For weeks the reopening of the Strait of Hormuz has been traded as a scheduling problem — a matter of a signature, a corridor, a date. On Monday the market stopped pretending. Brent crude rose 5% to $87.72 a barrel and West Texas Intermediate settled up 5% at roughly $82, not on any new attack or blockade, but on the plain reading of a statement Tehran had already made: the strait stays closed until the United States does six things, only one of which involves the sea.

Mohammad Bagher Zolghadr, secretary of Iran's Supreme National Security Council, laid them out over the weekend. The United States must never again threaten Iran or insult its beliefs. It must permanently end aggression against Iran and its allies in Lebanon, Palestine, Yemen and Iraq. It must lift the naval blockade and withdraw its ships and aircraft from the region. It must pay full reparations for two "imposed wars." It must lift all sanctions. And it must unconditionally release Iran's frozen assets. "The Supreme National Security Council will never back down," the statement read, "whether in war or in negotiations."

Read that list against the thing it is supposedly about — commercial tankers moving crude — and the mismatch is the story. A June memorandum had treated these same items as a phased, reciprocal process toward a final nuclear deal. Zolghadr has now converted them from the end of a negotiation into the price of admission for its first step. The waterway that carries roughly a fifth of the world's seaborne oil has been repriced as collateral against the entire American posture in the Middle East.

Washington's public position is narrower and, by design, cheaper. An unnamed U.S. official told Reuters the blockade of Iranian ports would lift once commercial shipping through Hormuz resumes "without impediments," with American steps tied to Iranian performance. Vice President JD Vance has said the administration expects Gulf flows to return to pre-war levels and that Iran privately signaled it would not impose tolls — while adding that the U.S. does not fully trust the assurance. Two governments are describing two different transactions and calling them the same deal.

The Oman-mediated corridor talks are the seam worth watching. Muscat calls them positive; Iran simultaneously warns that any corridor arrangement would not constitute a "reopening" in the full commercial sense. If the corridor can be operationalized independently of the six conditions, tankers move and the risk premium bleeds back out. If it cannot be separated, then the strait is now hostage to sanctions relief, reparations, and a regional troop withdrawal — none of which clears on an oil trader's calendar.

That is why Monday's tape matters more than its size. The S&P 500 slipped only 0.06% to 7,753.11 and the Dow eased 0.11% to 53,975.98, fractions off Friday's record. But the composition told the truth: energy led every sector while the 10-year Treasury yield climbed to 4.70%, up from 3.97% before the war with Iran. The market did not crash. It simply moved the Hormuz premium from the "resolved" column back to the "open" column, where Tehran has been insisting it belongs.

The through-line this newspaper has held all month applies cleanly here. Leaders announce; the world clears in physical fact. Gulf crude and condensate exports are still running well below pre-war levels — that is the fact. A president's confidence and a vice president's reassurance are the announcement. This week the two diverged loudly enough that a five-percent oil move did the arithmetic in public.

What to watch now is not rhetoric but transit counts. Do weekly vessel crossings stabilize or keep sliding? Does Iran's Supreme Leader endorse or soften the six-condition text? Does the EIA's next petroleum status report show further crude and distillate drawdowns? Those are the coordinates. Everything said from a podium between now and then is commentary on numbers that have already been written.

The pattern that runs through this issue — an announcement meeting a fact — has a second front this morning, and it is in Jerusalem. On Sunday, Prime Minister Benjamin Netanyahu told his cabinet: "I wish to be precise here: Israel does not accept the 15-point document." The document in question is the Board of Peace roadmap that President Donald Trump announced in late July as a "monumental step" toward ending the war in Gaza.

The dispute is sequencing, and sequencing is everything. Trump's plan envisions a phased withdrawal of Israeli forces moving in tandem with Hamas disarmament, supervised by an International Stabilization Force and a technocratic Palestinian administration. Netanyahu insists the order runs the other way: Hamas must undergo "genuine disarmament" and surrender all weapons before any pullback begins. "The army will not carry out any withdrawal until Hamas is genuinely disarmed," he said. Hamas, for its part, says it remains committed to the roadmap and has asked mediators to press Israel to hold to the terms.

What makes the rejection less than final is the calendar. Israel goes to a general election on October 27 — its first since the October 7 assault — and successive surveys show Netanyahu on course to lose. Analysts quoted by Al Jazeera and the ABC read his refusal as election positioning: a far-right coalition cannot be seen accepting a deal before the vote, which likely means little movement before late October and a real possibility that the next Israeli government inherits the implementation.

Washington's reaction was studied indifference. A senior U.S. official told Axios the White House "is not bothered" and views the statement as election-season theater. That composure has its own arithmetic: Trump faces November midterms with petrol and grocery prices elevated in part by a five-month war in Iran, an approval rating that has fallen, and a Republican Party bracing for losses. A visible foreign-policy win in Gaza would be useful to him too. Two leaders, two elections, one plan suspended between them.

The Board of Peace is not without leverage. It has secured $7 billion in reconstruction pledges from member states, and the United States has committed $10 billion of its own — a combined stake that gives Washington enormous say over Gaza's postwar political order. Money is patient. The question is whether a plan that both a sponsor and a signatory have reason to stall can survive two election seasons intact, or whether "breakthrough" joins "deal" as another word this year that described a press release rather than a place.

At 7:34 a.m. local time on Monday, the earth under western Colombia moved for two minutes. The magnitude 7.4 quake, centered near San José del Palmar in the Chocó department at a depth of about 110 kilometers, was felt across an enormous arc — into Ecuador, Panama and Venezuela — and, according to the U.S. Geological Survey, subjected roughly 1.5 million people to shaking strong enough to damage unreinforced masonry.

By Tuesday the confirmed death toll had climbed to at least 164, with more than 810 injured and over 3,200 people listed as missing — figures that rose through Monday from an initial 69 to 111 to 132 as rescuers reached collapsed buildings. Sixty people were killed in Pereira alone, three of them at Matecaña International Airport when it partially collapsed. At least 42 died in Valle del Cauca, the department that contains Cali, Colombia's third-largest city, where the mayor reported at least 20 buildings down with people trapped inside. President Abelardo de la Espriella declared a national emergency, assumed direct control of the response, and invoked special powers to reallocate budgets.

The damage reached the infrastructure a modern economy runs on. More than ten airports reported damage — Manizales, Quibdó, Armenia, Cartago, Buenaventura — and the civil aviation authority suspended flights at the affected fields. The Puracé volcano in Cauca emitted fumes after the shaking. More than 1,500 homes were damaged in the first count, a number that historically only grows.

There is a grim regional context. The quake follows back-to-back 7.2 and 7.5 tremors in Venezuela in late June that killed more than 5,000 people and, by some counts, over 6,000. Two of the deadliest natural disasters in the hemisphere's recent memory have now struck neighboring countries within seven weeks.

This paper spends most of its front page on the distance between what leaders declare and what the world executes. An earthquake is the limit case: a fact with no announcement, no negotiation, no press release to precede it. A depth of 110 kilometers is beyond the reach of any building code, sanction, or peace plan. What policy governs is only the aftermath — the speed of rescue, the honesty of the count, the rebuilding — and it is there, not in the seismograph, that Colombia's government will be measured in the days ahead.

Strip the drama out of Monday's tape and you are left with a market doing arithmetic in public. The S&P 500 slipped 4.53 points to 7,753.11, the Dow dipped 60.95 to 53,975.98, and the Nasdaq fell 85.26 to 26,605.36 — moves of roughly a tenth of a percent, a hair below Friday's records. That is not fear. It is a recalculation, and the variable that changed was oil, up 5% to $87.72 Brent on the fading hope of a near-term Hormuz reopening.

Why should a fifth of a percent in equities matter? Because of what it sits on top of. Higher oil pushes inflation up, and the 10-year Treasury yield climbed to 4.70% from 4.65% — a level that has already dragged mortgage and loan rates higher and is up sharply from the 3.97% that prevailed before the Iran war. Energy was the only sector to gain outright. Gold rose 1.14% to $4,390.85. This is the signature of a market hedging inflation, not celebrating growth.

The main event is Wednesday's July CPI. Consensus expects headline inflation to ease to 3.4% year over year from 3.5%, with core at 2.5%, and a soft monthly print of 0.1%. But the labor side has muddied the read: last week's jobs report showed the economy unexpectedly shed 23,000 jobs, cooling the case for a hike even as inflation stays above the Fed's 2% target. Traders now put September at roughly a coin flip — CME FedWatch shows about a 52% chance of a quarter-point hike, down from 67% a week earlier, while prediction markets lean toward a hold.

The complication is internal. The FOMC's July meeting drew three dissents in favor of a hike, and new Chair Kevin Warsh left conflicting signals. The Fed's own nowcast pegs July core CPI near 0.21% monthly — a mild reacceleration. If Wednesday runs hot on core services, the hawks gain a talking point six weeks before a September 15-16 meeting that is already, in the Fed's own framing, a toss-up.

The corporate tape carried its own tells. Intel fell 4% after announcing a dilutive $15 billion stock offering to fund AI and manufacturing — a company raising money the expensive way. Nvidia dropped 2.9% as investors eyed a $500 billion infrastructure funding package assembled with asset managers, worrying openly about how capital-hungry the AI buildout has become. Both moves rhyme with the bond market's message: capital is no longer free, and every ambitious plan now has a financing cost attached. That is the quiet story under a record index — not that stocks are high, but that the money underneath them has gotten expensive again.

Read the press release and it is a triumph of "physical AI." Read the structure and it is a confession. On Monday Boeing agreed to hand Archer Aviation three subsidiaries — the autonomous-flight maker Wisk Aero, the airspace-management software firm SkyGrid, and drone builder Insitu — in exchange for shares that will leave it holding roughly 16.5% of Archer after closing. No cash changes hands from Archer. Boeing is not selling these businesses so much as trading them for a minority stake in the company that will now run them.

The history sharpens the point. Boeing poured hundreds of millions into Wisk, including a $450 million injection in 2022, took full ownership in 2023, and settled a bruising trade-secret suit with Archer along the way. Wisk has flown six generations of eVTOL aircraft across more than 1,700 flight tests — genuine, hard-won engineering. And Boeing is now routing all of it into a former adversary's balance sheet, retaining only a cross-license so it can borrow Wisk's autonomy tech for its eventual 737 successor.

"We can now take all of the technologies and the IP and focus them back on our core Boeing products," said Brian Yutko, Boeing's product-development VP and Wisk's former CEO. Translated: Boeing has decided that autonomous air taxis and drones are not core, and would rather own a slice of someone else's execution than carry the cost of its own. The deal adds more than $200 million in annual revenue to Archer and hands it a ready pool of engineering talent for the hybrid-electric "loyal wingman" drone it is building with Anduril.

The market applauded the buyer more than the seller, which is usually the tell. For Archer — public since a 2021 SPAC, still pushing its Midnight air taxi toward commercial service under the White House's eVTOL pilot program — this is a step-change into defense and autonomy. For Boeing, it is the latest asset it has decided it doesn't want to build. Investors have spent years asking whether Boeing can still execute on ambitious new aircraft. On Monday, in the driest possible language, Boeing answered a narrower version of that question: on these three, it would rather not try.

On Monday Meta released Muse Glimmer, a 30-billion-parameter model with a 120,000-token context window, small enough to run on a Mac or PC with a single consumer GPU and licensed under the permissive Apache 2.0 terms. It is built for always-on local agents — coding, function-calling, task automation — and Nvidia says it can push 20,000 tokens per second on a single GPU. The weights are on Hugging Face today; Ollama, LM Studio, llama.cpp, MLX and vLLM support is landing in days. For developers who want a capable agent that never leaves their machine, this is a real gift.

Mark Zuckerberg wrapped it in roughly 6,500 words. The essay argues that advanced AI should empower individuals rather than concentrate in a few companies' hands, that distributing "superintelligence" widely could "begin a new era of personal empowerment," and that everyone will eventually have access to these tools. It positions Meta explicitly against OpenAI and Anthropic — labs that keep their frontier models proprietary — and frames open weights as both an American competitive answer to Chinese open-source models and a philosophy of governance.

The fine print is where the philosophy meets the org chart. Glimmer is essentially an open version of Muse Spark, Meta's most powerful closed model, which debuted in April. Spark 1.2, released August 5, remains closed-weight; Meta has promised to open its weights "in the coming weeks," a promise it has made and rescheduled before. As TechCrunch put it, Glimmer marks where Meta draws the line between the AI it wants people to own and the more powerful intelligence it keeps under its own control. Access, the piece noted, is not the same as ownership.

The business pressure is unsubtle. Meta shares are down about 10% this year as investors question capital spending that could reach $145 billion in 2026, and the stock ticked up 2.1% premarket on the announcement. Superintelligence Labs, formed last year, needs to show the money is buying something. An open model that runs on the hardware people already own is a strong answer to "what did the spending buy" — and a strong distraction from the question of what stays locked.

The generous reading and the skeptical one are both true. Meta has shipped a genuinely useful open model and made local, private agents dramatically more accessible. It has also demonstrated, in a single release, exactly how a company can be loud about distributing power while keeping the most valuable part of it in-house. The manifesto is about empowerment. The release notes are about which empowerment is free.

This is the summer the exodus reversed. Over the weekend, Spider-Man: Brand New Day and Christopher Nolan's The Odyssey crossed $1 billion at the global box office in near-lockstep — the web-slinger in twelve days, the second-fastest to the mark ever behind Avengers: Endgame, and the Homeric epic in four weeks, making it Nolan's highest-grossing film, past The Dark Knight Rises. They are the fourth and fifth billion-dollar releases of 2026, joining Toy Story 5, Michael, and The Super Mario Galaxy Movie. Five in a year — the most since 2019.

The numbers under the numbers are the interesting part. Brand New Day pulled $145 million domestically in its second weekend and has crossed $1.67 billion worldwide in ten days; no film has ever reached its domestic total faster. The Odyssey has minted $1.1 billion, and — the detail that should worry every streaming executive — it is now IMAX's highest-grossing domestic release of all time, with premium auditoriums near sellout in Los Angeles and New York through October. People are not just showing up. They are paying more to sit in a specific kind of room, then coming back to do it again.

Hollywood's year-to-date total sits at $6.6 billion, up 18.5% on last year. An R-rated three-hour art film about a Greek king and a superhero sequel are, together, out-grossing the combined take of Barbenheimer. The through-line this paper keeps returning to — that the world runs on physical fact, not announcement — has a gentle cultural corollary here. Streaming can announce infinite content. It cannot manufacture the one thing these films sold: an event that exists at a place and a time, that you have to leave your house to attend, and that is gone if you miss it.

Scarcity, it turns out, is the product. For a year defined by empty missile racks, closed straits and expensive money, the culture found its counter-move in the oldest format there is — a big screen, a fixed showtime, and 24 million people agreeing to be in the dark together. The couch had a great decade. This summer, the room won.

There is a trick of language in the AI industry that has gone almost fully unremarked, and Meta performed it beautifully this week. The trick is to attach the word "open" to whatever you have decided not to monetize directly, and then accept applause for generosity. Muse Glimmer is a fine model. It is also, by Meta's own description, the smaller sibling of Muse Spark — the one that stays closed. The essay is about empowering everyone. The release schedule is about keeping the best thing in the vault.

Let me be precise, because the accusation is precise. I am not saying open weights are worthless — they are enormously valuable, and a capable agent that runs on your own laptop and never phones home is a genuine advance in privacy and control. I am saying that when a company releases its second-best model under Apache 2.0 while keeping its best one proprietary, the correct headline is not "Meta distributes superintelligence." It is "Meta decides which intelligence is worth distributing, and it is not the good one."

The manifesto's rhetorical move is to make openness a moral property of the company rather than a property of a specific model. But you cannot download a philosophy. You can only download weights. And the weights that describe Meta's frontier — the thing it actually competes on — are exactly the ones you cannot have. "Spark 1.2 weights coming in the weeks ahead" is a sentence with a track record; Meta has promised and delayed before. A promise to open the frontier later is worth precisely what every "deal is close" on our front page this week has been worth: nothing until it clears.

Here is the tell that should end the debate. If open weights were genuinely the moral core of the enterprise, Meta would open the frontier first and let the small models follow. Instead it does the reverse, every time, and the reversal is not an accident — it is the strategy. Give away the models that build developer goodwill and Nvidia partnerships and a competitive moat against Chinese labs. Keep the model that would let a competitor actually replace you.

Call it open-washing, because that is what it is: the appearance of relinquishing power arranged so as to relinquish none of the power that matters. The reflexive gratitude the tech press extends to any weights-drop is the enabling condition. We should be able to say two things at once — thank you for the useful tool, and no, this is not the redistribution of intelligence you are billing it as. Meta has not opened superintelligence. It has opened the part of its lab it was comfortable letting you see, and written six thousand words to make you feel the difference was principle.

Every American account of the Strait of Hormuz treats Iran's six conditions as a tantrum — proof that Tehran is not serious, that the deal keeps almost closing only because Iran keeps unreasonably reopening it. This is a comforting story and an analytically lazy one. Sit in the Iranian chair for a moment and the six conditions stop looking like obstinance and start looking like a rational actor pricing its one remaining asset at what the asset is worth.

Consider what Iran has left. Its air defenses were degraded in the war. Its economy is under sanction. Its regional proxies are diminished. What it retains is geography: a chokepoint through which roughly a fifth of the world's seaborne oil must pass, and the demonstrated willingness to keep transponders dark and transit counts falling. When you have exactly one point of leverage, you do not spend it on a phased, reciprocal process where the other side controls the sequence. You attach it to everything you want and refuse to decouple, because decoupling is how you lose the only thing you hold.

That is precisely the move Zolghadr made. The June memorandum treated sanctions relief, reparations and troop withdrawal as parts of a gradual exchange. The August statement makes them preconditions for the first step. Washington calls this backsliding. It is not. It is Tehran recognizing that a phased deal gives America the ability to pocket the reopening and slow-walk the rest — and closing that exit. The strait is worth more as a held threat than as a spent one, and Iran will not spend it cheaply.

None of this makes Iran sympathetic. The regime is what it is; the demand to "never insult the sacred beliefs of the Iranian nation" as a condition for oil shipping is theater, and the human cost of a closed strait falls hardest on countries with no stake in the fight. But the mistake American policy keeps making is to confuse a strategy it dislikes with an absence of strategy. Iran is behaving like a weak power with one strong card, which is to say: it is refusing to play the card until the pot is worth it.

The implication for Washington is uncomfortable. There is no clever sequencing, no corridor half-measure, that gets the strait fully reopened without addressing the pile of things Iran has now bundled to it — because bundling them is the point. Either the U.S. pays something real on sanctions and assets, or it accepts a partial Omani corridor that Iran will explicitly refuse to call a reopening, and lives with the risk premium rebuilding into every barrel. The one option not on the table is the one the administration keeps announcing: a clean deal, close at hand, on American terms. Iran read the leverage. It is time Washington did too.