A week in which a four-year-old secret became a document, a ceasefire became a pressure cooker, and a presidential visit to China produced ceremony with a great deal less substance behind it

The Cipher Was Real

The document has a classification stamp. It is marked “secret.” Across the top it reads “no circulation.” Dated March 7, 2022, designated cable I-0678, it was transmitted from Pakistan’s embassy in Washington to Islamabad by then-ambassador Asad Majeed Khan. On the afternoon of May 17, 2026, Drop Site News published it in full and without redaction, for the first time since it entered Pakistani political history four years ago as the most contested document in the country’s postcolonial record

DropSiteNews Imran Khan Cypher Article

299KB ∙ PDF file

Download

Download

What the cable documents is a luncheon meeting between Asad Majeed Khan and Donald Lu, then the US Assistant Secretary of State for South and Central Asian Affairs. What Lu told the ambassador that afternoon was transmitted back to Islamabad as a classified diplomatic report. What followed, within weeks, was the no-confidence vote that removed Prime Minister Imran Khan from power in April 2022.

The Drop Site investigation, written by Waqas Ahmed, Murtaza Hussain, and Ryan Grim and based on leaked documents alongside interviews with former civilian and military officials, does not confine itself to the cipher alone. It traces the entire arc of the US-Pakistan relationship across five years: from the CIA director William J. Burns flying to Islamabad in June 2021 and waiting a full day before being told by phone that Khan would not see him, to the Field Marshal Asim Munir now described by Donald Trump as “my favorite Field Marshal.” The distance between those two points is the story.

Burns had come to Pakistan seeking territory for US drone bases to be used against targets in Afghanistan after the planned American withdrawal. He left with neither the bases nor the meeting. Khan had already told Axios co-founder Jonathan Swan, in plain language: “Absolutely not. There is no way we are going to allow any bases, any sort of action from Pakistani territory into Afghanistan.” Kabul fell to the Taliban weeks later. The US evacuation became a political catastrophe for the Biden administration. The rupture between Washington and Islamabad hardened.

What the cipher documents, in the reconstruction offered by Drop Site, is the next phase of that rupture. By early 2022, US officials had communicated to Pakistani interlocutors that “all would be forgiven” if Khan were removed. The cable records the specifics of that communication in classified diplomatic language. Drop Site released the document so that, in their words, “it becomes part of the public historical record.”

The political consequences inside Pakistan were immediate. Supporters of Khan, still imprisoned following multiple convictions that his Pakistan Tehreek-e-Insaf party describes as politically motivated, called the publication a vindication. What cannot be disputed by anyone reading the subsequent record is what followed Khan’s removal: under the military-backed government that replaced him, Islamabad reversed course on nearly every position Khan had held. Pakistan became a quiet supplier of artillery to Ukraine. It signed a mutual defence pact with Saudi Arabia that Khan had refused. The China-Pakistan Economic Corridor’s second phase stalled. Washington got everything it had wanted and could not extract while Khan governed.

Drop Site’s companion investigation includes a disclosure that Trump privately urged Asim Munir to “resolve” Khan’s detention during discussions last year. It also documents the irony embedded in the present moment: the same generals who engineered Khan’s removal to repair ties with Washington are now basking in international praise for mediating between the US and Iran, a role that required precisely the kind of independent regional positioning Khan was removed for attempting.

Drop Site also published an important caveat about the cipher’s provenance. The redactions next to the addressee list, as the outlet noted in a separate post, point both to the source of the leak and to who was not its source: the copy released by Drop Site did not originate from the prime minister’s office. The question of where it came from, and who held it for four years, is one the investigation has not yet fully answered.

The cipher’s release arrived on the same day that Pakistan’s Prime Minister Shehbaz Sharif told the Sunday Times that the country’s role as mediator in the US-Iran war is “one of the shining moments in our history.” Former Pakistani ambassador to the United States Masood Khan told the same paper: “I’ve had a long diplomatic career and I have never seen Pakistan on such a high pedestal.” The collision of those two narratives, the diplomatic triumph and the documentary exposure of the mechanism that produced it, defined the week more precisely than any other single development.

Read more

In January 2025, a company most Americans had never heard of released an open-source AI model trained for less than $6 million. Within weeks, $600 billion evaporated from Nvidia’s market capitalization in a single trading session. The model was called DeepSeek V3. Its reasoning variant, R1, published the same month, matched or outperformed the best available American AI systems on standard benchmarks. OpenAI’s Sam Altman had stated in 2023 that training a foundational model cost “much more” than $100 million. DeepSeek eventually published a peer-reviewed account of its R1 training cost in Nature: $294,000.

The standard interpretation of this moment focused on the AI race, on whether American companies had wasted billions, on whether export controls were working. That interpretation missed the structural point. DeepSeek’s engineers had been denied Nvidia’s most powerful chips. The October 2022 export restrictions barred China from H100s and A100s, and successive administrations tightened the controls with each cycle. What the constraint produced, unexpectedly, was a generation of Chinese AI engineers who learned to extract frontier-level performance from second-tier hardware through algorithmic innovation. Jensen Huang, Nvidia’s CEO, acknowledged the logic when DeepSeek V4 appeared in April 2026, trained partly on Huawei’s domestically produced Ascend processors: “The best AI researchers in the world, because they are limited in compute, also come up with extremely smart algorithms. The day that DeepSeek comes out on Huawei first is a horrible outcome.”

That day arrived. The chip Huawei produced was made by SMIC. The minerals processed for its components were refined in China. The export controls designed to slow China’s AI development had accelerated the vertical integration of China’s technology stack. Washington’s theory of leverage, applied for three years, produced the outcome it was designed to prevent.

This is not a story about AI. It is a story about what happens when a country that makes things is forced, by the country that designs things, to make better things.

The Bet

American manufacturing employment peaked at 19.6 million in June 1979. That same year, the Federal Reserve raised rates to 20 percent to break inflation, crushed export competitiveness, and accelerated the offshoring of industry that had been running since the mid-1960s. By 2023, manufacturing employed approximately 13 million Americans in an economy more than three times the size of 1979’s. Its share of US GDP had fallen from roughly 25 percent in 1970 to just over 10 percent today.

The policy consensus that produced this had intellectual architects and institutional champions. The argument, made with confidence throughout the 1980s and 1990s by the economists designing American trade policy, ran: manufacturing is the low-margin activity in any value chain. Design, intellectual property, financial services, and brand management are where returns concentrate. A wealthy country should specialize in high-value activities and let labor-abundant countries perform assembly. The factory was not something to mourn. It was something to export.

NAFTA in 1994 sent a portion of it to Mexico. China’s entry into the WTO in December 2001 sent the rest. Between 1991 and 2007, the value of Chinese imports into the United States grew by 1,156 percent. MIT economist David Autor and his colleagues documented what followed: Chinese import competition accounted for roughly 55 percent of the decline in US manufacturing employment between 2000 and 2007, attributing somewhere between one and 2.4 million jobs to the surge in Chinese goods, depending on methodology. The communities that lost those jobs have not recovered. Autor’s follow-up research showed that economic damage to exposed labor markets persisted intergenerationally, in ways no adjustment assistance program had anticipated.

The architects of the policy were not wrong that the design premium exists. Apple’s gross margin exceeds 40 percent. Nvidia’s valuation, before DeepSeek, reached $3.6 trillion. The premium on design is real. What the theory did not account for is the difference between capturing value in a supply chain and controlling one. A country that holds the design but not the manufacturing exists at the discretion of the countries that make what it draws. When those countries learn to draw it themselves, the premium evaporates.

What the Factory Built

In 1995, a chemist named Wang Chuanfu founded a battery company in Shenzhen using borrowed equipment and a staff of twenty. The company was BYD. By 2003 it had acquired a failing state-owned automaker. By 2024, BYD had sold more electric vehicles globally than any other company, including Tesla. It designs its vehicle software in-house. It manufactures its own battery cells using proprietary Blade Battery chemistry, processes battery materials through subsidiary supply chains, and builds cars in mega-factories across China, Thailand, Hungary, Brazil, and Turkey. The $143 billion that Chinese EV and battery companies collectively invested in overseas ventures between 2014 and 2025, per Rhodium Group, has planted that integration model across five continents.

Contemporary Amperex Technology, CATL, followed a parallel trajectory. Founded in 2011 on the back of China’s domestic EV push and state procurement, CATL achieved 37.9 percent of global EV battery market share by 2024. BYD held 17.2 percent. Together they control 55.4 percent of the global EV battery market, ahead of the combined share of LG, SK On, Samsung SDI, and Panasonic. China controls more than 70 percent of global electric vehicle production. EV market penetration inside China reached 48 percent in 2024, a figure no European market will approach for years. CATL’s annual production capacity reached 646 gigawatt-hours in 2024. It is constructing super-factories in Hungary and Spain to supply European automakers who have no competitive domestic battery alternative.

China’s solar manufacturers control more than 80 percent of global production capacity. DJI holds approximately 70 percent of the global commercial drone market. Huawei supplies telecommunications infrastructure to more than 170 countries. Each of these positions was built through the same sequence: state procurement creates a protected domestic market large enough to drive scale; scale drives costs below what competitors sustain; cost leadership captures export markets; export revenue funds the next generation of capability. The factory does not merely produce goods. It produces engineers who understand the production process well enough to improve it, supplier firms with specialized capabilities that seed adjacent industries, and institutional knowledge that compounds into technological leadership over decades.

Ha-Joon Chang, whose books on industrial policy were largely ignored by the Washington policymakers implementing their opposite, called this the ladder argument: every country that industrialized successfully did so by protecting infant industries, and every country that is already industrialized has since argued that protection is illegitimate. China read it as a prescription.

The Chip War as It Actually Stands

The standard American narrative of the semiconductor conflict holds that export controls are working: China cannot access EUV lithography machines from ASML, and without EUV, it cannot manufacture chips below 5 nanometers. The argument is technically accurate and strategically incomplete.

SMIC demonstrated 7-nanometer chip production in 2022 using older deep ultraviolet immersion machines running multi-patterning, printing the same circuit layer multiple times to achieve finer resolution than the equipment was designed for. Yields were poor: 20 to 40 percent, against TSMC’s 7-nanometer yields above 90 percent. The unit economics were bad. None of that stopped Huawei from releasing the Mate 60 Pro in August 2023 with a SMIC 7-nanometer Kirin 9000s processor, an event the US government, which had been publicly confident its controls had capped Chinese production at 14 nanometers, did not predict.

By early 2026, SMIC had entered pilot runs on a 5-nanometer-class process, targeting mass production for Huawei’s Ascend AI processors and Alibaba’s data center chips. An April 2026 American Enterprise Institute report by Ryan Fedasiuk and Julia Torres documented how SMIC had legally purchased approximately 90 ASML NXT:1980Fi deep ultraviolet immersion machines in 2024, worth $5 to $7 billion, before any restrictions on that equipment class were imposed, representing roughly 20 to 25 percent of the global installed base of those machines. China simultaneously aims to quintuple its 7-nanometer and 5-nanometer wafer output to 100,000 per month within two years, with a longer-range target of 500,000 wafers by 2030.

In a Shenzhen laboratory, a prototype extreme ultraviolet lithography machine, developed by Chinese state researchers, was completed in early 2025 and is undergoing testing. The machine is not factory-ready. EUV production tools involve engineering challenges the prototype stage does not yet test. The existence of the prototype, built without access to the Western supply chain for EUV components, confirms that China does not accept the technology cap as permanent.

Meanwhile DeepSeek V4, released in April 2026 and trained partly on Huawei’s Ascend processors made by SMIC, performs within three to six months of the leading American frontier models on standard benchmarks, at an inference cost of $3.48 per million tokens compared to OpenAI’s $30. China’s semiconductor self-sufficiency rate was approximately 33 percent in 2024 and is estimated at 50 percent in 2025. The export controls are slowing the rate at which China closes the gap. They are not preventing its closure. What they have additionally done is force the algorithmic efficiency breakthroughs, the software optimizations for constrained hardware, that DeepSeek now offers to the world at one-tenth the price of its American equivalents.

The Pentagon’s Supply Line

The military consequence of deindustrialization is the part of this story Washington has been slowest to absorb. Greg Hayes, Raytheon’s CEO, stated it plainly in 2023: “We have several thousand suppliers in China. Decoupling is impossible. We can de-risk but not decouple.”

A 2024 Govini analysis found more than 80,000 components across 1,908 Pentagon weapons systems that contain antimony, gallium, germanium, tungsten, or tellurium, five minerals whose global refining and processing China dominates. China controls 98.8 percent of the world’s refined gallium output. Until 2023, it accounted for 99 percent of global heavy rare earth processing. The F-35, the Virginia-class submarine, the Tomahawk cruise missile, the Predator UAV, advanced radar and missile guidance systems: all depend on supply chains running through Chinese refining infrastructure. The Navy’s exposure is sharpest, with more than 91 percent of its weapons systems containing at least one of the five critical minerals.

Beijing has turned this position into sequenced coercion. Export controls on gallium and germanium arrived in July 2023. Tungsten and tellurium in February 2024. Antimony in August 2024. Seven heavy rare earths, including dysprosium and terbium essential to the permanent magnets in fighter jet engines and missile guidance, in April 2025. The system is not a blunt embargo. Chinese authorities can approve, delay, or deny export license applications case by case, tightening or loosening supply in response to specific diplomatic pressure without triggering WTO violation proceedings. By May 2025, Chinese rare earth magnet exports had fallen 74 percent year-on-year. Bill Lynn, CEO of Leonardo DRS, testified that his company’s germanium reserves had reached safety stock levels. On April 7, 2026, China’s State Council published Order No. 834, creating a unified legal framework for supply chain security, monitored by fifteen government agencies, authorizing legal action against companies deemed to harm Chinese supply chains.

The Pentagon has committed $439 million to domestic rare earth supply chains since 2020. Its goal, set in the 2024 National Defense Industrial Strategy, is a complete mine-to-magnet supply chain by 2027. As of January 2025, the first domestically produced defense-grade dysprosium oxide, processed at a research facility in Colorado, was described by the producing company as a breakthrough. Stockpiles of critical minerals across the defense industrial base are measured in months.

The Arsenal of Democracy that Franklin Roosevelt invoked in December 1940 rested on industrial capacity that converted automobile plants to tank production within months, that outproduced the Axis in steel, aluminum, ships, and aircraft within two years. The country that holds that rhetorical inheritance is now producing its first dysprosium separation facility as a breakthrough, not as a baseline.

The Design Layer and Its Limits

The argument Washington makes to itself is that the design layer remains an insurmountable American lead. Nvidia, Apple, Qualcomm, and AMD hold chip architecture. OpenAI, Google, and Anthropic hold frontier AI. Synopsys and Cadence hold the electronic design automation tools without which no advanced chip can be designed. China cannot reach the cutting edge without EDA tools that are American, chip architectures that are American, and a foundry with capabilities that only Taiwan’s TSMC can currently provide at scale.

This is currently true, and it is becoming less true each quarter. Empyrean is building China’s domestic EDA ecosystem. SMIC is demonstrating 7-nanometer production and targeting 5-nanometer scale. DeepSeek V4 runs on Huawei Ascend, which runs on SMIC silicon, trained on an architecture developed under hardware constraint into something competitive with what years of unconstrained American spending produced. The AEI report on SMIC’s DUV lithography loophole noted that existing Bureau of Industry and Security controls are drawn at technical thresholds that SMIC’s NXT:1980Fi machines sit just outside, meaning the legal structure of control does not match the operational capability it was intended to restrict.

China’s manufacturing value-added reached $4.66 trillion in 2023, representing 28 percent of global output, more than the United States, Japan, and Germany combined. Manufacturing constitutes 25 percent of China’s economy. Every year spent at that share of GDP inside its own industrial base compounds the process expertise, supply chain density, engineering workforce, and equipment capability that eventually produces SMIC’s 7-nanometer yield, then its 5-nanometer pilot, then its EUV prototype. The gap narrows not linearly but through engineering breakthroughs, each enabled by the manufacturing base accumulated in the decade before it.

The CHIPS Act commits $52 billion to US semiconductor manufacturing over five years. China’s Big Fund has exceeded $50 billion in two tranches, a third estimated at $47 billion, alongside provincial and municipal subsidies that push the total substantially higher. TSMC’s Arizona fab, the flagship American reshoring effort, recorded a loss of $441 million in its first year of commercial production, with manufacturing costs running 50 to 100 percent above Taiwan equivalents. The supply chain ecosystem that makes Hsinchu the world’s semiconductor center, specialized material suppliers, purpose-trained engineers, chemical processing infrastructure built for this specific industrial chemistry, does not transplant because a president demands it.

What Constraint Produced

The lesson Washington has not absorbed is the one DeepSeek delivered twice. Denied the best chips, China’s engineers optimized algorithms to do more with less. The optimization now competes with, and in efficiency metrics surpasses, what American engineers built with unrestricted access to the world’s best hardware. DeepSeek charges $3.48 per million inference tokens. OpenAI charges $30. This is not a pricing strategy. It reflects a genuine difference in computational efficiency built under pressure, now offered to the world as open-source software that any country can run.

The same dynamic runs through the rare earth controls. When Beijing restricted gallium and germanium in December 2024, it forced US defense contractors to confront a dependency that had been predicted in policy papers since at least 2010, when the US Magnetic Materials Association first raised formal warnings, and that produced no structural response across four administrations and fifteen years. The crisis arrived exactly on schedule, for those who kept the schedule. It arrived as a surprise for those who did not.

In both cases, American policy imposed a constraint on China, and China responded by building the capability the constraint was intended to prevent, while revealing a structural American vulnerability the constraint had obscured. The factory that America gave up is now the weapon. The design advantage that America kept is now the target.

Whether a country can rebuild in five years what it gave up across fifty, before the country that kept the factory finishes building the capacity to design what it makes, is the question this decade will answer.

Prime Minister Narendra Modi was in Abu Dhabi on a state visit when UAE President Sheikh Mohamed bin Zayed Al Nahyan presented him with a Cerebras chip enclosed in glass. The chip is not decorative. It is the foundational processing unit of Condor Galaxy India, an 8-exaflop AI supercomputing cluster comprising 64 Cerebras CS-3 systems, to be deployed in India in partnership with India’s Centre for Development of Advanced Computing. The commercial framework was formalized on May 15, 2026, during Modi’s official state visit to Abu Dhabi, the same visit in which the UAE announced $5 billion in fresh investments in India spanning energy, defence, infrastructure, shipping, and advanced technology, and in which both governments reaffirmed a strategic defence partnership. The project had been in development since January 2026, when MBZ visited New Delhi and the two governments first announced the intention; the term sheet with C-DAC was signed at the India AI Impact Summit 2026, and Modi’s Abu Dhabi visit closed the commercial terms.

You present something in a glass case to someone whose future you have decided to share.

Pakistan does not appear in this press release, in G42’s announced commercial frameworks, or in any bilateral AI cooperation agreement between Gulf states and South Asian nations. The government in Islamabad issued no public statement on the Condor Galaxy India announcement. The Ministry of IT issued no statement. This is not a diplomatic oversight. It is a categorization, and categorizations do not respond to diplomatic protest.

The Cerebras CS-3 is the largest chip ever manufactured, occupying an entire 300-millimeter silicon wafer. Where conventional GPU clusters are throttled by the bandwidth required to move data between chips, the CS-3 eliminates that bottleneck entirely by housing 44 gigabytes of on-chip SRAM directly on the die. Each unit carries 4 trillion transistors, 900,000 AI cores, and produces 125 petaflops of peak AI performance. Each system can be configured with up to 1,200 terabytes of external memory. Sixty-four of them, networked through Cerebras’s proprietary SwarmX interconnect, produce 8 exaflops of AI compute. G42’s own executives describe what this architecture produces as “sovereign governed nation-scale intelligence.” The contractual framework does not foreclose expansion.

The sectors formally named in the G42-India framework are health and genomics, energy, and geospatial analytics. No military application is declared. None needs to be. Geospatial analytics on an 8-exaflop cluster is not the same instrument it was on the processors India operated a decade ago. The CS-3’s architecture, designed specifically to handle large-scale vision and language models simultaneously, makes it the natural platform for any program that needs to process satellite imagery at national scale, model terrain at resolution, or run autonomous targeting inference under time pressure. India’s defence establishment does not need to formally integrate Condor Galaxy India into military programs to benefit from the civilian researchers who will train foundational models on this infrastructure. The models move across sectors. Fine-tuning is a question of months, not years.

Pakistan’s most advanced supercomputing nodes, distributed across public universities and the National Centre for Physics in Islamabad, do not collectively approach the output of a single Cerebras CS-3. India is receiving 64 of them. This is not a competitive gap. It is a different order of existence in the coming AI age, and it is being installed sixty minutes by air from Lahore.

The UAE’s economic relationship with Pakistan is not hostile. It is structural, and the structure has always been the same: Pakistan exports labor, the UAE imports it, remittances flow home, and the arrangement is framed in the language of Islamic fraternity and Gulf solidarity. Millions of Pakistanis work in the Emirates. Remittance receipts serve as a chronic subsidy for an economy that cannot generate sufficient foreign exchange through trade or domestic investment. The arrangement works, at its level. Pakistan sends people. The UAE sends money. The relationship has never required the UAE to treat Pakistan as a strategic equal.

Modi’s Abu Dhabi visit made the distinction explicit. The UAE announced $5 billion in investments in India. It signed a strategic defence partnership. It finalized petroleum reserve and LPG supply agreements. It handed India 8 exaflops of Cerebras compute through G42 and C-DAC. UAE-India bilateral trade crossed $100 billion in fiscal year 2024-25, against a target to double that figure to $200 billion by 2032. The UAE is India’s third-largest trading partner. India is, in the UAE’s accounting, a peer in the emerging digital and economic order. Pakistan is a labor corridor, and the remittance corridor will continue, because it serves the UAE as much as it serves Islamabad.

This is not betrayal in Abu Dhabi’s accounting. It is categorization. A categorization is harder to reverse than a betrayal because it requires no repair of feeling, only a fundamental change in what Pakistan represents as an investment thesis. Pakistan has not begun to produce that change. It has not named the problem publicly. It has not committed to a sovereign AI program. It has not built the institutional infrastructure that would make it legible to G42 as a technology partner rather than a labor source. The categorization will hold until Pakistan does something to disturb it, and nothing in Islamabad’s current posture suggests disturbance is imminent.

G42’s trajectory over the past three years tells a story Islamabad has not publicly reckoned with. In 2023, the White House applied sustained pressure on G42 over its ties to Chinese technology firms. G42’s then-CEO Peng Xiao stated publicly that the UAE “cannot work with both sides.” What followed was documented: divestment from Chinese partnerships, removal of Huawei equipment from G42 data centers, and a sustained courtship of American AI hardware access. The Condor Galaxy India deployment is the return on that realignment. By anchoring to Washington’s technology architecture and routing frontier compute into India, G42 secured continued access to American AI hardware and positioned the UAE as the indispensable node in the Gulf’s AI supply chain.

This creates a structural problem for Pakistan that sits precisely at the intersection of its two most consequential bilateral relationships. Pakistan’s deepest infrastructure partnership is with China: BRI commitments, Huawei’s presence throughout its telecommunications backbone, Chinese engineering across CPEC’s road and energy corridors. China is the actor G42 explicitly distanced itself from to secure access to American AI hardware. Pakistan’s alignment with Chinese digital infrastructure, which is real and deepening, is the same alignment that places it outside the technology architecture Abu Dhabi chose.

China offers Pakistan an alternative through BRI’s digital corridor commitments and Huawei’s continued presence. But the Chinese AI compute ecosystem operates under its own constraints, shaped by American export controls on advanced semiconductors that have forced Beijing’s chipmakers onto alternative development paths. The chips China can transfer to Pakistan are not equivalent to what Cerebras has built. Pakistan’s AI partnership with China is genuine, but it is partnership inside a constrained ecosystem. The Condor Galaxy India deployment sits on the frontier. That distinction does not diminish with time. It widens.

Pakistan’s military planners are not naive about asymmetry. The concern in Rawalpindi is not that India will immediately field an AI weapons system derived from Cerebras hardware. The concern is the asymmetry of foundation. India will spend the next five years training sovereign AI models at a scale Pakistan cannot approach. Those models, regardless of their original sectoral designation, establish the computational vocabulary from which every future application is built. When Pakistan’s defence establishment eventually moves to build AI-assisted systems, it will be building on a foundation that is years behind, in a domain where the gap between early foundation-builders and late entrants is not linear. It compounds.

The sectors named in the G42-India framework compound the concern further. Geospatial analytics at 8-exaflop scale means satellite imagery processed at resolutions and speeds that transform border surveillance, terrain modeling, and logistics planning. Energy modeling at this scale means grid optimization and supply-chain simulation capacities with direct military application. Genomics at this scale means medical intelligence capabilities that affect force readiness. None of this requires a separate military AI program. The civilian infrastructure does the work. India’s defence planners understand this. They built the framework knowing it.

Pakistan’s response, as of May 16, 2026, has been silence.

Pakistan’s inability to invest in computational infrastructure is not a failure of vision alone. It is the consequence of an economic architecture that has spent decades servicing debt, sustaining a military that consumed 18 percent of government expenditure in 2024, and managing IMF conditionality that constrains public investment as a condition of each successive tranche. The Atlantic Council’s analysis of UAE AI strategy, published in May 2026, identifies the UAE’s approach as built on strategic alignment with the US technology architecture. Pakistan does not feature in that analysis. It is not mentioned in G42’s announced partnerships, in the Condor Galaxy network’s expansion logic, or in any bilateral AI cooperation framework between Gulf states and South Asian nations.

The loop is closed and self-reinforcing. Pakistan’s fiscal constraints limit its investment in computational infrastructure. That limit makes Pakistan unattractive to the kind of technology partnership G42 is building. The absence of that partnership limits the compute Pakistan can access. Limited compute limits Pakistan’s capacity to develop the AI economy that would generate the growth to loosen the fiscal constraints. Islamabad can negotiate trade agreements and security partnerships. It cannot negotiate its way into a compute capacity it lacks the institutional infrastructure to absorb. The Condor Galaxy India deal tightens this loop in a domain that will determine economic and military capacity for the next generation.

The IMF does not model AI compute capacity in its conditionality frameworks. The State Bank of Pakistan does not track the compute differential with India in its quarterly reports. The Ministry of IT does not maintain a sovereign AI program with a defined exaflop target. These absences are a document. They record what the Pakistani state has treated as peripheral across administrations, across crises, across the years when the infrastructure being installed in India was still affordable to compete with.

The institutional history of India’s Centre for Development of Advanced Computing is worth sitting with. C-DAC was established in 1988 specifically because Cray Research, under pressure from Washington, refused to sell India supercomputers. The engineers who built the PARAM 8000 in a Pune laboratory in 1991, when the world refused to sell India what it needed, built something more than a machine. They built an institutional reflex: technology dependency is a strategic vulnerability, and the answer to that vulnerability is technology sovereignty. That reflex has guided C-DAC’s institutional culture through every subsequent administration.

India spent three decades building toward a moment when it would not be refused. It now signs commercial terms for 8 exaflops of Cerebras compute, backed by UAE capital and cleared by Washington’s technology architecture, at an institution created specifically to ensure that 1988 cannot happen again. The Condor Galaxy India deployment is not India acquiring a tool. It is India completing an arc.

Pakistan has no equivalent arc. Pakistan’s technology institutions were not built around sovereign ambition of that kind. They were built, like most of Pakistan’s civilian institutions, around the management of dependency rather than its dissolution. There is no Pakistani institutional equivalent of C-DAC whose founding memory is the experience of being refused advanced technology and whose organizational purpose is to ensure the refusal cannot be repeated. In 1988, Pakistan and India both existed inside the same international technology denial regime. India’s response was to build. Pakistan’s response was to buy what it could and manage what it could not. The divergence from that moment has not closed. The Condor Galaxy India deal is what the open end of a thirty-eight-year divergence looks like.

For three decades, Pakistan counted on the Gulf states as a structural counterweight in South Asia. Islamic solidarity, labor ties, and financial dependency created a relationship that Pakistan’s foreign policy community treated as a reliable element of its regional position. The assumption was that the UAE and Saudi Arabia would not be neutral in a strategic contest between Pakistan and India because the contest carried civilizational weight and the Gulf’s Muslim populations created political constraints on alignment with New Delhi.

That assumption has not survived contact with the new digital order. The Condor Galaxy India deal, read alongside the UAE-India digital MOU of 2024, the strategic defence partnership signed during Modi’s Abu Dhabi visit, the $5 billion investment package, and G42’s documented pivot toward Washington’s technology architecture, records a Gulf realignment that has been developing below the level of official statements for years. Abu Dhabi is not choosing India over Pakistan in the sense of renouncing a relationship. It is building a technology and investment partnership with India that has no Pakistani equivalent, while maintaining the remittance relationship with Pakistan that has always defined the ceiling of what Pakistan represents to the Gulf. MBZ welcomed Modi as a brother and presented him a chip in glass. Modi invited MBZ to visit India. The choreography is its own argument.

Pakistan’s foreign policy establishment has not produced a public analysis of what this realignment means for its long-term position. The Ministry of Foreign Affairs issued no statement on the Condor Galaxy India announcement. The silence is its own kind of statement: that Islamabad does not yet have the vocabulary to name what it is watching happen.

The framework was formalized on May 15, 2026. Analysts of South Asian strategic competition will return to this date. They will note that India’s sovereign AI program was seeded with 8 exaflops of Cerebras compute, operated through C-DAC, backed by $5 billion in UAE capital and investment, cleared by Washington’s technology architecture, and signed with the language of joint R&D across health, energy, and geospatial analytics. They will note that Pakistan held no equivalent asset, sought no equivalent partnership, launched no sovereign AI program, and produced no public analysis of what it was watching happen across the border and across the Gulf.

The press release did not say any of this. Press releases never do. What it says is that 64 Cerebras CS-3 systems, each housing the largest chip ever manufactured, will be installed inside India’s national compute infrastructure, backed by an institution created in 1988 specifically so India would never again be refused. The math does not require commentary.

The question is not whether Pakistan can replicate what India signed in Abu Dhabi. It cannot. The question is whether anyone in Islamabad understands what not having it will cost.

A six-part investigation into the theological system that has governed US policy toward Israel, Palestine, and the wider Middle East for a century. From seventeenth-century English Puritan commentaries to a fraudster’s annotated Bible to the Jerusalem embassy, this series traces how a reading of scripture became the single most durable constraint on American foreign policy.

Part I: The Doctrinal Machine (1600s to 1909) Part II: From the Bible Conference to Truman’s Desk (1875 to 1948) Part III: Armageddon in the White House (1967 to 1992) Part IV: The Lobby Matures, Clinton to Bush (1993 to 2009) Part V: The Institutional Zenith, Obama Through Gaza (2009 to 2024) Part VI: Huckabee and the Destination

The Doctrinal Machine (1600s to 1909)

In February 2026, the United States Ambassador to Israel sat across from Tucker Carlson and was asked a direct question about the biblical land promises. Should Israel take all the territory God had promised Abraham, stretching from the Nile to the Euphrates, a territory that would on any honest map include Lebanon, Syria, Jordan, and large parts of Saudi Arabia and Iraq? Mike Huckabee, a Southern Baptist minister before he was an Arkansas governor, and a governor before he became US envoy to Jerusalem, did not hesitate. “It would be fine if they took it all,” he said.

He called it hyperbolic afterward. A joint statement of condemnation arrived from more than a dozen Arab and Muslim governments. An anonymous State Department official told Politico that Huckabee did not represent the administration’s better instincts on Israel. Donald Trump said nothing. Huckabee remained in post and continued conducting US foreign policy in a country whose military campaign in Gaza had by that point killed over fifty thousand Palestinians and was before the International Court of Justice on genocide charges.

The episode was not a scandal. It did not end a career. It produced one news cycle and was absorbed. This is the most significant fact in it: a US ambassador could announce a biblical mandate for Israeli control of the entire Middle East on a widely watched podcast and face no professional consequence. The absence of consequence is not negligence. It is the outcome of a century of institutional work by a theological movement that has, through church networks, Bible institutes, mass-market publishing, television ministries, electoral mobilisation, and direct congressional lobbying, embedded its assumptions so deeply into the vocabulary of American political life that positions which would have seemed extreme two generations ago now pass as standard Republican foreign policy.

To understand how that happened, the story begins not in Washington, not in modern Israel, and not in the twentieth century. It begins in England, in the 1600s, in the margins of Puritan biblical commentary, with the question of what God had promised Abraham and whether that promise remained in force.

The Puritan Foundations

Thomas Brightman was a Church of England minister in Bedfordshire who died in 1607, three years before his most important work was published. His commentary on the Book of Revelation appeared in a Latin edition in 1609 and in English translation in 1611. It argued against the dominant allegorical tradition in Protestant biblical interpretation. Where most of his contemporaries read the Old Testament prophecies about Israel as spiritual metaphors for the Christian church, understanding Israel’s restoration as the conversion of the Gentiles and Jerusalem as the heavenly city, Brightman insisted on a different reading. The promises made to Abraham’s descendants were literal and national. The Jews would physically return to the physical land of their ancestors. This was not a metaphor for something else. It was a concrete future event, prophetically necessary, awaiting fulfilment.

Henry Finch, a lawyer and Member of Parliament, developed the argument further in 1621 in a treatise titled “The World’s Great Restoration, or Calling of the Jews.” Finch drew on dozens of Old Testament passages to establish that the covenants God had made with Israel had not been transferred to the Christian church, that they remained binding, and that a literal political restoration of the Jewish people to Palestine was both scripturally mandated and historically approaching. King James I had Finch imprisoned briefly for the political implications: if the Jews were to inherit their ancient territory as a matter of divine decree, what did that imply about the Ottoman and Christian kingdoms currently occupying it? But the text circulated, was translated, was read, and planted arguments that would be developed for three centuries before they shaped modern policy.

What Brightman and Finch established was not a fringe position. Within Calvinist and Puritan theology, the question of Israel’s future occupied serious scholarly attention. The Puritans who sailed to Massachusetts in the 1630s carried this interpretive tradition across the Atlantic. John Cotton, Increase Mather, Jonathan Edwards, and subsequent generations of New England Calvinist divines all engaged with the question of Jewish restoration to the land as a matter of serious scriptural expectation. America was being built, in part, by people who believed they were living inside a prophetic timeline, and Israel’s future was a node in that timeline. The theological infrastructure for American Christian Zionism was in place before the United States existed as a nation.

The Victorian Synthesis: Lord Shaftesbury

The translation of theological expectation into political action, the move from believing that Jews would return to the land to actually lobbying governments to facilitate their return, happened in Victorian England. Its central figure was Anthony Ashley Cooper, the seventh Earl of Shaftesbury.

Shaftesbury is remembered by posterity as the great Victorian social reformer: the man who drove child labour legislation through Parliament in the 1840s and 1850s, who championed the regulation of lunatic asylums, who spent decades advocating for factory workers, chimney sweeps, and the ragged school movement for destitute children. He was the most prominent lay evangelical in England for most of his adult life, intimate with cabinet ministers, present at every major political table. He was also, simultaneously and without apparent tension, the most active Christian Zionist figure in the British political establishment of the nineteenth century.

In 1838, Shaftesbury wrote a memorandum circulated among British government ministers arguing that Britain should use its expanding influence in the declining Ottoman Empire to facilitate a Jewish return to Palestine. He worked through his stepfather-in-law, Lord Palmerston, then Foreign Secretary, to establish a British consulate in Jerusalem that year. The consulate served, among other functions, as a protective office for the small Jewish population of the city. Shaftesbury lobbied for the establishment of the joint Anglican-Lutheran bishopric in Jerusalem in 1841, with Michael Solomon Alexander, a convert from Judaism, as its first bishop. He wrote in newspapers, addressed public meetings, and raised money for Jewish missionary work throughout a career that lasted until his death in 1885.

The theology behind this activity was the restorationist tradition running from Brightman and Finch through the evangelical revival movements of the eighteenth century. Shaftesbury believed that the Jewish people had a divine title to Palestine, that Britain bore a special responsibility toward “God’s ancient people,” and that faithful discharge of this responsibility would bring divine blessing on the empire. The historian Donald Lewis, in his Cambridge University Press study of Shaftesbury’s Christian Zionism, documents how this framework turned Shaftesbury’s evangelical circle into active political lobbyists in a way that had no precedent in the two previous centuries of restorationist theology. The belief had always been there. Shaftesbury organised it.

The consequences ran forward for eighty years. The climate of opinion that Shaftesbury’s network created in British evangelical and political life was, by the early twentieth century, sufficiently strong to make the Balfour Declaration politically viable. Arthur Balfour himself, who wrote the 1917 letter committing Britain to a Jewish national home in Palestine, came from the same evangelical cultural milieu. The political Zionists who lobbied for the declaration, above all Chaim Weizmann, were navigating a cultural atmosphere that Christian Zionism had spent decades preparing. They did not create the sentiment. They harvested a crop that English Puritanism had planted in 1621.

John Nelson Darby and the Systematic Architecture

The theological system that would eventually produce American Christian Zionism as a mass political movement was built by an Irish Anglican minister who left the established church in the late 1820s and spent the next fifty years rebuilding Protestant eschatology from first principles.

John Nelson Darby was born in London in November 1800. His middle name came from his uncle, the British admiral who served under Lord Nelson. His father was a merchant. He studied law at Trinity College Dublin and was called to the Irish bar before, in 1825, abandoning a promising legal career to be ordained as a priest in the Church of Ireland in County Wicklow. In 1827 he suffered a serious riding accident that kept him bedridden for months. During that convalescence, by his own account, the theological questions he had been circling since ordination took decisive form.

What he concluded, lying immobile through the Irish winter, was that the institutional church was corrupt beyond reform and that the New Testament described a completely different model of Christian community: gathered believers, no clerical hierarchy, no formal liturgy, no establishment relationship with the state, reading scripture together and waiting for Christ’s return. He left the Church of Ireland in 1828 and began travelling among small independent assemblies in Ireland and England, connecting groups of like-minded believers into what became the Plymouth Brethren movement. The movement took its name from the assembly in Plymouth, Devon, which grew to around a thousand members by the early 1830s and gave the network its public identity.

The theological system Darby developed over the following decades was called dispensationalism. Its central argument was architectural. History, in Darby’s reading, was divided into distinct epochs or dispensations, each defined by a specific covenant between God and humanity, each ending in human failure. The current era, the Age of the Church, was itself a parenthesis in God’s primary narrative. That narrative had been interrupted at the Jewish rejection of Christ and would resume, with full force, after the church was removed from the world in an event Darby called the rapture. After the rapture, a seven-year period of tribulation would unfold, centred entirely on Israel. The great powers would gather against Jerusalem. The Antichrist would rule. Two-thirds of the Jewish people living in the land would be killed in the wars of the end times. The surviving remnant would recognise Christ, convert, and be saved at his physical return to the Mount of Olives.

The political implications of this system were not incidental. They were structural. If the Old Testament land promises were literal and permanent, then any political arrangement denying Jewish sovereignty over the biblical territory of Israel was in conflict with the divine plan. If the sequence of end-times events required a restored and sovereign Jewish state in the promised land, then that state’s existence and expansion served God’s purposes in a way that no secular argument could match. If the nations gathering against Israel in the final battle were to be defeated and judged, then aligning with Israel was not a geopolitical preference but a matter of covenantal obligation for any nation wishing to stand on the right side of history’s final verdict.

Darby made seven extended visits to North America between 1859 and 1877, speaking at churches, gatherings, and proto-conference circuits across the United States and Canada. His audiences were largely clergy, educated laypeople, and Bible teachers. But the system he transmitted was intellectually coherent, internally consistent, and addressed a question that American Protestantism was actively wrestling with: what role did the Jewish people play in the prophetic programme, and did the founding of the modern state they had anticipated make any difference to Christian expectation?

The answer Darby had ready was fully formed and required no improvisation. The programme was written. The timetable was running. Anyone who could read the Bible and read the newspapers could see the alignment. This is the key to dispensationalism’s political resilience over a century and a half: it is not a system that requires proof. It is a system that frames everything as confirmation. Every new development in the Middle East is a milestone. Every Israeli military victory is a prophetic fulfilment. Every international pressure on Israel is an omen. The framework converts current events into eschatological evidence and makes its own conclusion unfalsifiable.

Cyrus Scofield and the Annotated Mandate

The transformation of dispensationalism from an elite theological current into a mass popular movement required a delivery mechanism capable of reaching ordinary American Protestants who would never attend a Bible conference, read a work of academic theology, or follow the Plymouth Brethren controversy. That mechanism was a study Bible, and its creator was among the least credentialed and most compromised figures in the history of American religious publishing.

Cyrus Ingerson Scofield was born in Michigan in 1843. He served in the Confederate army, was admitted to the Kansas bar through the assistance of a politically connected acquaintance, and was appointed US District Attorney for Kansas in 1873. He lasted six months in the post before resigning in a documented cloud of scandal involving extortion and blackmail. He forged the signature of his political patron to secure funds. He was arrested in Wisconsin in 1879 under both his own name and the alias “Charles Ingerson,” and was escorted back to St Louis to face charges of criminal forgery. He spent six months in the St Louis city jail. He defrauded his mother-in-law of $1,300 in gold, her entire life savings, by forwarding her a mortgage document signed by a fictitious “Charles Best.” He abandoned his wife Leontine and their two daughters Abigail and Helen, providing no financial support for years. In December 1883, the Kansas court granted Leontine’s petition for divorce on grounds of willful abandonment and issued a decree declaring Scofield “not a fit person to have custody of the children.” He had already taken up with a young woman from the St Louis Flower Mission and would subsequently marry a third woman, Helen van Ward.

While in prison, Scofield had read Darby. He emerged claiming a conversion experience and began, almost immediately, working his way into the ministry networks connected to Dwight L. Moody. By October 1883, five days before Leontine filed her final divorce petition, he was ordained as a Congregationalist minister in Dallas, Texas. He later told Who’s Who in America in 1912 that he had been decorated for valour in the Civil War. He had in fact deserted. He called himself “Doctor” though he held no degree. By the documented record, he was a systematic liar who built his religious reputation on a series of fabrications and never made restitution to those he had defrauded.

In 1901, eight years before publishing his study Bible, Scofield was admitted to membership in the Lotos Club in New York, an elite private literary society whose rolls included Grover Cleveland, Theodore Roosevelt, Joseph Pulitzer, and Mark Twain. There is no plausible explanation in his public biography for how a mid-sized Texas church pastor with no significant publications, no academic credentials, and a past that still occasionally surfaced in newspapers gained entry to this club. What is documented is that one of the members who approved his candidacy was Samuel Untermyer, a prominent Jewish corporate lawyer and committed Zionist who would later play a significant role in the Brandeis Supreme Court appointment and other major American Jewish political efforts. Scofield subsequently made research trips to Oxford, apparently funded by sources his biographers have never fully accounted for.

In 1909, Oxford University Press published the Scofield Reference Bible. This was not a new Bible translation. It was the King James Version with Scofield’s commentary printed directly alongside the text, in the margins and at the bottom of pages, cross-referenced throughout, in a format that had not been used for an English Bible since the Geneva Bible of 1560. The effect was to make Darby’s dispensational system inseparable from the scriptural text itself. A reader opening the Scofield Bible to Genesis 12:3, “I will bless those who bless you, and whoever curses you I will curse,” found Scofield’s note telling him this was a standing divine law governing the behaviour of nations toward the Jewish people, with blessing contingent on support and judgment following opposition. The note was printed in the same typeface, on the same page, as the revealed word of God. For millions of readers who had no reason to distinguish between Paul’s letters and Scofield’s commentary, the margin was the message.

The 1909 edition was revised by Scofield himself in 1917 and revised again by a committee of dispensationalist scholars in 1967. It remains in print. It has sold in the tens of millions. When John Hagee built Christians United for Israel into a ten-million-member lobbying organisation, his foundational theological argument was Genesis 12:3 read through Scofield’s margin. When Mike Huckabee cited divine land promises in his Tucker Carlson interview, he was speaking a vocabulary that Scofield had printed in a study Bible in 1909, the same year the Ford Model T first went to market.

The machine was built in a margin note, published by a prestigious academic press, written by a convicted fraudster from Michigan. Everything that follows is elaboration.

Part II: From the Bible Conference to Truman’s Desk. How the Niagara Bible Conference circuit industrialised Darby’s system, how the fundamentalist-modernist fracture drove dispensationalism into a subculture where it hardened undisturbed, how radio carried it to twenty million households, and what Truman was actually deciding on May 14, 1948.

References

• Thomas Brightman, Apocalypsis Apocalypseos (Frankfurt, 1609; English trans. 1611)

• Henry Finch, The World’s Great Restoration, or Calling of the Jews (London, 1621)

• Donald M. Lewis, The Origins of Christian Zionism: Lord Shaftesbury and Evangelical Support for a Jewish Homeland (Cambridge University Press, 2010)

• Crawford Gribben, The Puritan Millennium: Literature and Theology, 1550-1682 (Four Courts Press, 2000)

• Crawford Gribben and Mark Sweetnam, J.N. Darby and the Roots of Dispensationalism (Queen’s University Belfast, 2024)

• R. Mangum and M. Sweetnam, The Scofield Bible: Its History and Impact on the Evangelical Church (Paternoster, 2009)

• Joseph M. Canfield, The Incredible Scofield and His Book (Ross House Books, 1988)

• Stephen Sizer, Christian Zionism: Road-map to Armageddon? (Inter-Varsity Press, 2004)

• George Marsden, Fundamentalism and American Culture (Oxford University Press, 1980)

• “Cyrus I. Schofield in the Role of a Congregational Minister,” Topeka Daily Capital, August 27, 1881

Air Force One put down at Ted Stevens Airport in Anchorage on a refueling stop. This was the night before the Beijing summit. Jensen Huang, Nvidia’s CEO, was not on the original delegation list. His …

Read more

Just yesterday the S&P 500 crossed 7,500 for the first time in its history. The Dow Jones Industrial Average simultaneously breached 50,000, a level it had not seen since February, when the Iran war …

Read more

The defining image of this summit was not from the bilaterals inside the Great Hall of the People. On Wednesday evening, Elon Musk, whose estimated $688 billion fortune makes him the world’s wealthie…

Read more

From 1993 to 2023, 890 cases of hantavirus disease were reported in the United States, mostly in western states. There have been an estimated 60,000 to 100,000 cases worldwide each year, with China a…

Read more

On June 6, 2018, a DEA-vetted counter-narcotics unit in the Honduran department of Cortés acted on a tip and stopped a convoy on the highway. When agents searched a Volkswagen in the group, they foun…

Read more

Twelve men. Aggregate market capitalization north of ten trillion dollars. Aggregate lobbying spend in Washington: hundreds of millions annually. Aggregate influence over American foreign policy, tra…

Read more

… expression of interest in purchasing more American oil as a gesture toward supply diversification. Western correspondents, following the familiar grammar of US-China summitry, filed it as a concess…

Read more

The rickshaw driver in Lahore fills his tank. Petrol is Rs415 per litre this week. Last month it was Rs393. Before the war it was Rs270. He does not know exactly what the Islamabad Process is. He kno…

Read more

On the morning of January 28, 2026, three weeks into his mayoralty, Zohran Mamdani stood in the Blue Room at City Hall with a PowerPoint presentation and a word no administration uses in its first mo…

Read more

Air Force One touched down at Beijing Capital International Airport at 7:50 p.m. local time Wednesday. Three hundred Chinese children in blue and white uniforms lined the tarmac waving American and Chinese flags. Chinese Vice President Han Zheng, Xi Jinping’s designated receiver for diplomatic ceremonies, waited at the bottom of the stairs. The red carpet was out. The pageantry was precise. Nine years had passed since a sitting American president last walked onto Chinese soil, and the two countries had spent most of those years in escalating trade war, technology embargo, and strategic antagonism. None of that was visible on the tarmac.

Read more

The last American president who arrived in Beijing believing he held the decisive advantage left with a photograph and a memorandum of understanding worth $83.7 billion, a number that exceeded West V…

Read more

There is a Chinese proverb that Western commentators rarely quote when writing about this summit: 欲速则不达. The faster you chase, the further the destination recedes. Donald Trump has been chasing since…

Read more

The phrase arrives in the second paragraph, without qualification: “the US-Israeli war on Iran.” Not a conflict. Not a military operation. Not Washington’s preferred formulation, which has spent seve…

Read more

In November 2017, Xi Jinping gave Donald Trump the Forbidden City. A private dinner in the Hall of Embodying Benevolence, a procession through Tiananmen, a ceremony at the Great Hall of the People un…

Read more

On February 19, 2026, with Donald Trump seated at the head of the table at the U.S. Institute of Peace in Washington and forty-two nations arrayed around him, the Board of Peace held its inaugural me…

Read more

On the morning of May 6, Secretary of State Marco Rubio stood before reporters at the White House and declared that Operation Epic Fury had concluded. The objectives of the US-Israel war on Iran which opened on February 28 with the assassination of Supreme Leader Ali Khamenei and proceeded, over 38 days, to destroy what Washington claimed was more than 85 percent of Iran’s defense industrial base had been achieved. The United States now preferred, Rubio said, “the path of peace.”

By Sunday, May 10, Iran had launched drone strikes on Qatar, the United Arab Emirates, and Kuwait.

The sequence follows from the structure of a war that has been declared over without being settled. The formal military campaign wound down, but no resolution has been reached, and what replaced combat operations is a dual blockade in which neither party can deliver a decisive blow but both retain the capacity to impose costs on everyone in range. The US Navy maintains a cordon around Iranian ports. Iran continues to strangle traffic through the Strait of Hormuz, the 21-mile passage through which, before February 28, roughly one-fifth of the world’s oil and gas moved each day. Since the US launched its “Project Freedom” convoy operation to guide stranded vessels through the strait, exactly two commercial ships have successfully transited under American naval protection. Major shipping lines have declined to send vessels through a corridor sitting within range of Iranian missiles, drones, and mines whose precise locations Iran has admitted, in one extraordinary moment of operational honesty, it can no longer fully track. Trump paused Project Freedom on May 6, the same day Rubio declared the war concluded, then posted on social media that bombing would resume “at a much higher level and intensity” if Iran did not accept terms. The press conference and the social media post arrived within hours of each other, each addressed to a different audience.

Read more