scarborough Shoal. Scarborough Shoal is a contested reef in the South China Sea, approximately 120 nautical miles west of the Philippines and well within its exclusive economic zone. China seized effective control in 2012 and has held it since, despite a 2016 international tribunal ruling that China’s claims have no legal basis. The Southern Theater Command is the PLA formation responsible for both the South China Sea and the Taiwan Strait — meaning any escalation it signals applies to the full western Pacific operating theater. “Combat readiness patrols” is a specific escalation in language from routine freedom-of-navigation activity. It signals that the PLA is not merely present at the shoal — it is advertising a wartime posture there, on the same day the US committed its combat power to the Persian Gulf.

3. A counter to Philippines-led joint patrols with extra-regional navies in the South China Sea. Over the past two years, the Philippines has deepened security cooperation with the US, Japan, and Australia, including joint naval patrols through contested waters. Beijing frames these as provocations by “extra-regional” powers — a deliberate word choice designed to delegitimize any non-ASEAN military presence in the South China Sea. The PLA Daily’s announcement that it is prepared to “counter” these patrols is a signal aimed at Manila’s partners as much as Manila itself: while the US is flying sorties over Tehran, China is reminding Washington’s Pacific allies that the Seventh Fleet cannot be in two oceans at once.

Chinese state media is editorialized with extreme precision. The co-placement of these three items is not an accident. Read together, the message structure is: America is overextended → we are asserting combat readiness at a contested feature in the Pacific → external powers operating in our waters can be countered. The editorial frames the strategic logic. The two military items demonstrate it.

The Capacity-Constraint Argument Is Real

The PLA Daily metaphor is not just rhetoric. The US government’s own documents support it:

• Syria withdrawal concurrent with Iran offensive. A CRS report updated two days before the strike reveals the US was preparing full withdrawal from Syria within two months — simultaneously winding down one Middle East commitment while launching a far larger one.⁹
• Munitions constraints. Years of Ukraine support, Israel resupply, and the June 2025 strike have depleted precision weapons stockpiles to the point where the air campaign timeline is supply-constrained, not strategy-constrained.
• Nuclear modernization costs. CRS estimates US nuclear force modernization at $95B per year.¹⁰ Adding a major conventional war creates competing demands on a defense budget already stretched by modernization.

China’s Energy Exposure

China imports approximately 40% of its crude oil through the Strait of Hormuz. This is existential exposure. But notably, no emergency measures have been announced — no SPR drawdown, no emergency pipeline pivot to Russia or Central Asia, no State Council energy security sessions. The restraint is a signal. Beijing may be waiting to see whether the Strait disruption is measured in days or weeks before triggering emergency protocols.

The depth of China’s economic entanglement with Iran makes the diplomatic restraint all the more notable. CRS data shows China’s share of Iran’s global trade expanded from 20% in 2010 to 31% by 2016. Chinese state-owned enterprises are building Tehran’s subway, developing Iranian oil fields, and operating a railway from China through Central Asia to Tehran as part of the Belt and Road Initiative.⁸ The US attack on Iran does not just threaten China’s oil supply — it threatens billions in Chinese infrastructure investment and a key BRI node.

China’s Most Likely Play

Beijing will exploit this diplomatically, not militarily:

1. Position as the voice of restraint while the US is the aggressor — accumulate soft power across the Global South
2. Quietly test US attention and response times in the South China Sea through increased patrols — already underway
3. Accelerate energy diversification from Hormuz-dependent routes toward overland Russian and Central Asian supply
4. Coordinate with Russia on a diplomatic counter-framework — the Wang Yi-Lavrov call happened on Day 1
5. Wait — if the US gets bogged down in a protracted Iran engagement, the Pacific deterrence calculus shifts structurally in China’s favor over months, not days

V. THE SUPPLY CHAIN EXPOSURE

This is the lower-probability of the two transmission mechanisms — but the one with the larger potential impact. The Hormuz → oil → rates chain is already active. The Pacific → TSMC chain is not. It is a conditional risk: it materializes only if China moves from rhetoric to action in the western Pacific, and the evidence from Section IV suggests Beijing is positioning diplomatically, not militarily. We include it because the dependency is real, the signals are new, and the financial exposure if it activates is severe.

In our CoreWeave report, we mapped the AI compute supply chain as a linear sequence of financial dependencies:¹

That chain terminates at Taiwan Semiconductor Manufacturing Company — a single foundry on an island 100 miles off the coast of mainland China. TSMC fabricates virtually all of NVIDIA’s advanced AI chips. NVIDIA ships those chips to CoreWeave. CoreWeave racks them in data centers, funded by $21.4B in debt. The dependency is structural and undiversifiable in the near term.

The question is not whether China will blockade Taiwan — it almost certainly will not in 2026. The question is whether increased PLA activity in the western Pacific creates enough maritime ambiguity to affect shipping costs, delivery timelines, or insurance premiums along the route from Hsinchu to Santa Clara. The same logic we applied to Hormuz applies here at lower intensity: a chokepoint does not need to be militarily closed to become economically impaired. It needs to become uncertain enough that the cost of transit rises.

From our CoreWeave analysis: a three-month supply chain disruption on $10B of GPU orders produces approximately $750M in incremental carrying costs — interest and depreciation accruing on hardware that cannot generate revenue.¹ CoreWeave guided at least $30B in FY2026 capex. Even modest delivery delays — weeks, not months — compound against a balance sheet already operating at 1.5x interest coverage.

This risk is not priced into CoreWeave’s equity or options chain, which reflect credit and execution risk. Whether it should be priced depends on how far Beijing pushes the “ten pots, nine lids” framing beyond rhetoric — a question the signals in Section VII are designed to answer.

VI. SECOND-ORDER EFFECTS

The first-order effects — oil prices and supply chain friction — are the most direct transmission mechanisms to the AI infrastructure thesis. But the Iran conflict introduces structural shifts that reshape the global operating environment over longer time horizons.

The Nonproliferation Collapse

Iran’s nuclear program has now been struck twice in nine months. The lesson every aspiring nuclear state draws is unambiguous: acquire the capability before the US strikes, not after. North Korea will point to this as vindication. Saudi Arabia’s nuclear ambitions gain stronger rationale. The nonproliferation regime — already fraying — may not survive this precedent.

The timing amplifies the damage. New START expired on February 5, 2026 — three weeks before the Iran attack. The US and Russia are operating without a binding nuclear arms treaty for the first time since the Cold War. Russia has deployed nonstrategic nuclear weapons to Belarus. China is expanding toward 1,000+ operational warheads by 2030.¹⁰ The Iran strike lands at the exact moment when the global nuclear order is most fragile.

The AI Capex Cycle at Risk

This report focuses on CoreWeave because it is the most leveraged and therefore most fragile link in the AI supply chain. But CoreWeave is not building this infrastructure alone. The entire hyperscaler complex is making enormous, time-sensitive capital commitments predicated on the same assumptions that Hormuz and the western Pacific now threaten.

Microsoft has $631B in Remaining Performance Obligations — contracted cloud services it has promised to deliver. Delivering them requires data centers, which require GPUs, which require TSMC silicon, which requires functioning trans-Pacific shipping lanes and stable energy costs. Amazon, Google, and Meta are making comparable bets. Combined hyperscaler capex commitments have exceeded $700B since mid-2023.¹ Every one of those dollars was budgeted at pre-Hormuz energy prices and pre-Hormuz interest rates.

The hyperscalers can absorb cost overruns that would bankrupt CoreWeave — that is the balance sheet advantage we documented in our previous report. But absorption is not immunity. Higher energy costs raise the operating expense of every data center in the fleet. Higher interest rates raise the cost of financing new construction. Higher oil prices feed into steel, concrete, copper, and every other input that goes into a 500 MW data center campus. If the AI infrastructure buildout was already the largest capital expenditure cycle in human history, an oil shock makes it more expensive at precisely the moment when the industry is trying to scale it faster.

The subtler risk is on the demand side. The hyperscalers are building capacity because their enterprise customers are committing to AI workloads. Those enterprise customers operate in the real economy — the one where $120 oil raises input costs, compresses margins, and forces CFOs to revisit discretionary spending. AI compute is increasingly non-discretionary for large enterprises, but it is not immune to budget pressure. A sustained oil shock does not kill AI demand. But it can slow the adoption curve at the margin — and at the margin is where the difference between $12B and $10B in CoreWeave FY2026 revenue lives.

European Energy Crisis 2.0

Europe is more oil-import dependent than the United States. A sustained Hormuz closure hits European economies harder — precisely when Europe is attempting to rearm and manage energy costs post-Russia. German Chancellor Merz visited Beijing days before the strike. The energy equation for Europe just got dramatically worse.

The Sovereignty Precedent

Beijing’s diplomatic framing deserves attention for what it signals about the future, not just the present. China is anchoring its condemnation in the principle that assassinating a foreign head of state violates the UN Charter. The subtext is clear: if the US will do this to Iran, the argument for nuclear deterrence strengthens for every state with adversarial US relations. This accelerates the multipolar fragmentation that makes supply chains harder to secure and geopolitical risk harder to price.

The Rules-Based Order Is Eroding from Both Sides

There is an irony running beneath all of this. Beijing’s condemnation of the Iran strikes rests on the claim that the US has “trampled” (践踏) international law and the UN Charter. But on the same day, China announced combat readiness patrols at Scarborough Shoal — a feature it occupies in direct defiance of a 2016 Permanent Court of Arbitration ruling issued under the United Nations Convention on the Law of the Sea, which China itself ratified. Beijing’s position is that the ruling is illegitimate and non-binding. Washington’s position is that assassinating a head of state falls within its national security prerogatives. Both powers are invoking international law selectively, discarding it when it conflicts with their strategic interests.

The implication is corrosive — and it extends beyond geopolitics into the financial infrastructure that this report is concerned with. The rules-based frameworks that underpin freedom of navigation, maritime shipping insurance, and the predictability of global supply chains are only as durable as the willingness of great powers to observe them. Right now, neither is. That is not a one-time risk event. It is a structural repricing of the assumption that goods flow freely across oceans — the same assumption on which CoreWeave’s just-in-time GPU delivery pipeline, and every supply chain mapped in our previous report, depends.

VII. WHAT TO WATCH

The single most important near-term signal for the CoreWeave trade is whether the Hormuz disruption persists long enough to change the Fed’s rate path. That answer arrives in weeks, not months. The Q1 CPI print — likely in mid-April — will be the first hard data point. If core CPI re-accelerates on energy pass-through, the “higher for longer” narrative is back, and every leveraged borrower in America — CoreWeave included — reprices.

VIII. WHAT COULD PROVE THIS WRONG

Hormuz reopens quickly. If the US Navy clears mines and secures transit within days, and commercial insurers resume underwriting tanker passages within two weeks, the oil spike is transient. Markets have short memories. A one-week disruption is noise, not signal.

The Fed looks through it. Central banks have historically treated oil supply shocks as transitory rather than structural. If the Fed explicitly communicates that it views energy-driven inflation as a supply shock that does not warrant a policy response, the rate path may not change. This is plausible but historically unreliable — the Fed said inflation was “transitory” in 2021 and was wrong for eighteen months.

China de-escalates in the Pacific. If the PLA Daily’s “ten pots, nine lids” framing is rhetoric rather than doctrine, and Scarborough Shoal patrols return to baseline within a week, the Pacific supply chain risk dissipates. The absence of a 钟声 editorial linking Iran to Pacific security would support this reading.

CoreWeave’s debt is entirely fixed-rate. Management stated that no debt matures until 2029. If the existing stack is entirely fixed-rate, rising market rates affect only new issuance, not the existing cost structure. This would limit the inflation transmission mechanism to refinancing risk rather than immediate coverage pressure. We will update the fixed-to-floating analysis at Q1 earnings.

AI demand overwhelms everything. If the compute buildout is large enough and urgent enough, the hyperscalers absorb oil-driven cost increases, rate increases, and supply chain friction without blinking. Microsoft’s $631B in RPO suggests this is at least possible. In this world, CoreWeave’s problems are real but temporary, and the geopolitical shock is a buying opportunity rather than an accelerant.

IX. THE BOTTOM LINE

We published our CoreWeave report four days ago, mapping the financial stress in the AI supply chain’s most leveraged link. The core finding was that CoreWeave sits at the center of a time arbitrage — borrowing money to buy GPUs, racing to deploy them before interest eats the margin — and that the arbitrage is losing to the clock.

We did not anticipate that within a week, two exogenous risks would emerge along the same chain. But the transmission mechanisms are straightforward:

If Hormuz stays closed and oil spikes, that is inflationary — which means higher rates for longer, which means CoreWeave’s blended interest rate does not decline as hoped, which means the interest coverage squeeze gets worse. This is the higher-probability mechanism. The Strait is already closed. The oil price is already moving. The only question is duration.

If China tests shipping lanes in the western Pacific — a lower-probability scenario, but one with new supporting signals as of today — that threatens the TSMC-to-NVIDIA-to-CoreWeave supply chain. GPU delivery delays would mean Construction in Progress grows and carrying costs compound against zero revenue.

This is not a prediction about the war’s outcome or China’s strategic intentions. It is a mapping of transmission mechanisms that connect geopolitical events to a specific balance sheet — one carrying $21.4B in debt, operating at 1.5x interest coverage, and dependent on a supply chain that starts in Hsinchu and passes through two bodies of water that are, as of this morning, under new pressure.

The chain we mapped does not end at TSMC. It extends through the Strait of Hormuz and into the western Pacific.

APPENDIX: METHODOLOGY & SOURCES

This report was produced within hours of the first strikes using four independent intelligence channels, each covering a distinct information domain. The analytical value comes from the intersections — where a PLA Daily cartoon confirms a GDELT sentiment spike, or where a CRS background paper contextualizes a breaking headline.

Source Channels

Chinese State Media Analysis. Direct scraping and translation of five Chinese-language state media outlets (People’s Daily, Xinhua, PLA Daily, Global Times Chinese, CCTV), scored on a 1-5 posture scale with historical anchors. Includes monitoring for six specific tripwire phrases that have preceded every major Chinese military action since 1962, and analysis of Chinese vs. English editorial divergence within the same outlets. Divergence score for this event: 1.5 — the diplomatic core was faithfully translated, but Chinese domestic framing led with Iran’s effective retaliation and US overstretch commentary that never appeared in English editions.

Global Media Sentiment (GDELT). Cross-lingual search across 65 machine-translated languages. 403 articles across 30+ countries in 24 hours. Chinese-language tone distribution of 404 articles confirmed intense negativity (peak at -5 to -7 tone) without mobilization-level extremes. GDELT API was intermittently unavailable during this analysis — likely overwhelmed by global traffic — representing a known data limitation.

Premium Western Financial Media. Interactive Brokers news feed (Barron’s, Dow Jones, WSJ) queried across energy and defense tickers for institutional market framing.

Congressional Research Service. Unclassified intelligence briefings for members of Congress, providing historical and structural context on Iran’s Hormuz capabilities (R42335), the 2019 Abqaiq market-impact precedent (IN11167), China-Iran economic dependency (IN10829), US nuclear modernization costs (IF12964), and concurrent Syria withdrawal planning (RL33487).

Cross-Referencing

Major findings were triangulated across channels. The Hormuz closure assessment was confirmed across Xinhua Chinese-language reporting, GDELT articles from Turkey, Italy, India, and Romania, and IBKR/Barron’s energy coverage. The China escalation assessment was validated by matching the CCP Pulse posture score against the GDELT Chinese-language tone distribution and confirming the absence of tripwire terms. The US capacity-constraint argument was sourced from PLA Daily editorials, corroborated by CRS reports on Syria withdrawal timing and nuclear modernization costs.

Disclosure: We are short CRWV via a September 2026 bear put spread. We have no position in any other security mentioned in this report. This is not investment advice. This report relies on public sources including SEC filings, Congressional Research Service reports, Chinese state media translations, and the GDELT global media monitoring system. Geopolitical analysis is inherently uncertain. Do your own work.