Category Archives: Security

Sailing, Security

Trump’s New Minesweepers for Hormuz Go MIA, Spotted in Malaysia

Leave a comment

Minesweeping Elvis has left the building.

Two-thirds of America’s Persian Gulf “mine countermeasure capability” just got busted by a shipspotter in Penang, Malaysia. Not declared by the Pentagon. Not briefed in Congress. A guy with a camera at a container terminal said, uh, what the hell.

Source: Twitter

USS Tulsa and USS Santa Barbara, two of three Independence-class Littoral Combat Ships known as the entire US Navy mine warfare force in the Middle East, were photographed docked at North Butterworth Container Terminal on March 15.

Their homeport is Bahrain. Bahrain is 3,500 nautical miles away. You know, where the US has been at war with Iran since February 28.

Iran’s mine doctrine is the reason these ships exist in their current “countermine” configuration.

The Gap

Again, as I’ve recently blogged, Trump had his four best minesweepers in January loaded onto a heavy lift ship and sent to Philadelphia for demolition. The Royal Navy decommissioned its last traditional minesweeper in Bahrain this year too.

And the official replacements are the Tulsa, Santa Barbara, and Canberra. These three LCS ships were just fitted with modular MCM mission packages at huge expense. Three large metal ships replacing four small non-metal ones, introducing an untested system, unfit for Gulf waters, against an adversary that has thousands of mines and decades of doctrine for deploying them in exactly the strait where traffic has now slowed to only what the Iranians decide.

Two of those three ships now sitting on the wrong side of the Indian Ocean is curious at least. Canberra’s location is “unknown”.

Finger Pointing

The War Zone reached out to CENTCOM, which directed them to Fifth Fleet. Fifth Fleet directed them back to CENTCOM. The Royal Malaysian Navy, which normally announces foreign warship arrivals on social media, said nothing. USNI News sources described it as a “logistics stop” and declined further comment.

Planet Labs satellite imagery shows no US warships in port in Manama since February 23, five days before strikes began. Clearing the port was prudent. Sending the mine warfare assets 3,500 miles east was… something else.

The Ol’ Switcheroo

Defense officials told reporters on Friday that missiles, definitely not the Iranian sea mines, are currently the largest threat to merchant shipping in the Strait of Hormuz. Read that slowly. Consider how many press conferences we have heard that Iranian missile capability has been “obliterated”. It’s not a reassurance here. The Navy sounds like it’s reframing the threat environment to match a capability it no longer has in theater. Or maybe the Navy realized the three ships have a critical vulnerability, beyond sucking a minesweeping.

At least twenty crude oil tankers and cargo ships have been struck by projectiles since February 28. Trump is failing to convince his former enemies, the NATO allies, to contribute ships for a convoy operation to reopen Hormuz. US officials say American escorts are unprepared, won’t be ready for weeks. So what’s with these new mine clearance ships sneaking off to Malaysia?

Are they being used to chase ships instead? Did someone decide a 40-knot trimaran with anti-ship missiles and boarding capability is more useful chasing Iranian sanctions traffic through the Malacca Strait than hunting mines it probably can’t find anyway?

LCS Can’t Find Itself

The LCS MCM mission package has been a decade late and plagued with reliability problems since inception. The concept envisions laboriously hunting individual mines one by one with unmanned systems, giving an estimated clearance rate of roughly two mines per hour against minefields that are expected to number in the thousands. Tulsa’s unmanned surface vessel, riddled with points of failure, infamously went rogue after its tow bracket quit.

Something Smells Rotten

The US appears to have pre-conceded capabilities it spent decades building. The Avengers were retired by Trump as a bureaucratic fait accompli a month before they were needed. The snowflake LCS MCM system was declared operational abruptly as an institutional checkbox. And when the contingency these systems were specifically designed for actually arrived, given headlines announcing Iran mining the Strait of Hormuz as everyone expected, the LCS fleet pushed off to somewhere else and nobody in the chain of command wants to say why.

The ships arrived in Penang on March 14 and departed March 16. A Russian task group had just left the same berth two weeks earlier. That’s probably just a coincidence. The rest of it isn’t.

Energy, History, Security

Trump Oil Bankrupt in Six Months: Promised Boom is Already Bust

Leave a comment

Trump bankruptcy is on the horizon again, this time on the ocean. Trump Steaks, Trump Vodka, Trump University, Trump Airline and now… Trump Oil.

Court filings tell the story. Trump applying the American military, like the mob flexes protection racket muscle, to monopolize a market isn’t what he thought it would be. Trump is spending tens of millions of taxpayer money grabbing and maintaining aging ocean tanker rust buckets that he can’t sell, holding oil he can’t offload, and continuing the program anyway into an expanding disaster.

Here’s a table for every tanker seized so far under his ill-considered “Operation Southern Spear”.

The Trump Junk Fleet

Tanker Seized Cargo (barrels) Est. Cargo Value Vessel Value Known Cost to U.S. Status
Skipper Dec 10, 2025 1.8M $120–$135M ~$10M $47M + $450K/mo + $5M pending Held; DOJ asking court to sell
Centuries Dec 20, 2025 ~2M ~$130M Unknown Unknown (moored at Galveston) Held
Bella 1 / Marinera Jan 7, 2026 Empty $0 Unknown Atlantic chase + ongoing Held; pure cost center
Sophia Jan 7, 2026 ~2M ~$130M Unknown Seizure costs; cargo returned Returned to Venezuela
Olina Jan 9, 2026 Loaded Unknown Unknown Seizure costs; cargo returned Returned to Venezuela
Veronica Jan 15, 2026 Empty $0 Unknown Unknown (moored off Puerto Rico) Held; pure cost center
Sagitta Jan 21, 2026 Unknown Unknown Unknown Unknown Held
Aquila II Feb 9, 2026 ~700K ~$45M Unknown 15,000 km pursuit + ongoing Held; not formally seized
2 additional (unidentified) Unknown Unknown Unknown Unknown Unknown Held per NYT

That’s just eight confirmed seizures already painting the obvious picture.

The NYT reports ten total with Venezuelan ties. Two (Bella 1 and Veronica) were empty when seized. Two more (Sophia and Olina) were returned to Venezuela. The U.S. absorbed the full operational cost of every seizure and got nothing back on four of them.

The Asset Trap

A tanker is not a seized bank account. It’s not a pile of gold. It is like a slumlord grabbing a condemned property, a decaying organism that consumes capital every second it sits unrepaired. Taking the decrepit hulls means the U.S. government has made itself into the world’s most expensive and insolvent shipping company.

For what?

The Skipper’s court filings are Trump Steaks all over again.

The U.S. government was forced to spend $47 million in three months on repairing and maintaining a vessel worth $10 million. Instead of all the things $47 million could have done domestically, it’s tangled up in acquired foreign debt.

Read that Trump businessman genius move again.

He’s blowing 4.7x a ship’s value just to keep it afloat in Texas. Oil storage runs $15,000 a day. Another $5 million is pending for insurance and crew. The DOJ’s own asset manager wrote that these costs “far outstrip standard assets.”

Grade school children understand the math showing this is bad, but not Trump. Previous American procedure was to seize the assets (oil) at sea with a siphon and let the liability (ships) sail on. Makes sense, right? The Trump model has been to take all the liability, immediately undermining the assets.

The Storage Bottleneck

Trump’s army of sycophants can’t simply sell the oil, deteriorating on old ships. These are civil forfeiture cases tied up in U.S. District Court in Washington. The Skipper’s cargo — worth $120 to $135 million — has been sitting unsold since December. At $450,000 a month in storage alone, a 12-month legal process would burn up $5.4 million before a buyer is found. Add the $47 million in catch-up maintenance and $5 million in pending costs, and nearly half the cargo value evaporates before a single barrel is sold.

The DOJ is now asking the court to allow an emergency sale of the Skipper’s oil before the massive losses become obvious to the public. That’s the Trump circus creating emergencies by admitting their strategy is hemorrhaging money faster than they can bully people into covering it up.

Net Recovery Projection

Only the Skipper has detailed cost data. But the Skipper is the template. These are all aging, end-of-life shadow fleet tankers that were past commercial retirement when they were seized. If the Skipper’s costs are even roughly representative, here’s what the full fleet of eight held tankers looks like over time.

Assumptions: maximum recoverable cargo across the fleet estimated at $500 million. Initial repair costs averaged at $20M per tanker (conservative — the Skipper hit $47M). Ongoing monthly costs per tanker estimated at $2–3.5M (maintenance, crew, insurance, storage). Neither scenario includes military operational costs, legal fees, or cargo depreciation.

Scenario Initial Repair (fleet) Monthly Burn (fleet) Total Cost at 6 Mo. Total Cost at 12 Mo. Max Recoverable Cargo Net at 12 Mo.
Conservative ($20M avg repair, $2M/mo per tanker) $160M $16M/mo $256M $352M ~$500M +$148M
Skipper Rate ($40M avg repair, $3.5M/mo per tanker) $320M $28M/mo $488M $656M ~$500M –$156M

Under the conservative scenario — which assumes each tanker costs less than half what the Skipper actually cost — the operation barely breaks even at 12 months. Under the Skipper rate, the operation goes underwater at roughly month 6 and never recovers. By month 12, the U.S. has spent $156 million more than the oil is worth.

Month six!

The Risk Nobody’s Pricing: Environmental Liability

Everything above is the optimistic scenario. It assumes nothing goes wrong with the ships themselves. That assumption deserves scrutiny.

These are single-hull, end-of-life “ghost fleet” tankers. They were built over two decades ago. They have been running intentionally dark, spoofing locations, skipping important inspections, and operating without valid safety certifications for years. So Trump has targeted absolute worst junk assets, with the least chance of positive return, for seizure.

Several were already rusting through, for obvious reasons. The Skipper’s $47 million in immediate repairs were totally avoidable by not seizing it.

I suspect the people who never maintain anything and have no concept of safety are the ones assuming all ships are equally valued.

Seizing unfit vessels on the verge of disaster actually makes the U.S. government the “responsible party” under the Oil Pollution Act of 1990. OPA 90 imposes strict liability on the owner or operator of any vessel from which oil is discharged into U.S. waters.

Bush signed OPA 90 in response to the Exxon Valdez disaster. But as the Netflix documentary The White House Effect now documents using his own presidential library memos, his chief of staff John Sununu was simultaneously running a back channel with Exxon to neutralize every environmental commitment the administration made.

Perhaps that’s the Trump plan too.

The filmmakers found never-before-seen correspondence between oil executives and the White House chief of staff — memos in which, according to director Jon Shenk, Sununu openly bullied the President. EPA chief Bill Reilly told the filmmakers that even he was shocked by the tone.

The oil industry’s reaction to the Valdez spill was not remorse. It was to circle the wagons — applying the tobacco industry playbook of deny, counter, and split the electorate. Exxon wrote directly to Sununu as their line into the government. He convened a confidential “Global Warming Scientific ‘Skeptics’ Meeting” stacked with climate contrarians funded by coal companies. And the Bush White House forced NASA scientist James Hansen to alter his own congressional testimony to downplay climate risks.

The law survived the Bush corruption that is responsible for growing climate change disasters we experience today. The intentions behind it didn’t, perhaps by design. And now that same OPA 90 framework — strict liability, uncapped when safety regulations are violated — is the one that’s governing Trump’s seized tanker fleet. What are the chances it holds?

“Strict” means no-fault, so if the oil spills, the responsible party pays. And the current OPA liability cap for a single-hull tank vessel over 3,000 gross tons is the greater of $4,000 per gross ton or $29.6 million. But the cap vanishes entirely if the spill resulted from “violation of an applicable Federal safety, construction, or operating regulation.” These ships have no valid classification, no current safety certificates, and no double hulls. The cap would not survive a normal courtroom.

Here is what the uncapped liability looks like.

Spill Scenario Volume Historical Comparable Cleanup Cost Range Total Liability (incl. damages)
Minor hull breach (1 tanker, partial cargo) ~500K barrels Larger than Exxon Valdez (262K bbl) $2–4 billion $3–7 billion
Major structural failure (1 full tanker) ~1.8M barrels Approaching Deepwater Horizon scale $5–15 billion $10–25 billion
Cascading failure (2+ tankers at anchorage) 3–4M barrels No historical precedent $15–40 billion $25–65 billion

The numbers have precedent. Exxon spent roughly $2.5 billion on cleanup alone for 262,000 barrels — about $9,500 per barrel spilled. BP’s total Deepwater Horizon liability exceeded $20.8 billion in settlements, with total costs above $65 billion. The Skipper is sitting in the Galveston Offshore Lightering Area with 1.8 million barrels of heavy Venezuelan crude — nearly seven times the volume of the Exxon Valdez spill — in a hull that required extensive repairs so it wouldn’t wreck Texas.

And the government plans to add even more debt from captured Iranian tankers to this fleet. Iranian shadow fleet vessels are notoriously among the worst-maintained ships afloat. By seizing them, the U.S. takes the environmental and safety liability another step deeper. One major hull breach in a U.S. port turns a hundred-million-dollar waste into a multi-billion-dollar ecological disaster. Talk about sunk cost.

The Ledger

Trump is pushing deranged reports of gross cargo value, to generate $130 million headline figures, as money he magically made. That’s clearly not how anything works. The actual balance sheet looks very different.

Line Item Headline Number Actual Number
Gross cargo value (all held tankers) ~$500M ~$500M (if every barrel is eventually sold)
Emergency repairs (fleet) Not reported $160–$320M (based on Skipper rate)
Ongoing maintenance, crew, insurance Not reported $16–$28M per month, compounding
Oil storage Not reported ~$3.6M per month (est. across loaded tankers)
Military operations (carrier groups, SEALs, 160th SOAR, CG cutters) Not reported Classified / buried in defense budget
Legal fees and court costs Not reported Unknown; 10 separate forfeiture cases
Empty tankers (Bella 1, Veronica) “Seized!” Pure liability; $0 revenue
Returned tankers (Sophia, Olina) “Seized!” Sunk cost; $0 revenue
Environmental tail risk (OPA 90) Not mentioned $3–65 billion per incident, uncapped
Net position at 12 months “Financial boon” +$148M (best case) to –$156M (Skipper rate)
Net position if one hull fails –$3 billion to –$65 billion

Every day Trump’s seized liabilities sit in U.S. waters, the gap between artificially gross headlines and the balanced reality ledger widens.

The one number that should keep the DOJ’s asset manager awake at night is the OPA 90 tail risk of a single-hull structural failure in a Texas anchorage, which doesn’t appear in any press conference. Bush signed that law. Sununu gutted the intent. And now Trump is parking the exact category of vessel it was designed to eliminate — single-hull, uncertified, end-of-life tankers loaded with heavy crude — in American waters, on the American taxpayer’s tab, with the American coastline as collateral.

This is what Trump Oil looks like, just like every other Trump bankruptcy, as court filings reveal the disinformation behind his toxic press releases.

Security

Where’s Ed: Anthropic Told Court $5 Billion But Public $19 Billion

Leave a comment

Ed Zitron just published two pieces on Where’s Ed: “The Beginning of History” (March 10) and “Why Are We Still Doing This?” (March 17).

They land a clean hit on Anthropic’s hallucinations in financial storytelling. The math is simple enough that it can’t be denied. Let me show you how.

What Anthropic Told the Court

Anthropic’s Chief Financial Officer Krishna Rao filed an affidavit on March 9, meaning he swore it was true, in their lawsuit against the Department of Defense. It stated that Anthropic’s total revenue “to date” was “exceeding $5 billion.” That’s all the money Anthropic has ever made, from its founding day through March 9, 2026.

What Anthropic Told Everyone Else

Throughout 2025 and into 2026, Anthropic repeatedly announced its “annualized revenue”, which doesn’t match the affidavit.

Say you run a lemonade stand. In July, you sell $10 worth of lemonade. Someone asks how your business is doing. Instead of saying “I made ten dollars, that is true,” you say: “I have a $120-a-year pace no doubt!” That’s annualized revenue. You take one month of actual money, multiply by twelve prediction months, and report the biggest number you can.

Here are the annualized revenue figures Zitron compiled from Anthropic’s own announcements and press coverage, with sources:

Date Annualized Revenue Implied Monthly Revenue
January 2025 $1 billion $83 million
March 11, 2025 $1.4 billion $117 million
March 30, 2025 $2 billion $167 million
May 30, 2025 $3 billion $250 million
July 1, 2025 $4 billion $333 million
July 31, 2025 $5 billion $417 million
October 2025 $7 billion $583 million
December 2025 $9 billion $750 million
February 12, 2026 $14 billion $1.17 billion
March 3, 2026 $19 billion $1.58 billion

The right column is the story. If annualized revenue means “this month times twelve,” then dividing by twelve gives you what they actually made each month.

The Addition

Now when we add up the monthly revenues we see a problem. This is where Where’s Ed earns his keep. Anthropic told the world at ten different points how much it grew. Each implies starting from an actual monthly number. Add them up, estimate the gaps between announcements, and you get a total.

Zitron’s figure: roughly $6.66 billion in implied cumulative revenue through early March 2026.

The CFO’s sworn figure: “exceeding $5 billion.”

Those numbers are not a match.

Ten reports across fourteen months, each covering a distinct measurement period. Zitron’s gap-period estimates are even conservative for the uncovered stretches between reports. He uses the lower ARR figure rather than interpolating upward. The $6.66 billion is a floor way higher than $5 billion, not a ceiling.

The Speedometer and the Odometer

You don’t even need the full table to see the problem. Just take the last four months. December 2025 through early March 2026, using Anthropic’s own ARR figures, implies roughly $4.25 billion in revenue ($750M + $750M + $1.17B + $1.58B). That would have to mean everything before December 2025, which would be the entire prior history of the company, produced less than $750 million. Look at the table again. That’s impossible.

Think of ARR like a speedometer. It reports how fast you’re going right now. Total revenue is your odometer. It tells you how far you’ve actually traveled.

If the speedometer says you’ve been doing 100 miles per hour for the last hour, but the odometer says you’ve only gone 40 miles, the speedometer is lying. Or more precisely: when you hit 100 for a split second you reported it as your cruising speed.

Rao’s low revenue word choice matters here. He said “exceeding $5 billion” and not “nearly $6 billion,” not “approaching $6 billion.”

In a filing where Anthropic was trying to impress a federal court with its commercial scale, Rao is expected to use the biggest number he can. “Exceeding $5 billion” tells you his real figure is much closer to $5 billion than to $6.

That puts overstatement implied by the ARR figures somewhere around 25–35%!

Which Anthropic Do You Believe?

If the $5 billion lifetime figure is the truth, as sworn under oath in federal court, then the annualized revenue figures don’t mean what they are meant to say. “Annualized” at Anthropic may not mean “last month times twelve.” It might mean best week times fifty-two. Or best day times three-sixty-five. Or something else entirely that makes the number as large as possible.

There is one charitable reading. ARR sometimes counts signed contracts with money promised, not money received. Rao’s “revenue to date” likely means recognized revenue, only money actually earned. If Anthropic has billions in contracts where service hasn’t been delivered yet (possible given the huge boost in February), the ARR looks huge while total revenue stays lower.

But if that’s the explanation, Anthropic was reporting unearned contract value to the press as though it were operating revenue, while reporting actual revenue to the court. That’s not an accounting distinction. That’s two different stories tuned for two different audiences.

Either way, the conclusion is the same.

Every headline that reported Anthropic’s annualized revenue as though it indicated actual business scale was wrong. Every valuation model built on those figures was fed inflated inputs. Every investor who used ARR trajectory to justify Anthropic’s $380 billion valuation was working with a disinformation number.

The lemonade stand making $10 in July that told everyone it is a $120-a-year business? Anthropic somehow screwed that reporting up despite regulators, sophisticated investors and the global financial press. Oh, and despite having AI as its core product. Or is it because of AI?

To believe both the ARR headlines and the CFO’s affidavit, you have to believe that Anthropic’s business was essentially non-existent for its first four years and then suddenly processed 85% of its entire lifetime volume in the last 100 days. Wow.

Zitron cleverly asked for proof.

Anthropic math implies flawed integrity. Just like AI.

History, Security

Yglesias Defends Big Tech Bros Fleecing the Poor: “Let Them Eat Shovels”

Leave a comment

Matthew Yglesias runs a Substack called Slow Boring where he routes every problem in American political economy through zoning reform. His latest piece asks why Silicon Valley hasn’t done more for most Americans, and his answer is: not enough apartments near Cupertino.

Facepalm.

Paul Krugman had pointed out that tech generates a negative externality by producing billionaires who corrupt democracy. True.

Yglesias called this “puzzling” and used it to change the subject to housing density.

The problem is that Yglesias doesn’t seem to know the history of the examples he’s citing. Yup, I said it. HISTORY. Pull up a chair because I’m about to open a can of whoop-history on Yglesias.

He invokes Chicago in 1900, Detroit in 1920, the California Gold Rush, and Shenzhen. My God. He pulls all of that to our attention without appearing to notice that every one of these is a well-documented case study in the exact failure mode he’s ignoring.

Imagine being the guy who says Germany 1938 is a great example of how broken windows can fuel the economy.

Yeah, that bad.

The Fabian Society was founded in 1884 specifically because the industrial boomtowns Yglesias romanticizes were producing spectacular wealth for owners and squalor for everyone else. It’s like Krugman was so right that he didn’t even have to use history to know it, but if he had it would have cemented his point even more. Meanwhile Yglesias responds by walking through a minefield of his own examples and stepping on every one.

Here’s what Yglesias says, and what Fabians discovered over a century ago. Like explaining water is wet, I humbly present now, something hopefully obvious.

What Yglesias Says vs. What Fabians Would Say

Dimension Yglesias (Slow Boring) Fabian 1880s Critique
Why hasn’t tech helped most Americans? Housing constraints prevented a megacity from forming around Silicon Valley Private capture of publicly-funded innovation prevented democratic benefit
Proposed mechanism for shared prosperity Build denser housing near tech campuses so service workers can “sell shovels during the gold rush” Graduated taxation, public ownership stakes, municipal enterprise, democratic governance of technology
Role of the state Get out of the way — remove zoning restrictions Capture monopoly rents, fund universal public goods, regulate concentrated power
Who creates value? Tech founders and employees, radiating outward through spending Public universities, government-funded research, workers, infrastructure — tech founders captured value others created
What “the boom” looks like Population growth, construction, rising property values — Shenzhen, 1900s Chicago Rising wages, universal healthcare, public education, democratic workplace governance — postwar Britain
The billionaire question Not addressed — Krugman’s point about political corruption is replaced with a housing supply argument Billionaires are a policy failure. Concentrated wealth is concentrated political power. That’s the point.
Historical model invoked Industrial-era boomtowns (Chicago, Detroit) — workers flocking to capital The very boomtowns that produced child labor, tenement squalor, and Pinkertons — prompting the Fabian movement in the first place
Utopian vision Apartment towers in Marin County (cites Star Trek: Picard) Star Trek’s actual economy: no money, no landlords, replicators are public goods
What’s invisible Ownership. Power. Democratic control. Who decides what gets built and for whom. Nothing, these are the starting questions
Treatment of Krugman’s argument “Puzzling assertion” claim to dismiss the political corruption claim and jazz hands into housing Krugman understated it. The corruption is the business model. It’s not external.
If you force enough Stanford kool-aid into the mix, does it even matter what else exists?

Let me just reiterate that Shenzhen is government-owned land, state-directed investment, and party-controlled development. It’s literally the Fabian model, as the state captured the land value. Yglesias completely inverts reality and cites his error as his evidence for removing zoning restrictions.

Similarly, “selling shovels during the gold rush” is famous precisely because the miners with shovels went broke. It proves Yglesias wrong. Sure, Levi Strauss and Sam Brannan got rich by being smart while hard workers lost everything and died as nobodies. We’re supposed to want that? But the real lesson not to avoid is the abject cruelty, like the man who built the university at the center of Silicon Valley who got rich through government fraud, racism and genocide. That’s some devastatingly real harm Yglesias is romanticizing.

But what do I know. I’m not on Substack.