Tatsu’s Newsletter Podcast

Tatsu Ikeda

Obvious Truths Everyone Seems to Miss tatsuikeda.substack.com

  1. Sep 27

    Trump Couldn't Take Tehran, So He Took Caracas

    Bloomberg: $35/month. Financial Times: $42/month. The Economist: $17/month. Original geopolitical analysis by Tatsu with deep footnotes: $8/month. Share this preview with anyone pricing Gulf risk right now. Six weeks ago the president tried to charge the world a toll to cross the Strait of Hormuz. The toll collapsed in a day. The instinct behind it did not. Since then the strait has closed rather than reopened. Vessel traffic through Hormuz has fallen by roughly 95 percent, from about eighty-five ships a day before the war to around five, and Iran now says nothing crosses except "in coordination with Iran."[1] Gulf crude exports have been cut nearly in half.[2] The American air campaign was sold as the thing that would keep the oil moving. The oil has stopped moving, and Tehran holds the valve. When the instrument of force fails this completely, an honest government has two options: escalate to something that works, or stop. President Trump has spent six weeks demonstrating a third, and it is the most revealing pattern of the entire war. He is trying to bill the war, not win it. Then, on Friday, he made the pattern unmistakable. With the war against Iran frozen in stalemate, Trump announced that the United States had taken majority control of more than 65 billion barrels of Venezuelan oil reserves, a deal he called "the biggest oil deal in world history," secured "at no cost" to the American taxpayer.[3] A war that cannot be won in the Gulf is being monetized in the Caribbean. The bombing was supposed to reopen the strait. Instead Iran closed it, and the president went looking for something to sell. Below the paywall: * How the Venezuela seizure actually works: the 55 percent stake, the 100-year concession, and the regime change that made it "free" * The strait Iran now controls: 95 percent of traffic gone, fifteen burning tankers, and a Kharg terminal strangled rather than struck * Selling the shortage: why the president is marketing Texas crude to the buyers his own war stranded * The $9 billion accusation: what Iran says the president's envoys did with inside knowledge of the talks, and how Vance answered * Escalation with no endpoint: the "Economic D-Day" sanctions, the threat to bomb the mediator, and Iran's move toward US bases in Europe $8/month, 20 footnotes, no talking points. Follow how a war nobody is winning gets invoiced. Toll That Could Not Survive a Press Cycle Start where the pattern started, because it set the template for everything after. On July 13, Trump declared the United States "the Guardian of the Hormuz Strait" and announced it would be "reimbursed, at the rate of 20% on all cargo shipped" through it.[4] It was a demand that the world pay Washington a percentage of the value of roughly a fifth of all seaborne crude, in exchange for protection against a threat Washington was actively failing to suppress. Within about twenty-four hours the United Nations' International Maritime Organization stated flatly that there is no legal basis for charging vessels to transit a strait used for international navigation.[5] A regional outlet ran the entire critique in one word: piracy.[6] Trump withdrew the toll and replaced it with vague "trade and investment deals" the Gulf states would supposedly sign instead.[7] A protection racket that folds the next morning was never a policy. It was a trial balloon with a price tag. The retreat mattered more than the proposal, because it showed the administration reaching for monetization before it had any legal or coalition foundation to stand on. That is what governments do when they are improvising under pressure rather than executing a plan. Strait That Closed Instead of Opening The toll failed. The strait got worse anyway, which is the part the daily coverage keeps underplaying. Attacks on Gulf shipping did not stop; they became routine. The Abu Dhabi National Oil Company says fifteen of its vessels have been struck by missiles and drones since the conflict began, one crew member killed and around twenty injured, with the UAE blaming Iran's Revolutionary Guard for what it, too, called piracy.[8] Every hull that burns reprices the risk on every hull that has not. Brent crude spiked toward $105 a barrel when the summer ceasefire collapsed, then drifted into the high $80s and low $90s as Gulf producers cautiously resumed partial loadings, a range wide enough to tell you the market has no idea what a war with no ceiling is worth.[9] One correction to the record, because precision matters here. Trump has spent weeks threatening to obliterate Iran's Kharg Island terminal, which handles roughly 90 percent of Iranian crude exports. He has not done it. Kharg's oil infrastructure is offline not because it was bombed but because a US naval blockade and the collapse of insurance have idled it, with loadings halted for stretches at a time.[10] The terminal that anchors Iran's oil economy is being strangled, not struck, which is a slower and more deniable way to break the same thing. Selling the Shortage Here is where monetization stops being a metaphor. With the strait choked, the president found a sales pitch. People are finding "great alternatives," Trump said, "including Texas, including Alaska," and Louisiana, and "people are coming to the United States to get oil."[11] Against the map, the meaning is plain: the country whose military activity closed the world's most important oil corridor is now marketing its own crude as the workaround, to buyers it helped strand. The closure is a demand-side event for American barrels, not a problem the administration is racing to solve. The market data is more careful than the rhetoric, and I will be too: there is no confirmed record surge in US oil exports you can pin cleanly to Hormuz, only Trump's framing and the general price lift.[12] But the framing is the point. You do not hurry to reopen a passage while you are advertising yourself as the detour. A government that wanted the strait open would not narrate the crisis as a sales funnel. Biggest Oil Deal in World History Then came Venezuela, and the pattern graduated from rhetoric to seizure. On Friday, Trump announced that a private joint venture had been granted a 100-year concession over Venezuelan fields holding 65 billion barrels of oil, with the US government controlling 55 percent of the venture through equity and the right to take crude at cost. He called it the biggest oil deal in history and said it more than doubles American reserves, second only to Saudi Aramco among the world's owners of proven oil. It comes, he stressed, "at no cost" to the taxpayer.[3] "At no cost to the taxpayer" is doing a great deal of work in that sentence, given that the background to the deal is a US military operation that replaced Venezuela's government earlier this year. Set it beside Iran and the through-line is obvious. The administration could not take Tehran, so it took Caracas. The war it is losing in the Gulf pays for itself in the Caribbean, and the reserves it could not secure by bombing Iran it secured by regime change in Venezuela and a joint venture with an interim president installed after the fact. The acquisition was the objective. The war was the pretext. Who Is Trading the War The corruption thread is the part that turns a strategic failure into a scandal, and it has hardened since July, though not into anything a court has tested. During US-Iran talks earlier this summer, Iranian negotiators sent a private message to Vice President JD Vance warning that Special Envoy Steve Witkoff and the president's son-in-law Jared Kushner were exploiting inside knowledge of the negotiations to profit in financial markets. Tehran has since escalated, quantifying its accusation at roughly $9 billion in gains, about half of it attributed to alleged market manipulation, and demanding a Congressional investigation.[13] Vance has called the claim "completely bogus" and denied ever receiving the message.[14] The denial arrived faster than any evidence, which is its own kind of tell. Asked earlier about the conflict-of-interest concerns, Kushner offered a line worth preserving intact: "What people call conflicts of interest, Steve and I call experience and trusted relationships that we have throughout the world."[15] That is a rebrand, not a denial. I want to be exact about the evidentiary status, because it carries weight. Iran is an interested party, its documentation has not been independently adjudicated, no US regulator or prosecutor has opened a formal probe, and the sharpest figures rest on Iranian claims relayed through non-mainstream outlets. What is on the record is that the warning was delivered, that the pattern of Kushner and Witkoff blending private business with public diplomacy has been reported by mainstream outlets, and that the administration has defended the arrangement rather than dismantled it. That makes the question legitimate. It does not convict anyone, and this piece does not. Escalation Without an Endpoint Everything above sits inside a war that has no exit and no plan for one. The framework that was supposed to end it was never a ceasefire. The June understanding opened a 60-day window for talks, and that window expired in mid-August with no deal and Iran declaring the whole thing dead after what it called gross American violations.[16] Washington's answer was not diplomacy but a new economic front: on August 24, Treasury Secretary Scott Bessent unveiled "Operation Economic Outcast," a sweeping campaign of US secondary sanctions targeting Iran's gold, aviation, shipping, and digital-asset networks, with country-by-country ultimatums to cut ties or lose access to the dollar.[17] "Operation Economic Outcast" is what you name a policy when the shooting has not produced a headline. The military track, meanwhile, keeps widening rather than resolving. Trump has privately threatened to bomb Oman, the very country trying to broker a Hormuz arrangement, if it "g

    Trump Couldn't Take Tehran, So He Took Caracas
  2. Sep 22

    Day 207: Washington Is Feeding Its MQ-9 Reaper Fleet to the Strait of Hormuz

    Bloomberg: $35/month. Financial Times: $42/month. The Economist: $17/month. Original geopolitical analysis by Tatsu with deep footnotes: $8/month, 14-day free trial. This post is public. Share it with anyone still calling the Strait of Hormuz "contested." The number that decides this war is an inventory, not a casualty count or a headline. As of mid-August the United States had lost roughly a quarter of its entire MQ-9 Reaper fleet to the Iran war, at least 45 aircraft at up to $50 million each, according to the Washington Post.[1] That was five weeks ago. The bleeding has not stopped, and the place it is bleeding fastest is the sixty kilometers of water Washington has spent seven months failing to reopen. This morning Iranian air defenses claimed another American drone over the Strait, and OSINT reporting put the current loss rate at four to five Reapers a week.[2][3] Run that math forward and the arithmetic is brutal in its simplicity. At four to five a week, the remaining Reaper inventory is a six-month problem, and month one has already started. Start a 14-day free trial to read the paid structural analysis when it drops. $80/year if you stay, less than two Bloomberg sandwiches. Twenty-Five Percent of the Reaper Fleet Is Gone The MQ-9 was built to loiter over people who cannot shoot back. Iran can shoot back. Tehran's Revolutionary Guard says it has been downing Reapers over and east of Hormuz with a new domestic air-defense system, and while US Central Command declines to confirm the individual claims, the Post's fleet accounting does the confirming for them: the arsenal is measurably smaller than it was in February.[1][4] Consider the exchange rate. A Reaper runs into the tens of millions of dollars. The surface-to-air missile that kills it costs a rounding error by comparison. Every week the United States is trading multi-million-dollar airframes for the privilege of watching a waterway it does not control, and the footage of the wreckage goes out on Iranian channels within the hour.[5] General Atomics is about to post the best sales quarter in its history, and the customer is attrition. None of this buys a single mile of open water. The drones are surveillance and strike platforms, not minesweepers or escorts. They can watch the Strait close. They cannot pry it open, and now they cannot even watch it for very long. Iran Is Sinking Tankers Back The Strait has become a reciprocal shooting gallery, and the United States started this particular exchange. In early September, US forces struck Iranian oil tankers for the first time, retaliation for attacks on American warships operating in the Gulf.[6] Tehran answered in the only currency that moves oil markets: hulls. This morning Iranian forces struck a tanker attempting to transit the Strait, and separate maritime warnings flagged a gas carrier hit by debris from unidentified projectiles.[7][8] These are not one-off incidents anymore. The UK Maritime Trade Operations desk has become a running ticker of vessels hit by "unknown projectiles" in the world's most important oil chokepoint, and the US naval blockade layered on top keeps traffic thin and prices high.[7] Six weeks ago the argument was whether Iran could close the Strait. Now the argument is how many tankers a week it can hit while doing it. This is the same machine I traced through Venezuela, running in reverse. There, a war that produced no victory got monetized into an oil concession. Here, a war that produces no victory is monetizing the closure itself, in insurance premiums, in freight rates, in the price of every barrel that still dares the passage. Yemen Strike That Never Launches While the Reapers burn, the administration keeps rehearsing a second front it cannot bring itself to open. Over the weekend, US strikes on the Houthis in Yemen were close enough to real that commanders had approved the target lists and crews were loading bombs onto the aircraft, before Trump called the whole thing off at the last minute following a personal plea from the Saudi crown prince.[9] Axios confirmed the same reversal from the administration side.[10] This is the second telegraphed-then-aborted Yemen operation in a week. The prior version was a Camp David meeting billed as a review of strike options.[11] The options keep reviewing well and launching poorly. OSINT commentary on the pattern was blunter than anything a briefing room would allow: if you are going to do it, do it, and if you are not, stop announcing it.[12] The restraint is arithmetic of a different kind, not principle. Opening Yemen risks the Bab al-Mandab, the other chokepoint, where the Houthis already operate with something close to freedom of movement. Two closed straits instead of one is not a threat Riyadh wants to test, which is why the last voice in Trump's ear was Saudi, not American. Pressure Theater Everywhere the Guns Are Not Where the battlefield offers no wins, the administration has moved the war to places where it can still issue a press release. Treasury Secretary Scott Bessent went on CNBC to announce that every Iranian airline would be "shut down around the world" within two days, and the compliance came fast: Iraq agreed to suspend sanctioned Iranian carriers starting at dawn, Turkey grounded its Iran routes into 2027, and Georgia barred them outright.[13][14] An Iranian member of parliament responded by filing an emergency bill to withdraw from the Non-Proliferation Treaty, while the foreign minister routed through Doha on his way to lecture the UN General Assembly.[15] The airlines are grounded. The Strait is not. That's the whole war in one sentence. Washington can still close an airspace, freeze a bank, and lean on a neighbor to cancel a landing slot. What it cannot do, after seven months and a quarter of its drone fleet, is move a tanker safely through sixty kilometers of water, which was the entire point. A war you are winning does not get fought this way. You do not substitute sanctions on airlines for control of the sea lane when the sea lane is the objective. You reach for the airlines precisely because the sea lane is beyond you, and you need something, anything, to put in the day's briefing. The Reapers will keep going up, because the alternative is admitting the surveillance picture has gone dark. The tankers will keep getting hit, because deterrence in a chokepoint runs both ways and Iran learned the lesson faster. And somewhere on a flight line, ordnance crews will load another Yemen package that will not fly, billed, as always, to a separate appropriation. 14-day free trial. Cancel anytime. $80/year if you stay, or $8/month. The war is not being won, it is being invoiced, one Reaper and one tanker at a time. Follow the money with me. Notes [1] "U.S. military has lost roughly 25% of its Reaper drones as Iran war depletes arsenal." The Washington Post, August 13, 2026. At least 45 MQ-9 Reapers lost during the war with Iran, roughly a quarter of the fleet, at up to $50 million per aircraft, with heavy use and heavy losses around the Strait of Hormuz. [2] "Iran claims US MQ-1 drone shootdown over Hormuz amid MQ-9 Reaper losses." Army Recognition, September 2026. Documents the pattern of Iranian shootdown claims over the Strait and the accelerating US drone attrition, including the MQ-1 downing. [3] OSINT intelligence capture (47,768 views): analysis estimating the United States is losing four to five drones a week to Iranian air defenses in the Strait of Hormuz, which at the stated rate would exhaust the remaining Reaper inventory within roughly six months. Attrition rate is OSINT-sourced; the fleet baseline it builds on is the Washington Post accounting in footnote 1. [4] "IRGC shoots down hostile US MQ-9 drone over Hormuz." IRNA, August 2026. Iranian state confirmation that the Revolutionary Guard Aerospace Force downed a US MQ-9 over the Strait using a new air-defense system. US Central Command has not confirmed individual incidents. [5] OSINT intelligence capture (1,079 views): IRGC-released footage of an MQ-1 drone reported detected and shot down over the Strait of Hormuz the morning of September 22, 2026. [6] "U.S. strikes Iran oil tankers for first time in retaliation for Hormuz strikes." Axios, September 2, 2026. First direct US strikes on Iranian tankers, framed as retaliation for attacks on American vessels. See also "U.S. strikes Iranian oil tankers after Navy ships targeted." NPR, September 5, 2026. [7] "Vessel struck in Strait of Hormuz, UKMTO says, as prospects for U.S.-Iran diplomacy appear elusive." CNBC, September 13, 2026. A tanker struck by an unidentified projectile, fire aboard, crew safe, amid a US naval blockade keeping shipping thin and oil prices elevated. Iran said its forces struck a ship attempting to "illegally" transit. [8] OSINT intelligence capture (7,356 views): UKMTO warning of a late-reported attack on an LPG tanker struck by debris from unknown projectiles in the vicinity of the Strait; the vessel was not damaged and the crew was reported safe. [9] "Houthis push for control of Yemen highlands as Trump is said to have called off strikes." The Japan Times (New York Times wire), September 22, 2026. Planned US strikes on the Houthis were halted at a late stage, with target lists approved and crews loading bombs onto aircraft, after a fresh plea from the Saudi crown prince. [10] "Trump weighed strikes against the Houthis in Yemen but decided to hold off." Axios, September 21, 2026. Administration-side confirmation of the last-minute reversal following the bin Salman conversation. [11] OSINT intelligence capture (47,243 views): report citing CNN national security correspondent Alex Plitsas that a Camp David meeting was convened to deliberate and review strike options for Yemen. [12] OSINT intelligence capture (45,629 views): commentary on the repeated public signaling of Yemen strike options without execution, arguing that telegraphed operations forfeit both deterrence and surp

    Day 207: Washington Is Feeding Its MQ-9 Reaper Fleet to the Strait of Hormuz
  3. Sep 22

    Legal AI in 2026: How the Sausage Gets Made, and Whether to Rent or Build

    Welcome back. I took August off, the first real break since I started this, and the world was decent enough not to end while I was away. Did you miss me? If so, there is a concrete way to say it: the paid tier runs $8 a month, about the price of one coffee, and it is what keeps these pieces coming. One note before we start, half good news and half not. This newsletter just crossed 2,100 subscribers, a number I still do not quite believe when I look at it. The less cheerful half: revenue is actually down even as the list has grown, which is the diplomatic way of saying most of that new growth is reading for free. That is what the free tier is for, and I am glad you are here. But if the work has earned a place in your week, a paid subscription is the difference between me spending weeks on an investigation like this one and spending an afternoon. If you can help, please do. Bloomberg: $35/month. Financial Times: $42/month. The Economist: $17/month. Original analysis by Tatsu with 18 footnotes: $8/month. Share this with anyone about to sign a legal AI contract. Every major legal AI vendor has, at one point or another, sold its product on a version of the same promise: it will not make things up. LexisNexis advertised "100% hallucination-free linked legal citations." Thomson Reuters said it avoids hallucinations "by relying on trusted content within Westlaw." Casetext, the company behind CoCounsel, said its tool "does not make up facts, or hallucinate."[1] In 2024, a team at Stanford's RegLab decided to test that promise against reality. They wrote 202 legal research questions, ran them through the leading tools, and had legal experts score every answer. The study was preregistered and later survived peer review in the Journal of Empirical Legal Studies.[1] It is the closest thing this industry has to an independent audit, and its finding is not ambiguous. The leading grounded legal research tools hallucinate between 17 and 33 percent of the time.[1][2] Broken out by product, the spread is worse than the headline. LexisNexis's Lexis+ AI was the best performer tested, and it was still fully accurate only about 65 percent of the time, hallucinating on more than one answer in six. Thomson Reuters's Westlaw AI-Assisted Research was accurate roughly 42 percent of the time and hallucinated on nearly one answer in three. Thomson Reuters's Ask Practical Law AI was accurate on fewer than one in five questions, and it managed that partly by refusing to answer 62 percent of the time. A tool that declines to answer two questions out of three is technically not hallucinating, in the same sense that a witness who takes the Fifth is technically not lying.[3] The vendors said hallucination-free. The study printed the error rate directly beneath the marketing copy. This matters because the downside is sanctions. Lawyers have been fined and referred to bar discipline for filing briefs with citations that turned out to be invented,[4][5] and "the software told me it was real" is not a defense that has worked for anyone. When a firm buys a legal AI tool, it is buying a promise about citation integrity, and that promise is the thing the leading study specifically measured and found wanting. To be fair to the category, grounding does help enormously. An earlier Stanford-affiliated study, "Large Legal Fictions," ran more than 800,000 queries through general-purpose models and found hallucination rates of 58 percent for GPT-4, 69 percent for GPT-3.5, and a spectacular 88 percent for Llama 2.[6] Against that baseline, a specialized tool that is right two-thirds of the time is a real improvement. The problem is the gap between "meaningfully better than raw ChatGPT" and "hallucination-free," because only one of those two phrases appeared in the sales deck, and it was not the accurate one. That is the scandal you can measure. The structure underneath it, the part that determines who you actually pay and what you actually get, is the one worth the subscription. Below the paywall: * Why the Harvey-versus-Claude frame is rigged, and the two incumbents it quietly deletes * Who really owns the grounding behind every "trusted" answer, and why Harvey subleases its credibility from a direct competitor * The DPA-versus-ZDR trap: how firms get their data retained after being told it was safe * The "enterprise-grade" model with the weakest data passport in Europe * How to actually procure one without overpaying on the parts nobody benchmarks $8/month, 18 footnotes, no vendor deck. 14-day free trial, cancel anytime. Harvey Versus Everyone Is the Wrong Question Walk into most legal AI procurement conversations in 2026 and you will be handed a binary: buy a managed legal platform, meaning Harvey, or deploy a frontier model, meaning Claude, and build the legal layer yourself. It is a clean frame. It is also a false one, and the two options it quietly deletes are the two that should worry the incumbents most. On the managed side, Harvey is a leader, not the category. Its most direct competitor is CoCounsel, owned by Thomson Reuters through its $650 million Casetext acquisition[7] and grounded in Westlaw.[8] Sitting right next to it is Lexis+ AI, the first-party product from LexisNexis, built on the very research corpus that Harvey pays LexisNexis to license. Then there is Legora, a fast-growing entrant that is particularly strong in Europe,[9] plus a specialist layer, Robin AI and Spellbook and Luminance for contracts, Paxton AI and vLex Vincent for research. On the frontier side, "native model" has been treated as a synonym for Anthropic, which is strange, because the single most-deployed enterprise model is OpenAI, and it is one of the models running inside Harvey.[10] The honest native column is Claude, OpenAI, Google Gemini, and Microsoft Copilot, evaluated on where they differ rather than on the governance baseline they all now share. Any evaluation that names only Harvey and Claude has skipped the incumbents with the deepest moats and the vendor with the largest install base. That omission is the shape of the market's own marketing. Rent Duopoly Nobody Puts in the Deck Here is the fact that reorganizes everything else. Citation integrity is primarily a property of the research corpus the model is grounded in, not the model itself, and that corpus is a duopoly: Thomson Reuters's Westlaw on one side, LexisNexis on the other. KeyCite and Shepard's, the two citators that tell you whether a case is still good law,[11] are each locked to one of those two houses. CoCounsel grounds on Westlaw. Lexis+ AI grounds on LexisNexis. And Harvey, the independent, grounds through a commercial partnership with LexisNexis,[12] which happens to sell a competing product built on the same data. Harvey pays LexisNexis for the grounding that makes it trustworthy. LexisNexis also sells its own tool that does the same thing. Somewhere inside that sentence is a renewal negotiation. For a buyer, this has two consequences the vendor will not volunteer. A tool grounded on a single corpus inherits that corpus's coverage gaps and its citator's judgment calls about which authority still counts. And a tool that grounds through a partnership rather than owning the data is exposed to the commercial terms of that partnership in a way a first-party product is not. You are not just choosing an interface. You are choosing a landlord, and in Harvey's case, a landlord who subleases from a competitor. "Hallucination-Free" Is the Tell, Not the Feature Return to the marketing claim, because its persistence is diagnostic. No vendor has published a benchmark showing zero hallucinations. They have published the sentence. The Stanford team went looking for the evidence behind "hallucination-free" and found the phrase unsupported, which is why the study quoted it by name. The deeper hole is this: the two products most relevant to an actual Harvey purchase decision, CoCounsel and Harvey itself, have no independent hallucination benchmark at all. The Stanford study tested Lexis+ AI, Westlaw AI-Assisted Research, Ask Practical Law AI, and GPT-4. It did not test CoCounsel, and it did not test Harvey. So when a competing deck cites "the Stanford numbers" against CoCounsel, it is misattributing figures that belong to a different Thomson Reuters product.[3] The two most-sold platforms in the category are, empirically, unmeasured. A newer benchmark from Vals AI in October 2025 showed real improvement across the tools it tested, but the market leaders, Westlaw CoCounsel and LexisNexis and vLex, opted out of the legal research portion, so those flattering numbers describe smaller tools, not the ones a firm would shortlist.[13] The takeaway is that "hallucination-free" is a claim the leading study disproved, and any procurement process that accepts it at face value has skipped its one job. Contract You Signed Is Not the Contract You Think Now to the governance fine print, where a subtle conflation costs firms real exposure. Two contractual objects get treated as one, and they are not the same thing. A Data Processing Addendum limits how a vendor uses your data and secures how it is processed. It typically still permits the vendor to retain that data for a window, for abuse monitoring and debugging. Zero Data Retention is the different promise: that prompts and outputs are not persisted at all. ZDR is a separate, sales-approved, endpoint-by-endpoint amendment. It is not a default. OpenAI's own documentation says as much in plain language, and enterprise chat and API products default to controlled retention measured in weeks, not zero.[14] The contract everyone signed says the vendor will not train on your data. It does not say the vendor will not keep it. Those are different promises, and firms tend to learn the difference in that order. This is where Harvey has a genuine, defensible edge that survives scrutiny: it contractually requires both no-training and Zero Data Retention across all of its underlying

    Legal AI in 2026: How the Sausage Gets Made, and Whether to Rent or Build
  4. Aug 2

    Geopolitical Risk Assessment: May 2026

    CLASSIFICATION: Professional Analysis, Founding Member Tier PREPARED: 2026-05-28 | COVERAGE: May 1 to May 28, 2026 | WORD COUNT: ~11,500 REPORT TYPE: Monthly Strategic Assessment Phase 1: Regional Stability Rankings Region | Stability | Trend vs. April | Key Driver --------------+-----------+-----------------+----------------------------- Middle East | 1/10 | no change | Ceasefire collapsed May | | | 27-28; US strikes Bandar | | | Abbas; Iran strikes Kuwait | | | airbase; PGSA | | | institutionalized Russia/Ukr. | 3/10 | no change | Sarmat ICBM test; Victory | | | Day parade without tanks; | | | rival 2-day ceasefires; | | | 1,300+ drone nights Europe | 3/10 | no change | German recession confirmed; | | | Norway joins French nuclear | | | umbrella; 5K troop | | | withdrawal expanded North America | 3/10 | ▼ -1 | War Powers Resolution dodge | | | codified; 61% Americans call | | | war a mistake; Cuba carrier | | | threat sustained Asia-Pacific | 5/10 | no change | Trump-Xi summit | | | underwhelming; China emerges | | | as Iran's guarantor; | | | Pakistan opens 6 land | | | corridors Latin America | 4/10 | no change | Cuba aircraft carrier threat | | | (Abraham Lincoln); Venezuela | | | transition continues | | | failing; Mexico USMCA stress Africa | 4/10 | no change | Africa Corps Kidal | | | withdrawal aftermath; Sahel | | | security partnership | | | collapse; ECOWAS realignment Oceania/Pac. | 7/10 | no change | AUKUS holding; LNG premium | | | pricing sustained; no direct | | | conflict exposure Phase 2: Executive Summary Bottom Line The global risk posture remains at ELEVATED (Level 4 of 5) for a third consecutive month. May 2026 was the month the Iran war's architecture became institutional. The Trump administration formally argued to Congress that "hostilities have terminated" while operational tempo escalated to direct US-Iran naval combat in the Strait of Hormuz, Iranian ballistic missile strikes on a US airbase in Kuwait, and US strikes on Iranian missile launch sites at Bandar Abbas. The April 8 ceasefire, declared dead by Iranian leadership in late April, was finally treated as dead by every actor by May 28.[1] The defining strategic transition of May is institutional. Iran formally stood up the Persian Gulf Strait Authority (PGSA) with tolls payable in Iranian rial to Iranian-state banks and a stated email address (info@PGSA.ir).[2] The Trump administration sanctioned the PGSA within a week of its founding, validating the institution's reality by attacking it. Chinese vessels began transiting Hormuz under a face-saving framework Beijing and Washington both called "environmental fees." Pakistan opened six land border crossings with Iran for over 3,000 containers. China activated rail corridor cargo trains. Saudi Arabia proposed a non-aggression pact with Iran on May 14 (Financial Times). The United Arab Emirates exited OPEC+ to ramp output from Fujairah and Khor Fakkan ports outside Hormuz. The two-tier maritime governance architecture that emerged in April hardened into operational permanence in May. The Trump-Xi summit in Beijing on May 13 was the month's most overhyped event. Trump arrived with a $11.5 trillion CEO delegation (Musk, Cook, Huang). Xi did not meet him at the airport. The summit produced a Boeing order for 200 aircraft (down from Trump's floated 500, sending Boeing shares down 4%) and rhetorical commitments on Iran that East Asia Forum characterized as "largely rhetorical rather than substantive."[3] The substantive deliverable was operational normalization for Chinese vessels through Hormuz, framed as a US win while functioning as Iranian sovereignty recognition. The most consequential disclosure of May was the exposure of the hidden coalition. Saudi Arabia and Kuwait conducted unpublicized strikes on Iran during the active phase of the war (Reuters, WSJ, Times of Israel).[4] Netanyahu announced on May 13 that he made a secret visit to Abu Dhabi during the war to meet with MBZ, with Israel transferring Iron Dome batteries and personnel to the UAE.[5] The UAE denied the meeting took place. The IDF chief, Mossad chief, and Shin Bet chief all visited the UAE during the war per Israeli government statements. The war that did not officially exist had allies who did not officially exist using weapons systems that were not officially deployed. Key Strategic Signals 1. The War Powers Resolution interpretation became operational doctrine. Trump's May 1 notification to Speaker Mike Johnson and Senator Charles Grassley argued the 60-day clock does not apply because hostilities "terminated" on April 7. By May 28, the United States had launched strikes on Iranian launch sites at Bandar Abbas, Iran had struck Ali Al-Salem Airbase in Kuwait with ballistic missiles (15 American servicemembers wounded according to US military), and JCS Chair General Caine had publicly admitted Iran has attacked US forces "ten times since the declaration of the ceasefire."[6] The administration has not retracted the termination claim despite operational reality. The legal precedent now applies to every future executive. Confidence: 95%. Timeframe: Permanent. 2. The Persian Gulf Strait Authority became a sovereign institution. Iran's PGSA operates with tolls in Iranian rial, Iranian-bank settlement, banned vessels from Israel-linked operators, and a stated email contact. The US Treasury sanctioned the PGSA on May 8, treating it as a state institution to be sanctioned rather than a wartime measure to be dismissed. China publicly criticized "militarizing the Strait of Hormuz or a tolling system" while its tankers received operational access under "environmental fee" framing. Two-tier maritime governance is now the operational equilibrium with no path back to the pre-war regime. Confidence: 90%. Timeframe: Structural. 3. The Beijing summit confirmed China as Iran's guarantor. The summit produced Boeing announcements and Iran rhetoric while delivering Hormuz normalization for Chinese vessels. The joint China-Iran statement issued through Foreign Minister Wang Yi during Araghchi's earlier visit ("China will provide Iran with full support in defending its right to protect itself") became operational policy when Beijing accommodated Chinese tanker transits without breaking the toll regime. Putin landed in Beijing on May 20, five days after Trump departed. The triangular Beijing-Moscow-Tehran axis now operates as a coordinated structure rather than a parallel set of bilateral relationships.[7] Confidence: 85%. Timeframe: Multi-year. 4. The hidden coalition's exposure has reshaped Gulf diplomacy. Reuters confirmed on May 12 that Saudi Arabia and Kuwait conducted unpublicized strikes on Iran during the active phase. Times of Israel cited Western officials and Iranian sources confirming Saudi strikes "in retaliation for attacks carried out in the kingdom during the war." Kuwait launched attacks on Iranian-backed Iraqi militias from its own territory per WSJ. The Iron Dome transfer to UAE during the war was confirmed by Netanyahu's office and denied by the UAE. The credibility cost of having maintained official neutrality during the war is now being paid by every Gulf state simultaneously. Saudi Arabia's May 14 non-aggression pact proposal with Iran represents an attempt to reset before further exposure. Confidence: 90%. Timeframe: 90 to 180 days. 5. Hezbollah's fiber-optic drone capability inverted Israeli ground operations. Hezbollah released footage on May 7 and May 8 showing FPV strikes on Iron Dome battery launchers at Jal al-Alam.[8] The IDF acknowledged it lacks reliable counters to the fiber-optic drone threat. Israeli military rushed "thousands of meters of fishing nets" to troops in southern Lebanon as a stopgap. Master Sgt. Alexander Glovanyov, 47, was killed in a drone attack near the Lebanese border on May 21, the 18th Israeli soldier killed since the Lebanon front reopened. The technology cost ratio (drones at $4,000 versus Iron Dome interceptors at $50,000 to $100,000 versus Merkava IV tanks at $7 million) has structurally broken Israeli ground operations doctrine. Confidence: 95%. Timeframe: Immediate. 6. Iran's capability disclosures contradict American strategic claims. New York Times reporting confirmed Iran retains access to 90% of underground missile sites; only 10% were permanently disabled by 64 days of US and Israeli air strikes. Iran's Defense Ministry stated 80% of collapsed tunnel entrances have been restored. The Department of Defense Logistics Agency confirmed insufficient tungsten reserves for sustained kinetic operations and began sourcing from Brazil and Vietnam (NBC). US Tomahawk inventory is over one-third depleted; half of THAAD and Patriot interceptor stockpiles were expended in 64 days of active combat. China is conducting its own after-action review with these numbers. Confidence: 90%. T

  5. Jul 7

    A Federal Judge Just Read a Man's AI Chats Into Evidence

    In February 2026, a fraud defendant in the Southern District of New York did something that felt, at the time, entirely reasonable. He had received a grand jury subpoena. He knew he was a target. So he opened the consumer version of Claude and worked through his own legal exposure: what the government might have, what he could argue, where the facts cut against him. Judge Jed Rakoff ordered those documents produced to the prosecution.[1] The ruling was a question of first impression, and the reasoning was blunt. The AI is not a lawyer, so nothing the man typed into it was a privileged attorney-client communication. The provider's own terms of service defeated any reasonable expectation that the exchange was confidential. And because he acted on his own rather than at the direction of counsel, the work product doctrine did not protect it either. The tool he trusted to analyze his risk became the evidence against him. That case is not an outlier. It is the shape of things. Yesterday, a client asked me a simple question. Are my AI chats private? I spend most of my time at the intersection of technology, capital, and risk, and I expected to give a quick reassurance. Instead the honest answer turned out to be so alarming, and so poorly understood even by sophisticated people, that answering it properly became a practice. Today I am proud to introduce Chatham by THV (chatham.techhealthventures.com), the AI-governance and data-privacy practice of Tech Health Ventures LLC. Chatham helps private-capital principals understand exactly what their AI use exposes them to, on privacy and on legal discovery, and builds the governance to control it, working alongside their own lawyers. It is the productized version of that first alarming answer. This post is public. Share it with anyone whose AI chats contain things they would not want read aloud in a deposition. Rules Changed in 2026. Habits Did Not. Most people picture an AI chat as something between a private diary and a search query. It is neither. It is a written record, stored on someone else's servers, produced under someone else's policy, and increasingly reachable by courts and counsel. Here is what actually changed while everyone was busy being impressed by the technology. * AI chats are discoverable business records, and a court has said so. Heppner is the first federal ruling to say it plainly, but the logic is portable to any civil dispute, regulatory inquiry, or investigation: what you typed is a document, and documents get produced. * Providers can be compelled to hand over chats at scale. In the copyright litigation against OpenAI, a federal judge affirmed an order requiring the company to produce roughly twenty million de-identified ChatGPT conversation logs, overriding OpenAI's own objection that doing so would invade its users' privacy.[2] The privacy of the person typing is not the provider's priority when the provider is the one being sued. * Consumer AI keeps your chats, and can now hand them over. Anthropic's consumer tiers (Free, Pro, and Max) train on your conversations unless you turn that off, retain them for up to five years when training is on, and, under a policy effective July 8, 2026, permit disclosure to law enforcement on the company's own good-faith belief, without a court order.[3] Commercial and enterprise tiers are governed separately. This is not a knock on one company; it is the direction the whole consumer market is moving. * Connected AI is an exfiltration surface. The moment you link an AI tool to your email or your drive, through connectors or the emerging plumbing people call MCP, a single prompt-injection attack can quietly reach everything that connection can read. The convenience and the exposure are the same wire. * The folk remedies do not work. Incognito mode is still retained, commonly around thirty days. A VPN hides an IP address, not a logged-in identity. Deleting your chats and opening a fresh account does not purge the provider's copy. And once litigation is reasonably foreseeable, deletion stops being hygiene and becomes spoliation. It converts a privacy worry into a sanctions problem. The rules changed in 2025 and 2026. Most people's habits did not. That gap is the entire reason Chatham exists. Nobody Can Sell You Immunity From Discovery Let me be precise about what governance can and cannot do, because the distinction is the whole business. Nobody can sell you immunity from discovery, and anyone who claims to is selling you a sanctions motion. There is no product, no setting, and no clever deletion routine that makes a legitimate legal request go away. What exists instead is lawful, ordinary, and boring: minimizing what you needlessly create, keeping genuinely privileged material inside the channels where privilege actually attaches, and running defensible retention and legal-hold procedures so that when you are asked, you can answer cleanly rather than looking like you hid something. None of this is legal advice, and Chatham is not a law firm. It is governance and technical exposure work that makes your lawyer's job faster and your position defensible. The goal is not to beat discovery. The goal is to never be the easiest, most careless target in the room. That is the work, and it starts at chatham.techhealthventures.com. What Chatham Does The work runs as a ladder, and you can step off at any rung. Exposure Assessment. A fixed-fee diagnostic, roughly two weeks. We map every AI tool, account, and connector in use, trace where your sensitive data actually ends up, and deliver a written Exposure Report with a prioritized risk map. The report cleanly separates the governance findings, which are ours to fix, from the questions that belong with your lawyer. Governance Program. The build. An AI-use policy with real data tiers, a written information security program, a data-retention and legal-hold procedure, communications hygiene across every channel you actually use, and a counsel handoff pack so your attorney finalizes the legal layer from finished drafts instead of a blank page and a running clock. Ongoing advisory. AI providers rewrite their terms constantly, as the July 8 policy change illustrates. Chatham keeps your program current so last quarter's governance does not quietly expire. Everything is fixed fee, quoted before you sign. Never hourly. You will always know the number before the work starts. Why Chatham Is Different The framing nobody else leads with. Consultancies sell "AI adoption for wealthy families." Law firms publish client alerts. Both are useful, and neither starts where you actually are. Chatham leads with the inverted, evidence-backed message: you are already exposed, here is precisely where, and here is the governed way out. When I scanned the market for anyone else planting a flag on that exact ground, I found it unclaimed. Not a law firm, on purpose. Chatham does the governance and technical work and hands every legal question to your own counsel, packaged and pre-briefed. There is no conflict with your existing lawyer. Your lawyer receives finished drafts to redline, not a new bill to open a new matter. We practice what we sell. The Chatham site stores no lead data anywhere. An inquiry exists only as two emails. There are no cookies, no trackers beyond an anonymous aggregate visit counter, and no third-party form vendors sitting quietly in the middle of your first confidential message. Sensitive drafting happens on local models. The privacy pitch is not marketing. It is the operating model. Fixed fees and boutique scale. You get a named advisor who read your file, not a rotating bench that re-learns your situation on your dime. Who It Is For Emerging fund managers and syndicate leads. Single-family offices. RIAs and advisers. Angel investors. Founders holding sensitive cap tables. In short, anyone whose AI chats contain investor names, capital commitments, deal terms, valuations, or their own private legal worries. If you have ever pasted something into a chatbot that you would not want read aloud in a deposition, this is for you. What This Costs Everywhere Else I will not quote Chatham's fees here, because the number belongs in a private conversation about your specific scope. But the market gives you the shape of it. Comparable fixed-scope privacy and compliance assessments start around eight thousand dollars and run far higher; published GDPR gap assessments are a useful public benchmark. Fractional chief information security officer and compliance-officer retainers are commonly quoted between fifteen hundred and ten thousand dollars a month. Law-firm review of the same ground runs at partner hourly rates, and much of that time is spent producing the very first drafts that Chatham's process hands to your counsel already written. The honest comparison is this. One Chatham engagement typically costs less than the billable hours a lawyer would spend just getting oriented in this space, and the deliverables arrive lawyer-ready. Chatham sits well below what this work costs anywhere else, without being cheap, because cheap is not what you want standing between your chat history and a subpoena. How to Reach Me There is one path in: chatham.techhealthventures.com. The form asks four things: your name, your email, your role, and your single biggest concern. A real person, me, reads every request. If it is a fit, you receive a private booking link for a confidential call. You are covered by an NDA from first contact. There is no sales sequence, no drip campaign, and no junior associate assigned to warm you up. A federal judge has already demonstrated what happens when you treat a consumer AI like a private advisor. The delete button is not going to save the next person, and neither is a VPN. What works is knowing exactly what you have exposed and governing it before anyone asks. That is the entire job. If it is on your mind, start at chatham.techhealthventures.com. I write about technology, capi

    A Federal Judge Just Read a Man's AI Chats Into Evidence
  6. Jul 7

    Trump's Special Envoy Told Italian Television He Knew Epstein Had Underage Girls

    Bloomberg: $35/month. Financial Times: $42/month. The Economist: $17/month. Original analysis by Tatsu with 40+ footnotes: $8/month, 14-day free trial. This preview is public. Share it with anyone who still thinks the Epstein story is about one dead man. On April 19, 2026, the Italian state broadcaster RAI aired a segment of its investigative program Report titled "La guerra di Epstein," Epstein's War. In it, a man named Paolo Zampolli sat for an interview about his decades inside the international modeling business and his long friendship with Jeffrey Epstein. He was asked about the young women in Epstein's orbit. His answer, recorded and broadcast on national television, was this: "I knew he had the girls, but they weren't mine. They weren't even models, they were young girls, masseuses." The phrase "young girls" is doing a great deal of work in that sentence, and Zampolli appears to have understood exactly how much.[1] His lawyers sent RAI a formal cease-and-desist, a diffida, demanding the network not air the interview. RAI aired it anyway, on the grounds that what a sitting envoy of the United States says about a child-trafficking network is news.[2] That is the part worth holding onto. Paolo Zampolli is not a private citizen describing a regrettable old acquaintance. Since March 2025 he has been the Special Representative for Global Partnerships, appointed by President Trump, traveling the world cutting deals in the name of the United States government.[3] An Italian public-broadcasting unit asked him on camera what the American Department of Justice never managed to ask him at all. 14-day free trial. The documented record below: the FBI interview, the $195 million passport scheme, the ICE deportation, and the deal he inflated by twelve billion dollars. $80/year if you stay, less than two Bloomberg sandwiches. Modeling Agent, Then Friend of the President The basic biography is not in dispute. Zampolli, born in Milan in 1970, learned the modeling trade from John Casablancas of Elite and from Jean-Luc Brunel, the agent whose MC2 agency was funded by Epstein and who died in a Paris jail cell in 2022 awaiting trial on rape charges.[4] Zampolli founded his own New York agency, ID Model Management, and in September 1998 he introduced a Slovenian model named Melania Knauss to Donald Trump at a party during Fashion Week. He has been a fixture of Trump's circle ever since, which is itself notable. The president is not famous for keeping the same friends for twenty-eight years.[4] What Epstein thought of him is preserved in Epstein's own correspondence. In a 2011 email warning an Emirati business contact, Epstein wrote: "Be careful, zampoli is trouble. Lots." The line is widely cited in mainstream profiles of Zampolli, though the exact court exhibit index has not surfaced in the public databases.[5] When Jeffrey Epstein is the one warning you that a man is trouble, the assessment carries a certain authority. The relationship was not only social. Epstein patronized ID Models, and in 2004 the two men partnered on a joint bid to buy Elite Model Management outright. The bid failed, but the ambition is the point: Epstein and Zampolli wanted to own a modeling agency together.[4] What a modeling agency provided, in the architecture Brunel and Epstein built, was a lawful-looking pipeline of young women with visas, apartments, and a reason to be in the room. "Sleazy," in the FBI's Files The released Department of Justice production includes an FBI interview report that puts Zampolli's conduct on the record in law-enforcement language. The 302, drafted on August 26, 2019, during the Southern District of New York's Epstein investigation, documents a former ID Models signee.[6] She had gone to Paris at seventeen. At eighteen she signed with ID Models and was placed in a model apartment on Varick Street with two other girls. The report's characterization is blunt: Zampolli was "sleazy" and dated models. He made her get a short haircut, then let her go. She lasted about a year, during which, she said, she mostly sat around the apartment.[6] That is one interview, redacted, of a kind that recurs across the modeling-world testimony in the files. It describes the ordinary machinery: young women imported, housed together, kept available, discarded on a haircut. The same production contains a 2010 email in which an associate forwards Zampolli's United Nations gala material directly to Epstein, with the note "Paolo asked me to forward this to you."[7] The channel between the two men ran straight through the diplomatic world Zampolli was busy building for himself. Ghislaine Maxwell's Ocean Charity In 2013, Zampolli became one of only four core partners in the TerraMar Project, Ghislaine Maxwell's ocean-conservation nonprofit.[8] TerraMar presented itself as an advocate for the high seas. It was, by the assessment of international-development specialists, an influence-peddling vehicle that gave Maxwell a podium at global forums and a charitable reason to stand next to powerful people. The organization was dissolved in both the United States and the United Kingdom within days of Epstein's federal arrest in July 2019, which is the timing one expects of a genuine conservation charity.[8] Zampolli ran his own version of the same instrument. His "We Are The Oceans" and "Save Our Shark Coalition" events, co-hosted with his then-wife Amanda Ungaro, drew the same crowd. Maxwell attended a We Are The Oceans reception in New York on March 10, 2016.[9] RAI's reporting indicates that the oceans branding was load-bearing in a more literal sense: the operation is now under examination for its role in a financial-fraud matter involving the United Nations Office for Project Services, the allegation being that the moral glow of conservation was used to wave funding past the auditors.[10] A Sovereign Passport Business, Litigated in New York The diplomatic credentials were real, and Zampolli monetized them. In 2011 he was named Minister-Counsellor to the Commonwealth of Dominica's UN mission, and in October 2013 Dominican Prime Minister Roosevelt Skerrit elevated him to United Nations Ambassador and Ambassador for Oceans and Seas. The following June, Grenada appointed Ungaro its UN Ambassador for Youth Affairs. The pattern of converting social access into sovereign titles was, by this point, a method.[11] What the titles were worth is documented in a New York courtroom rather than in a blog. Zampolli sued the government of Dominica, Kempinski Hotels, and the Dubai-based Range Developments in the Appellate Division of the New York Supreme Court, seeking a finder's fee he said he was owed for aligning the parties behind the Cabrits Resort and Spa Kempinski Dominica.[12] His own filings describe how the resort was financed: through the sale of roughly 800 family passports under Dominica's Citizenship-by-Investment program, at about $225,000 each, for an estimated $195 million in state revenue. Dominica's opposition leader, Lennox Linton, confirmed publicly that Zampolli was chasing approximately $7.5 million, a five percent cut, and that the government had paid $30 million to the developers while leaving his commission unpaid.[12] Strip away the conservation galas and the UN letterhead and the business underneath is selling national identities to anyone with a quarter million dollars, then suing when the commission check is late. The man marketing the passports now markets the United States. ICE as a Custody Tool The clearest illustration of how Zampolli uses state power is what happened to Amanda Ungaro, the Brazilian former model who was his partner for roughly two decades and the mother of his teenage son. As their custody battle escalated, Zampolli contacted a senior Immigration and Customs Enforcement official, David Venturella, and reported her immigration status. Ungaro was detained and, in October 2025, deported to Brazil.[13] The next court date in the custody case would reportedly have moved their son toward her. Instead the federal deportation apparatus removed her from the country. The ethics watchdog Citizens for Responsibility and Ethics in Washington has called for the Department of Homeland Security Inspector General to investigate a Trump ally using ICE to detain the mother of his child.[14] A man with the President's phone number arranged for the government to deport the woman he was fighting in family court. Whatever else this is, it is a demonstration of capability. "Brazilian Women Are Programmed" The RAI broadcast did not only capture Zampolli on Epstein. Asked about Ungaro, he offered a theory of her nationality. The exchange, reported by the Brazilian press, went like this:[15] Zampolli: "Brazilian women cause trouble with everyone, right? It's not like this was the first one. Brazilian women are programmed." Interviewer: "To extort?" Zampolli: "No, to cause trouble." He also, according to Brazilian government statements and regional reporting, called Brazilian women a "cursed race," raça maldita, programmed to behave this way.[16] The remarks detonated a diplomatic incident. Brazil's Ministry of Women, under Minister Márcia Lopes, issued a formal condemnation calling the comments misogyny and hate speech. The country's First Lady, Rosângela "Janja" Lula da Silva, denounced him publicly and noted that the man insulting Brazilian women as a class had been accused by one of them of domestic, sexual, and psychological abuse.[17] A sitting United States envoy generated a head-of-state-level protest from the largest country in South America by going on Italian television and describing its women as a cursed, defective race. The Global Partnerships portfolio is off to a strong start. Twenty Billion Dollars in Twenty Minutes, Minus Twelve Billion Zampolli's actual diplomatic work follows the same pattern as his charities: the branding vastly outruns the substance. He has described himself as "Boeing's number-two salesperson in the

    Trump's Special Envoy Told Italian Television He Knew Epstein Had Underage Girls
  7. Jun 30

    How Trump's Family Picks Which Country to Loot Next

    Bloomberg: $35/month. Financial Times: $42/month. The Economist: $17/month. Original analysis by Tatsu with 30+ footnotes: $8/month, 14-day free trial. Share this preview with others. On December 15, 2025, Jared Kushner publicly withdrew from a project nobody outside the Western Balkans knew he had been negotiating. The site was the bombed-out former headquarters of the Yugoslav General Staff in central Belgrade, Serbia, leveled by NATO in 1999 and left as a politically charged ruin for twenty-six years. Affinity Partners, Kushner's private-equity vehicle, had spent most of 2024 quietly building a deal to convert the site into a luxury hotel and residential complex. In November 2025, the Serbian Parliament passed a special law to enable the privatization. In December 2025, the Serbian Prosecutor for Organized Crime charged four senior officials, including a sitting government minister, with abuse of office and document forgery to facilitate the development.[1] Kushner withdrew within the week. His public statement said that "significant projects should unite, not divide."[2] It was a diplomatic phrasing of "the prosecutor reached the file." Six months later, an almost identical Affinity Partners project is the subject of seven straight nights of mass protest in Tirana, Albania. The Sazan Island resort is, on paper, the Belgrade deal again. Bombed-out military site. Sovereign-level political facilitation. Special legislation. Gulf-state co-financing. Foreign-government regulatory leverage. Same private-equity vehicle, same family principals, same business template, two neighboring small states twelve months apart. The two projects diverged on one variable. In Serbia, an independent prosecutor's office reached the file and the project died. In Albania, the prosecutor's office reached the file last week, froze approximately $195 million in connected accounts, and the prime minister went on television to denounce the seizure as "arbitrary and negative" and to vow that the project will not stop "as long as I am here."[3] The qualifier was load-bearing. The Albanian Special Prosecution Office Against Corruption and Organized Crime (SPAK) is, by Western reform standards, the equivalent body to the Serbian prosecutor that killed the Belgrade deal. The procedural sequence so far is identical. What is different is the political math around SPAK. Edi Rama's Socialist Party won 83 of 140 seats in the May 11, 2025 parliamentary election.[4] That is a constitutional supermajority. The prime minister has the legislative votes to amend any law, reorganize any institution, and outwait any prosecutor for the remainder of a four-year term. SPAK can investigate. SPAK can freeze accounts. SPAK can file charges. What SPAK cannot do is win a political argument with Rama, because there is no opposition coalition large enough to make one. This piece is about the variable that determined the outcomes in Belgrade and Tirana, and about what that variable predicts for twenty-three other Trump-family foreign real-estate projects under development in 2026. The structural argument is in three parts. First, what Affinity Partners actually is, including the documented "direct financial partner" arrangement with the Albanian and Serbian governments that gave foreign sovereigns what the United States Senate Committee on Finance called "potentially coercive control" over the next president's family's investments.[5] Second, the four-variable scorecard for predicting which foreign-state Trump projects will go the Belgrade way and which will go the Albanian way. Third, the broader pattern this scorecard generalizes to, including the parallel case of Peter Thiel's data-and-mineral arrangement with Argentina's Javier Milei government, which is the same mechanism with a different operator. Start a 14-day free trial. Full structural analysis below. $80/year if you stay, less than two Bloomberg sandwiches. What Affinity Partners Is, on Paper and in Practice Affinity Partners was incorporated in Miami, Florida, in July 2021, six months after Jared Kushner left his White House role as senior adviser to the president of the United States.[6] It is structured as a private-equity firm wholly owned by Kushner. By mid-2026, Affinity manages approximately $6.16 billion in assets, according to congressional investigators.[7] Ninety-nine percent of that capital comes from foreign sovereign entities.[7] The anchor commitment is $2 billion from Saudi Arabia's Public Investment Fund, made in 2021, paying Affinity a guaranteed 1.25% annual management fee through August 2026. The fee runs regardless of investment performance, and Senate Finance Committee documents confirm it was guaranteed for the full five-year period "without exception."[8] Between 2021 and 2024, Affinity collected approximately $157 million in cumulative management fees from foreign clients, including roughly $87 million paid directly by the Saudi government and over $110 million total in Saudi-source fees through 2026.[8] Internal Saudi PIF documents obtained by Senator Wyden's investigators show the PIF Investment Committee unanimously rejected the commitment on grounds of "inexperience," "unsatisfactory" operational review, "excessive" fees, and "public relations risks." The PIF Board, chaired by Crown Prince Mohammed bin Salman, overruled the committee.[8] The Investment Committee was presumably not invited back. By any reasonable corporate-governance standard, that sequence does not happen. It happens when the investment thesis is not financial returns. It happens when the investment thesis is access. By the end of 2023, Affinity had deployed less than 18% of its capital. As of late 2024, the firm had generated, in Senate Finance language, "no return on investment" and had "not distributed a penny of earnings back to clients."[8] Affinity charges a management fee on capital it has largely not invested. The fee is the product. What Affinity buys with that fee structure is sovereign-government partnership terms that no purely private investor could secure. In a March 19, 2026 letter to Affinity, Senate Finance Committee Ranking Member Ron Wyden and House Oversight Committee Ranking Member Robert Garcia wrote that Affinity had confirmed in writing that the Albanian and Serbian governments would be "direct financial partners" in Kushner's real-estate developments, with "full responsibility for obtaining all necessary approvals, permits and licenses."[9] Wyden's letter characterized this arrangement as giving the foreign governments "potentially coercive control" over the personal investments of the family of the sitting president of the United States.[9] That sentence is the heart of the legal-and-political problem. In every prior version of this scandal, the foreign state was a customer. Here, the foreign state is a counterparty. The "direct financial partner" language is what distinguishes this entire structure from ordinary real-estate development. The Albanian government is not zoning Kushner's land. The Albanian government has committed, on paper, to a financial stake in the development of Kushner's land. The same is true of Serbia. Affinity's exposure to those two governments is not regulatory; it is contractual. When the Belgrade prosecutor charged the four government officials with abuse of office, what they were prosecuting was not generic corruption. They were prosecuting the precise mechanism by which the foreign government had agreed to be Affinity's partner. The Belgrade case collapsed because the prosecution attacked the structure itself. How Sazan Was Sold The Albanian end of the same structure runs through one law and one administrative action. The law is Law 21/2024, adopted by the Albanian Parliament on February 22, 2024.[10] It was introduced not as a government bill but as a "members' bill," sponsored by Socialist Party deputy Fadil Nasufi and eleven other majority-party MPs.[11] The members'-bill route bypasses standard public consultation, mandatory environmental impact review, and strategic review by independent expert panels. The sponsors argued publicly that prior conservation law from 2017 created developmental "deadlocks" by placing 21.5% of Albania's territory off-limits to intensive economic activity.[11] Article 14 of Law 21/2024 explicitly legalizes five-star and higher luxury tourism developments inside core protected areas, with the operative phrase being "regardless of whether the protected area's own founding decree allows it."[10] Articles 6 and 8 require municipal administrations to assume management of at least 20% of the surface of protected areas within their borders, decentralizing reclassification authority to municipalities.[10] Additional provisions empower the Council of Ministers to redraw the boundaries of national parks by executive decree.[10] A coalition of 56 Albanian environmental and civil-society organizations submitted a formal letter to the Assembly demanding the bill's withdrawal.[12] Mirjan Topi, executive director of Bird Guide, testified before the parliamentary Committee on Productive Activities, Trade, and the Environment that the law was "in open opposition to the guidelines of the international community for the protection of nature."[12] Taulant Bino of the Albanian Ornithological Society told the committee that the draft law was designed to "treat protected areas simply and only as economic areas," equating "irreplaceable national ecosystems" to "non-natural municipal flower gardens."[12] The Socialist majority blocked all amendments and passed the bill on a party-line vote. The administrative action followed ten months later. On December 30, 2024, the Albanian Strategic Investment Committee, chaired by Prime Minister Edi Rama, granted "Strategic Investor with Special Procedure" status to Atlantic Incubation Partners LLC, an affiliate of Jared Kushner's Affinity Partners.[13] The designation provides fast-tracked permits, acce

    How Trump's Family Picks Which Country to Loot Next
  8. Jun 23

    Day 116: Trump Divorced Netanyahu and Iran Walked. I Wrote This in February.

    Bloomberg: $35/month. Financial Times: $42/month. The Economist: $17/month. Original analysis by Tatsu, footnoted and on the record: $8/month, 14-day free trial. This post is public. Send it to anyone who spent 116 days telling you the Iranian regime was about to fall. On February 28, the day Israeli and American strikes killed Ali Khamenei in the opening hours of the war, I published a piece with a headline that got me called every name in the book. It was three words long after the colon: "They Killed Khamenei. It Won't Matter."[1] It did not matter. One hundred and sixteen days later, the war is over. The Iranian government is intact. It kept its enrichment program, its missile arsenal, and its grip on power. The United States lifted oil sanctions and signed the end of the war at a dinner in the Palace of Versailles.[2] And the country that started this to topple Tehran, Israel, ended it cut out of the room, fighting alone, and politically on fire. I am not going to pretend this is humility. I called the shape of this war on the day it started, and I called it again every week for fifteen weeks while the experts predicted a collapse that was never coming. This is the scorecard. It includes the one thing I am still refusing to sell you, even now, because that refusal is the whole reason the rest of it was right. Start a 14-day free trial. The structural analysis that called this war while the New York Times was still quoting the Foundation for Defense of Democracies. $80/year if you stay. Receipt One: Decapitation Was Never Going to Work The entire Western theory of this war was that you could kill the head and the body would die. On the day the strike landed, the Foundation for Defense of Democracies published "Regime change in Iran is underway, and it won't be easy." I published the opposite, in plain language: "Regime change. False. The regime was not overthrown. It was replaced by a military council that is more authoritarian, more militarized, and more capable of domestic repression than the clerical system it replaced."[1] That is exactly what happened. The succession to Mojtaba Khamenei consolidated during the war rather than fracturing under it, which I documented on Day 97, nine days before the deal that proved it.[3] The Foundation for Defense of Democracies predicted Iran's collapse for eighty-nine consecutive days. Iran spent the eighty-ninth day firing ballistic missiles at a US airbase. Decapitation theory did not survive contact with an actual state. Receipt Two: The War Would End Like Korea On Day 55, April 23, while the commentary was still split between "Iran is finished" and "World War Three is here," I wrote that both were wrong and the war would freeze into a Korea-style stalemate: "A frozen conflict in the pattern of Korea, with active economic warfare as the defining texture."[4] I gave it the highest probability on the board. That is the war we got. One hundred and sixteen days of attrition, no decisive victory for anyone, ending not in surrender but in a memorandum that resolves nothing structural and freezes the lines roughly where they sat. The ceasefire that finally arrived is Panmunjom, not Appomattox: a signature on top of a conflict that never actually stopped, which is why Israel was still bombing Lebanon the night the deal was announced. Receipt Three: Hormuz Becomes Iran's Toll Road On March 27, I wrote that Iran had not closed the Strait of Hormuz in the way everyone described. It had done something smarter: "It has converted the Strait from an international common into a regulated toll road, charging up to $2 million per voyage for safe passage."[5] I wrote that before Iran formally stood up the authority to do exactly that. The signed memorandum reopens Hormuz "without tolls," which sounds like Iran giving the toll road back. Then watch what Tehran actually did this weekend. The moment Israel kept bombing Lebanon, Iran re-closed the Strait, its navy ordering vessels to stay clear or be targeted.[6] The text says no tolls. The behavior says the chokepoint is still Iran's switch, and Tehran will flip it whenever the deal is violated. I told you eleven weeks ago that the Strait was the leverage. It still is. Receipt Four: Iran Negotiates From Victory, Trump From Weakness On Day 43, April 11, the headline was "Iran Is Negotiating From Victory. Trump Is Negotiating From Truth Social," and the line under it was blunt: "Iran's preconditions are being met one by one. Iran has not conceded anything. This is what negotiating from victory looks like."[7] On Day 87 I wrote that the uranium question had already been resolved by operational fact and Iran would keep its stockpile.[8] The signed deal: oil sanctions waived, frozen funds on the table, the missile program and the proxies struck from the agenda, the enriched uranium still sitting in Iran pending a future talk. The side that was supposedly losing does not get those terms. I said so in April. The paper arrived in June. Receipt Five: Netanyahu Would Sabotage the Peace, and Trump Would Cut Him Loose This is the one that matters most for understanding today's headlines. On Day 48, April 15, I published "Netanyahu Killed the Ceasefire. Here's How," with a one-line thesis: "Three parallel negotiations. One man torpedoed all of them."[9] I argued that Netanyahu's entire method was to separate the fronts, kill the Lebanon condition, and open his own track to exclude Iran. Then on Day 106 I named the endgame: the Trump-Bibi exit asymmetry, the senior partner wanting out and the junior partner wanting forever-war, resolving with the senior partner getting the exit, calling the junior partner to inform rather than consult, and the junior partner folding.[10] Look at what just happened. Trump signed the deal over Israel's head. Vice President Vance publicly lashed out at Israeli officials who urged Netanyahu to ignore it. Netanyahu announced Israel was "not a party" to anything, and his ministers said Israel was "not bound" by the American agreement.[6] Trump ended his own ally's war, against his own ally's wishes, and then dared him to do something about it. The divorce I described in April was finalized this month at a French dinner table. What I Am Still Not Going to Sell You Here is where the victory lap stops, on purpose, because the discipline is the point. Every outlet now is running the same headline: Iran won everything. It is the Iranian state media line, and it is wrong, and I am not going to repeat it just because the broad strokes flatter my own record. Iran did not get everything it wanted. It got a real win that is already coming apart. Iran's central demand was a ceasefire in Lebanon. Israel violated it within hours, which is why Iran had to re-close the Strait to enforce a deal it supposedly won.[6] Trump is dangling the money, threatening that Iran gets no funds during the sixty-day window and to "hit Iran very hard again."[6] And inside Iran, the regime is fighting itself over whether the negotiators sold out: a member of parliament went on state television accusing the team of ignoring the Supreme Leader's orders and signing anyway, after which the broadcast was cut, the network sued him, and the channel's director resigned.[11] A government that won everything does not put its own negotiators on trial. The honest scorecard is not "Iran won." It is the thing I have written since Day One: Iran survived, which under the circumstances was the only victory that mattered, and the deal that ratifies it is as fragile as every ceasefire that came before it. I called the regime's survival, the Korea outcome, the Hormuz leverage, the favorable terms, and the Trump-Bibi split. I am calling the next part too: this deal will be tested, probably in Lebanon, probably soon. The Strait already closed once this weekend. It will close again. Israel Did Not Lose the War. It Ruined Itself. The deepest call, the one underneath all the others, was about Israel. A state that went to war to prove it could reshape the region by force, and instead proved the opposite. It killed the enemy's leader on day one and changed nothing. It expanded its physical footprint into Lebanon, Gaza, and Syria by roughly a thousand square kilometers, the largest land grab since 1967, and made itself a permanent occupier of three hostile borders in the process.[6] It got cut out of the deal that ended the war it started. Its politics convulsed, with former prime ministers calling it the greatest strategic failure in the country's history. That is not defeat in the old sense. Israel was not conquered. It defeated itself, by mistaking military dominance for strategic victory, and discovering far too late that you can win every airstrike and still lose the war. The ministers shouting that "all of Lebanon should burn" are not describing strength. They are describing a country that has run out of moves and is reaching for the only one it has left, which is more violence that buys nothing. Why the Scorecard Matters I am not doing this to take a bow, or not only. I am doing it because the gap between what I wrote and what the major papers wrote for fifteen straight weeks is the entire argument for reading this newsletter instead of them. The New York Times had nineteen reporters on this war. I had a spreadsheet, a set of footnotes, and a willingness to say the unpopular structural thing on the day it mattered rather than the comfortable consensus thing a month after it was obvious.[12] The war is over. The next phase, what a fractured American hegemony looks like when it can sign deals it can no longer enforce, is the subject of the next piece, and that one is for subscribers. The receipts above are free. The forecast is not. 14-day free trial. Cancel anytime. $80/year if you stay, or $8/month. I called this war on Day One and every week since. The next call is behind the paywall. Notes [1] Tatsu Ikeda, "They Killed Khamenei. It Won't Matter," February 28, 2026. P

    Day 116: Trump Divorced Netanyahu and Iran Walked. I Wrote This in February.

About

Obvious Truths Everyone Seems to Miss tatsuikeda.substack.com