‘Digger’ is the year’s big climate movie. This climate reporter hated it.
If you scroll down on the list of “Carbon Majors,” the 178 producers that are responsible for most of the world’s fossil fuel emissions, you will eventually get to a company called Continental Resources. This firm is ranked around 150, well after giants like Saudi Aramco and ExxonMobil, but it has still produced almost 600 million tons of carbon dioxide since 2005 alone, equivalent to the combined carbon footprint of around 2 million Americans over that same time. The founder of this company is a man named Harold Hamm, the spitting image of an American oil executive — he grew up dirt-poor in Oklahoma and worked his way into fracking riches, but he still speaks with a Southern drawl. He now uses his wealth to influence the government, raising millions of dollars for Republican candidates and funding crusades against green energy.
If you squint, the protagonist of Alejandro González Iñárritu’s new film Digger could pass for Hamm, or for the Hamm-like archetype that has bled into the American psyche. The film centers around an oil executive named Digger Rockwell, a Texas-accented plutocrat who has built his father’s wildcat oil company into a global concern with offshore rigs in Greenland and Alaska. As depicted by Tom Cruise, he is the ideal of the Southern businessman-ideologue, partial to monogrammed private jet seats and roaring rants about his own genitalia. Other than making money, his sole concern seems to be the welfare of his white cat Maggie, whom he is trying to keep alive against veterinary advice even as she decomposes in his arms.
Digger examines two questions. First, how much responsibility does such an individual bear for the deaths and displacements caused by global climate change? Second, what is the right way for an artist to consider the first question? If you think you already know the answers to these questions, this may be the movie for you, though I can’t promise you’ll enjoy watching it. In trying to construct a fable about the hubris of oil titans like Hamm, Iñárritu has produced a film that says nothing except what the viewer already knows. He is all too aware that political parables about corporate greed are of limited artistic value, but he can’t manage to escape the confines of parable. The result is a mess of shrieking and spittle that will leave even the most climate-conscious viewers reaching for an Advil.
After an opening scene in which the audience is given the chance to marvel at Cruise’s physical metamorphosis into a flabby oil executive, the plot arrives with the delicacy of a tornado siren. A technician on Digger’s rig in Greenland calls to tell him that a methane leak has opened up a large crack that threatens the stability of the island’s enormous ice sheet. Before Digger can even finish a monologue about how he’s not going to shut down a billion-dollar platform because of a tiny crack, another group of company officials arrives on his doorstep to tell him that the rig has blown up and the Greenland ice sheet has already collapsed. Now a massive iceberg is headed straight for Europe, with the potential to flood the continent and kill hundreds of thousands of people. Digger confers with the president (John Goodman) to figure out what to do, and they quickly decide to keep drilling.
This opening act achieves a level of heavy-handedness that remained out of reach even for Adam McKay’s Don’t Look Up, the last major film to satirize those responsible for global warming, which likened climate change to a comet about to hit the earth. Digger’s proposed response to the crisis is to deploy a digital tool called the “firewall” to censor all mentions of “methane,” “iceberg,” and “climate change” from the Internet, while the president and his conservative cronies make predictable suggestions about bombing the iceberg or praying to God for deliverance. Whereas McKay opted for CNN-toned realism, Iñárritu’s aesthetics are a more interesting mix between Wes Anderson and Thomas Hart Benton — Digger’s boardroom is full of Art Deco murals depicting oil discoveries, and above the table hangs an obvious sculpture of a snake eating its own tail.
Digger Rockwell (Tom Cruise) meets the president in a boardroom to discuss the iceberg approaching Europe. Courtesy of Warner Bros. Pictures
The dialogue is full of jokes about climate change that will likely be catnip for the small minority of viewers who already spend most of their days thinking about it. There are references to the outsize global warming potential of methane and to a supposed “COP30 fiasco” at which the U.S. alienated its allies by withdrawing from the Paris Agreement. (In reality, they were almost happy to see us go.) There’s even a reference to thawing Arctic permafrost that will be familiar to devoted Grist readers. The vehicle for these revelations is a series of dialogues between Digger and his in-house scientist Ganesh (Riz Ahmed), each of which devolves into a fusillade of worn-out culture war tropes. Ganesh rejects Digger’s offer of delectable sashimi on the grounds that the bluefin tuna population has declined by 97 percent, “even though this administration won’t label them endangered.” He then tries the sashimi and loves it.
Whether or not this satire provides an accurate picture of the world economy, it does not make for enjoyable moviegoing. Well over half the movie consists of people screaming their lungs out about methane or the stock market, culminating in an unbearable scene in which Digger berates the European parliament, calling them hypocrites for using his product while condemning his company. The Danish and German legislators howl at him from the benches, accusing him of genocide for allowing the iceberg to hit Europe, and Digger screams back for several minutes about how the world needs his crude. “You want me to keep producing,” he says. “You demand, and I supply!” If you’ve ever worked in climate change, participated in climate politics, or gotten into a debate about climate change on social media, the whole exchange will feel familiar, and not in a good way.
If I hadn’t been assigned to review the film, I might have walked out after the European parliament scene — not because of what Digger was saying, but just because it was ugly, loud, and unpleasant. Regardless of whether you agree with Digger or with the Europeans, who wants to listen to all that? In a Cruise-appropriate twist, the most pleasant moment of the entire film is when the U.S. Air Force scrambles a squadron of fighter jets to carpet bomb the iceberg. It’s almost the only moment in the movie where everyone shuts up.
The movie includes a twist, or perhaps better to say a reveal, that is intended to show that Iñárritu is up to more than McKay was in Don’t Look Up. At the risk of giving too much away, this isn’t just a parable about climate change, it’s a parable about a parable about climate change. The apparent intention of this narrative complication is to ask what might be achieved by subjecting a viewer to a ham-fisted metaphor like the one the audience has just endured.
McKay’s obvious hope in Don’t Look Up was that such fables might spur people to action, and we can see how well that worked out a few years later — Donald Trump, a climate denier, received almost 80 million votes and triggered a global trend of “climate hushing,” or refusing to look up. Watching a movie about a tyrannical oil executive isn’t likely to make anyone buy an electric vehicle any more than reading the list of the largest emitters would.
So why bother? Iñárritu seems to want to make the case for the value of chronicling our state of affairs, even if that chronicle drifts into caricature. A few guys really have gotten rich selling a product that has enabled global welfare while also threatening the continuity of that welfare. Those guys are much more inclined to continue selling their product than to stop. Isn’t there some value in preserving that truth for our descendants? During a scene whose context I won’t reveal, one character says the following: “It’s ridiculous, and it’s scary, and it’s funny, and it’s stupid. It makes no goddamn sense, but it’s whathappened.”
Be that as it may, a film still needs a heart, and Digger doesn’t have one any more than Digger himself does. Despite the shots of Cruise looking longingly at his disintegrating cat or at an oil rig sunset, it doesn’t even achieve the level of emotional resonance of a Marvel movie. It has about as much human warmth as a Twitter debate about permitting reform, and its final suggestion that stories can provide solace in times of great calamity is a classic instance of an artist telling where he should be showing.
If Iñárritu’s point is to show through metaphor that corporate greed has abetted the climate crisis, well, he has made that point only too well. If the point is that this kind of metaphor is simplistic and of dubious political utility, he has also made that point. Even so, viewers looking for stories about the sins of big business may wish they had spent their hard-earned wages on a showing of Coyote vs. Acme, the acclaimed recent Looney Tunes revival. At least there, when the TNT crate drops, you don’t have to think before you laugh.
A note before we start. What follows works through the innovation loops of the semiconductor world, sub-sector by sub-sector, with eight diagrams along the way. They are complicated, despite a fair amount of effort to make them otherwise.
Anyone who wants to skip the loops and the slides can find the conclusions in two places: this opening section and the closing one about 4,000 words below, where the points that matter for money are stated plainly. That said, the loop diagrams are fun for those who relish complexity or those who want new reasons to be negative.
A core GenInnov belief is that innovation is the only true growth theme left in markets. Listed companies in growing numbers have realized that the key to better valuations in an AI world is to talk about new products, new services, and new business areas. Using the latest technology to shave a few points off costs is table stakes, and it wins nothing for long.
So the number of companies attempting real innovation keeps climbing, first inside technology and now well outside it, at a pace that used to belong to a certain kind of venture company. One difference stands out. Most of what gets announced is a plan. Even in the best case, most plans need many years and a great deal of money before anyone can tell whether they worked.
The industry that follows innovation is changing with it. The days when a three-year forecast, extended from current trends, counted as analysis are ending. Innovation writing has moved from a handful of blogs into mainstream analyst reports.
What has traveled less well is the habit of the corner. Most commentators pick a patch they can picture, and the patches that are easy to picture collect the attention: a new agent, a new chat interface, a new consumer gadget. A harder patch gathers a following, too, usually a specialized one that reads only itself.
Two failure modes arise from this. The first is underrating how quickly a good idea gets copied when nothing protects it, as persistent agent products like Muse, Grok Bot, and Instinct are demonstrating publicly this week.
The second is quieter. Innovation attempted is innovation achieved only some of the time. Physics does not yield to applause at a conference, and plenty of announced things will never ship, whatever the research or however charming the speaker. It is worth noting how often the word quantum turns up when somebody wants to signal that they are thinking big.
For anyone who enjoys a loop diagram – and the pessimists do love a loop diagram – the real difficulty lies elsewhere. Looking at one technology in isolation and assuming that everything going right inside it leads to money has become dangerous.
Whether an innovation makes money now depends heavily on other innovations in the next field along, and often in fields far away. Some of those are prerequisites. Some are substitutes waiting to make the whole effort pointless. Some are customers whose own success decides whether anything gets paid for.
Drawing arrows for money moving among a dozen companies takes an afternoon and proves very little. Drawing what has to be true in six other places before one technology pays off is more complicated, but likely more important.
This piece stays inside one industry, semiconductor manufacturing and the infrastructure built directly on it. Similar loops run through equipment and materials, through models, through applications, and through biotech, and they deserve their own pieces. What follows starts with a master diagram of the whole spiral, then takes its seven buckets one at a time. Once again, anyone who tires of the technology is welcome to jump straight to the conclusion, about 4,000 words and eight diagrams below, where the points that matter for money are gathered.
The spiral and why it turns
The master diagram is the last tidy thing in this piece. Its seven stations – compute, memory, packaging, connectivity, power and thermal, qualification, and capital – organize the problem. In practice, several move at once and progress in one can change what is needed from the others.
Take memory. An accelerator waiting for data gains little from being able to calculate faster. Feeding it better can make more of its existing capability useful. But designing a system around that improvement brings other requirements into view: how the memory connects to the processor, how the assembly is manufactured, and how its heat is removed. The promised gain depends on several industries delivering together. Each has its own development schedule.
That dependence is what gives the spiral its shape. A constraint attracts effort and money. Relieving it makes a more ambitious system possible, which puts pressure elsewhere, and the money that arrived to relieve the first constraint has usually already committed itself before the second one appears. The next round begins at a higher level of bandwidth, density, power and capital intensity, sometimes before the previous investment has earned its keep.
There is a less comfortable route around the picture too. Customers can respond to an expensive constraint by finding ways to need less of it. A memory producer may expand supply, a chip designer may change the architecture and a model developer may reduce the workload’s memory needs. All three can make technical progress. Their business plans may be less compatible.
This is where the arrows become useful for investing. A supplier’s opportunity depends on how long its contribution remains necessary, how readily customers can substitute for it, and whether the rest of the system will be ready in time. Scarcity can create pricing power. It also gives customers a reason to finance an escape.
Read the diagrams with those competing responses in mind. The maturity labels matter because an available workaround can shape purchases while a more elegant solution is still being qualified. A place on the same page does not mean a place in the same purchasing cycle.
Compute has one incumbent and many forks
The compute diagram is crowded because several different arguments are taking place at once. Some challengers want to build a better general-purpose accelerator. Others want a chip tailored to a narrower workload. Still others aim to get more useful work from the hardware already installed. Success in one route changes the market available to the others.
Start with what the challengers are challenging. The cadence is now annual, and the product being sold has grown from a chip to a rack of dozens of them, tied together by a proprietary fabric. A competitor therefore has to match a system – its memory, interconnect and software – which is a far larger undertaking than matching a part. That explains three things on the diagram: why the same few design houses sit behind most custom programs, why a national stack built under export controls counts as a separate branch and why some challengers end up licensed into the incumbent instead of beating it. Absorption pays differently than winning or failing.
Consider the pull toward custom silicon. At sufficient scale, a cloud operator has a reason to design around its own workloads and seek better economics. But the finished design still needs manufacturing capacity, memory, packaging, chip connections and software to make the system useful. The commercial question extends well beyond whether the chip works. It includes how much of the hoped-for saving survives those requirements and whether the workload remains suitable when the system arrives.
Specialization makes that timing question sharper. A design optimized for one pattern of computation may perform very well, and its investment case depends on that pattern lasting long enough to recover the development cost. A chip that wins by hardwiring a model architecture is exposed to any change in model design, and the shift toward sparse models, where only a fraction of the parameters is active, is already underway.
The efficiency routes further complicate the calculation. Lower precision, sparse models and separating the stages of inference seek gains by changing how the work is done. Their implications differ across workloads and they force the same forecasting question: How much hardware will be needed for a given amount of useful output? Lower costs may then encourage greater usage, so a gain in efficiency alone does not settle the demand forecast.
Several competing designs may draw on the same constrained suppliers. Others may change the type or amount of memory and connectivity required. Counting announced accelerators tells us little about how those demands will add up.
That takes us to memory, where efforts to supply more and efforts to need less meet in the same diagram.
Memory is where the scarcity sits
Memory used to improve in one direction. A new process packed more bits into a die, and everything else followed. High-bandwidth memory now advances along five vectors at once. Faster raises the signaling rate on each connection. Wider doubles the interface to 2,048 connections, buying bandwidth at the cost of routing density. Taller moves stacks from eight dies to twelve and sixteen, adding capacity with no additional footprint and shifting the problem to thinning, bonding and assembly yield. Smarter builds the base die at the bottom of the stack on a logic process, so it can absorb control, test, and repair functions. Repartitioned moves the boundary between accelerator and memory subsystem itself, a system architecture decision wearing a memory costume and the one our earlier piece, Memory by Design, followed farthest.
The five vectors overlap. One generation ships in volume while the next ramps; samples of the one after that reach customers; and more ambitious concepts appear at conferences. They compete for engineering effort and capital on very different schedules. A customer choosing a system today must work with the memory that can be delivered and qualified in time.
Scarcity adds a second set of responses, and part of its cause is physical. A stacked part consumes roughly three times the wafer area per bit of ordinary DRAM, so every order for it squeezes everything else made in the same fabs. Memory producers work to supply more bandwidth and capacity. Their customers work to get by with less of the expensive kind. The diagram follows both, because success on either side changes the opportunity on the other.
Some responses stay within HBM, using shorter stacks and spreading the workload across more accelerators. Others change where data sits. Placement software decides what must remain close to the processor and what can move into cheaper memory or storage. Model designers compress the key-value cache and reduce the memory needed for weights. Each route has limits: a saving in scarce memory may require more data movement, more hardware elsewhere, or a compromise in performance.
This is where the loop becomes commercially awkward. A supplier expanding capacity must judge how much demand will remain by the time that capacity arrives. A customer developing a workaround must judge whether the saving will still justify the extra complexity. Greater availability and lower prices can weaken the case for substitution. Persistent scarcity can help a workaround become an established design choice.
Taller stacks put pressure on bonding, cooling and assembly yield. Moving data farther from the accelerator increases the burden on connectivity and placement software. Memory innovation therefore changes both how memory is built and how the rest of the system uses it. Packaging sits at the intersection of several of those choices.
One exclusion is worth naming. The storage side of the hierarchy appears here only where it reaches back toward the processor, in flash placed behind a memory interface and in context offloaded to SSDs. Everything below that line has been left out: NAND scaling past four hundred layers; cells and control circuits built on separate wafers and bonded together; controller design; and the interfaces that move data in and out. Those are loops in their own right, with their own substitutions and their own suppliers, large enough to need a diagram of their own. Treat this picture as the part of the hierarchy that competes directly for the accelerator’s attention.
Packaging became the computer
Packaging used to be treated as the last step before a chip left the factory. It now decides how much computer fits in one place, giving it a scaling roadmap as ambitious as anything in logic. It also brings more of the system’s value into a single assembly, where a manufacturing problem can become costly.
The lateral race is about area. Larger packages provide more compute and memory, provided the connections between them can be manufactured reliably. Current parts already span several times the area of a single lithography field and the published roadmaps go beyond nine, with a dozen or more memory stacks, before the end of the decade. Those are supplier roadmaps, and the whole industry is planning against them. Alongside sit alternatives with different economics: local silicon bridges, larger manufacturing panels, glass substrates and additional capacity at outsourced assembly houses. Each offers a route to a larger area or lower cost, and each must establish that it can deliver enough good packages to make the savings real.
The vertical race is about pitch. Bringing dies closer together allows denser connections and makes surface quality and assembly precision more demanding. Copper-to-copper hybrid bonding opens possibilities beyond conventional micro-bumps. A fine-pitch research demonstration, a production tool, and a process qualified for a customer’s product are separate achievements. Progress in one does not establish the other two.
Nor does the most advanced method have the timetable to itself. If a change in permitted package height gives an existing bonding process more room, customers may be able to reach their next capacity target without adopting its successor. A relatively modest change elsewhere in the system can extend the commercial life of an older technology. The new process then has to compete against an incumbent that has bought itself more time.
What moves into the package makes the choices harder. Memory stacks bring height and thermal requirements. Optical engines bring connections that must be assembled and tested and kept reliable. Changes in power delivery alter the electrical and physical design and more ambitious cooling proposals bring the thermal solution into the assembly itself. Each changes what the package must accommodate and how it can fail.
More integration promises better system performance, encouraging larger and more complex assemblies, whose manufacturing demands raise the value of bonding, inspection, testing and thermal control. By the time a package is assembled, it carries a great deal of known-good silicon, so a late failure scraps everything bonded with it. If cost or yield disappoints, customers may change the design, retain an older process, or place some functions outside the package. Those choices reshape demand for the capacity and equipment being built to serve them.
The durable advantage may therefore lie with whoever can co-design the interposer, bonding, thermal path and inspection as a single yield-learning system. The evidence is how reliably an ambitious design becomes a shippable product, and at what cost.
Connectivity has many paths to light
Optics has been the future of short-reach links for twenty years. It is arriving along several paths at once, which makes the investment question harder than simply predicting more light.
The driver is the amount of time an expensive system spends waiting. More compute within a node increases the demand for communication between processors. Faster chips deliver less than their promised gain if data movement and synchronization cannot keep up. Packaging brings more compute together; connectivity determines how effectively it works together.
As electrical lane speeds rise, loss and the power needed to recover signals make longer copper paths harder to sustain. At around 200 gigabits per second per lane, that cost becomes hard to justify over any distance, and the date at which each link crosses that threshold sets the timetable for everything else on the diagram. Optics moves closer to the chip through competing designs with different compromises. Conventional pluggables preserve easy replacement. Linear-drive optics remove processing from the module and place more demand on the electrical channel. Near-package optics shorten that channel while retaining some separation. Co-packaged optics bring the optical engine into the package, making optical assembly and repair part of a much more expensive system. These routes can coexist across different uses while competing for the same design decisions, and putting their bandwidth numbers side by side can make the contest look more settled than it is.
The manufacturing dependencies multiply as optics moves inward. Silicon photonics, drivers, fiber attachment and thermal control must work together, with different suppliers responsible for different parts of the result. Lasers deserve separate attention. They are built on indium phosphide, their capacity is tight and they are the component most likely to limit how quickly more integrated designs can scale. One reason a socketed engine with an external laser source exists at all is to keep that constraint outside the expensive package.
Standards add a second contest over how processors talk to each other inside a rack. This is the compute argument reaching its next station. A proprietary fabric ships in volume today. Two open alternatives are working to displace it. The question is whether an open standard can arrive in time to matter while the incumbent is already selling complete systems. Optical circuit switching sits alongside all of this, reducing packet-switching requirements where traffic patterns allow.
Faster links allow larger compute domains, whose communication demands encourage another increase in link speed. Moving optics closer reduces the electrical burden while raising the importance of alignment, attachment yield, laser reliability, testing, and repair. Progress transfers some of the difficulty from moving signals to manufacturing and maintaining the system.
Copper keeps changing the timetable through active cables and shorter electrical paths. Every extension of its useful life gives customers a reason to postpone an optical transition, and a workable intermediate optical design can delay a more integrated successor in the same way. The investor therefore needs a view on which links change first, which components those designs require, and how long they remain useful before the architecture changes again.
Power runs through every one of those choices.
Power and thermal set the schedule
Rack power is climbing, and the arithmetic is unforgiving. A megawatt at 54 volts draws about 18,500 amps; at 800 volts, about 1,250. For the same conductor resistance, losses rise with the square of current. Higher distribution voltage reduces the burden on copper, while moving major conversion equipment outside the compute rack frees space inside it.
The diagram separates two changes that should not share a maturity label. Sidecar power racks relocate conversion equipment and are close to deployment. Solid-state transformers attempt a deeper redesign, converting medium-voltage AC to the DC supply the system needs. This requires switching devices, magnetics, protection, controls and cooling to advance together. Progress in power semiconductors helps, but an available device is still several qualifications away from a usable power system.
The electricity then becomes a cooling problem. Heat must cross the chip, the cooling interface, and the facility before it reaches the outside world. Direct liquid cooling is now the standard for dense racks, specified with warm water so mechanical chilling can be reduced where conditions allow. What follows is genuinely open, with two-phase cooling, immersion and coolant channels inside the silicon, all shown on the diagram, and none of them settled. Whichever wins will need connectors, distribution equipment and service procedures that customers trust, which is a slower business than improving the thermal physics.
The underlying difficulty is synchronizing historically different clocks. Accelerators, cooling, converters, and rack interfaces increasingly have to meet the same deployment date. Utility connections, protection standards and lifetime qualification have not kept pace with semiconductor cycle times. Their decisions nevertheless have to be made against the compute roadmap.
The largest loop on this diagram runs through the grid. Denser computing raises power demand, while delays in grid access encourage builders to seek their own generation or secure existing supply. Those alternatives bring their own equipment queues, fuel requirements and approvals. Batteries can bridge interruptions and shift demand, and they still need an energy source. Flexible workloads offer another bargain: Accept some curtailment in exchange for an earlier connection.
Each response changes the next investment decision. Better cooling makes a denser rack feasible, and its power requirement may change the site or the opening date. On-site generation can shorten one wait while adding another. What matters is how soon the whole installation can operate reliably, and how much useful compute it can deliver once it does.
Qualification is the real gate
Capacity counts only once it is qualified. A new packaging line or memory supplier adds little usable supply for a customer until its parts pass that customer’s requirements. As packages combine more dies, the distance between installed capacity and dependable output can grow. A factory opening is easier to announce than a reliable yield.
The expense accumulates during assembly, and the order of operations makes it worse. Memory stacks go in early, so everything bonded afterward sits on top of a part that was screened before the expensive work began. A defective die or bond discovered late can put the entire package’s value at risk. Known-good-die screening catches problems before more silicon is committed, and testing between assembly stages checks what the next step might otherwise bury. Burn-in stresses parts to expose early failures, and some of it is moving from finished systems to the wafer, shifting the work and spending to a different point in the chain. System-level test runs parts under realistic workloads, because passing an electrical check does not establish that a processor will behave reliably in a server. Testing high-power parts then requires substantial power delivery and cooling.
Inspection has to reach places that assembly hides. X-ray and acoustic methods look inside stacks for voids and delamination. Finer bonded interfaces demand tighter measurement and defect detection, and automated classification helps sort the resulting flood of images. The difficulty is finding consequential defects quickly enough to keep production moving. Longer tests improve confidence while reducing the output of a given set of machines.
Qualification also allocates market share. A memory supplier that clears a customer’s requirements earlier can secure orders while a rival’s installed capacity remains unavailable to that program. New substrates, optical assemblies and liquid-cooling systems face their own reliability tests. A second source offers little bargaining power until the customer can actually use it.
For equipment suppliers, spending arrives at different times, and the split is now visible. Electrical test is collecting orders today because it scales with production that is already running. Demand for each inspection tool depends on which bonding processes reach volume and which measurements customers require in production. Some inspection steps become more intensive, while others must become faster or more selective to remain economical. More complexity creates opportunities, and it does not make every proposed test a permanent addition to the line.
The feedback continues after shipment. Large fleets reveal silent data errors that factory tests missed. Those findings inform tighter screens and revised designs, which improve reliability while adding test time or reducing the share of parts that passes. The supply number changes without a single factory being added or removed.
The scarce resource is therefore reliable output at an acceptable cost. Installed capacity is only the starting figure.
Capital follows scarcity
Every bottleneck in this spiral draws money. Scarcity supports high returns; those returns encourage spending, and new capacity changes the conditions that justified the investment. The factory takes years to build. Its selling price can change a good deal faster.
The financing now reaches well beyond cloud cash flows. Bonds, sovereign capital, private equity, and private credit spread the cost across more balance sheets. Accelerators become collateral. Multi-year leases and take-or-pay contracts turn expected demand into commitments that lenders can use to lend. Suppliers invest in customers, and developers contract power before their buildings exist. These arrangements help projects proceed earlier, while tying repayment to decisions elsewhere in the system. A long contract allocates risk; it does not eliminate it.
The money reaches memory fabs, leading-edge foundries, packaging lines and power equipment. Each expands on a different schedule. A completed building may wait for electricity; installed equipment may wait for qualification. Meanwhile, alternatives developed during the shortage can change what customers want by the time the new capacity arrives. Capital can relieve a constraint and still earn a disappointing return.
The first difficult question is how long the assets earn. An accelerator’s physical life, its accounting depreciation and its competitive earning life are three different things. Older hardware can remain useful while commanding a lower rental price or consuming more power per unit of work than its successor. Whether it earns adequately for three years or six matters to its owner, its lender and anyone relying on its collateral value. Extending a depreciation schedule changes reported earnings. It does not secure another customer.
The second question is circularity. Supplier investment and vendor financing can help customers adopt useful technology. They can also make orders look more independent of continued funding than they are. Calling the arrangement circular settles neither point. The work is to follow the cash: who ultimately pays for the service, whether that revenue covers delivery and financing costs and which participant must raise more money if it does not. Contracted revenue is valuable, and its quality depends on the customer’s ability to honor the contract. In several cases, that ability depends on the customer raising fresh capital from investors who read the same contract as evidence of demand.
The hopeful outcome is that cheaper capability brings enough additional paid usage to support the investment. The harder outcome is that usage grows while prices fall faster than costs, leaving impressive activity and inadequate returns. Both can accompany rapid technical progress. The spiral can keep turning even when some of the capital that financed it does not come back.
The conclusion, for those who skipped
Eight diagrams later, the honest summary is that the picture is a mess, and the mess is the finding. By the time you finish reading, a few more daring plans will have been announced and congratulated as achievements. Asking about the probability of delivery, at the proposed cost and on the promised date, can spoil the occasion. Mentioning a competing announcement that makes the whole undertaking less useful is positively rude.
The vocabulary makes it easier. A roadmap becomes a schedule, a sample becomes a shipment, an architecture becomes a product and a demonstration becomes a capability, all without anyone saying anything untrue. Physics has not agreed to any of it. Physics does not attend conferences, has never been moved by a keynote and remains unimpressed by the word quantum, which turns up most reliably when somebody wants credit for thinking big.
Several of these technologies can succeed while undermining one another’s business plans. Scarcity finances both more supply and ways to need less of it. Today’s indispensable supplier may be tomorrow’s expense to engineer out. Success buys resources for the next contest. It buys no exemption from it.
So we want evidence before commitment, and room to change our minds afterward. Qualification results, repeat orders, yields and cash collection carry more weight than a launch date. Waiting costs some upside. Paying a valuation that assumes every unfinished step will work can cost considerably more.
Keeping the exits open matters at least as much. A sound investment can become unsound because somebody else succeeded, which is a strange way to lose money and a common one. So we keep checking what customers require, what competitors can replace, and whether the original economics survive. Loyalty to a thesis is expensive in our Super-Moore era.
That is the particular attraction of public markets. They let us stay current with an evolving technology landscape and move our exposure as the evidence changes. No technology or product endeavor stays certain of success for long, and the ones that look certain are usually the ones being priced as though the next six dependencies are already solved.
And what a future it is attempting. This is an extraordinary era of innovation, increasingly built through scientific and engineering undertakings far beyond the garage and the modest seed check. New materials, intricate manufacturing, optical links and power systems require vast capital and coordination across disciplines. The ambition deserves celebration. Much of what these teams are attempting is astonishing.
We can applaud the attempts while recognizing that a meaningful number of announcements are turning more bombastic and impractical.
Nilesh Jasani is the founder and CEO of GenInnov Pte Ltd Singapore. Vibhu Sharma is a veteran of the semiconductor industry and a strategic technology advisor for GenInnov.
US issues new sanctions over Iran military procurement networks
The US Treasury Department is seen in Washington, DC, on August 20, 2026. [Mehmet Eser – Anadolu Agency]
The US Treasury Department on Tuesday sanctioned 10 individuals and entities in multiple jurisdictions that it said procured weapons and weapons components for Iran’s Ministry of Defense and Armed Forces Logistics (MODAFL), Anadolu reports.
“Treasury will not tolerate any support to the regime and will continue to identify, expose, and isolate Iran’s enablers,” Treasury Secretary Scott Bessent said.
The department said MODAFL is responsible for weapons research, production and acquisition for Iran’s armed forces and oversees organizations developing Iran’s ballistic missiles and unmanned aerial vehicles (UAVs).
Among those designated is Seyyed Asghar Alizadeh Tabatabai, whom Treasury described as a MODAFL representative in Beijing who “coordinates the procurement of finished weapons systems and dual-use components in China on Iran’s behalf.”
Trump ads paid for by US government allegedly violate anti-propaganda laws
Broadcasters should be ordered to stop running Trump commercials that appear to violate US laws, consumer advocacy group Public Citizen said today in a complaint to the Federal Communications Commission and Federal Trade Commission. The ads appear to violate laws against using taxpayer funds for propaganda and a law that restricts political activity by federal government employees, the group said.
“The most recent of these taxpayer-funded campaign ads, entitled ‘The Final Battle,’ is a copied version of an ad that the Trump campaign committee issued in the 2024 election,” Public Citizen wrote. “In dead serious black-and-white footage, Trump is seen marching down a corridor with a stern look, while the audio blares: ‘This is the final battle. With you at my side, we will demolish the deep state. We will expel the warmongers from our government.’ A notice at the bottom of the screen announces: ‘Paid for by the US government.’”
Trump also says in the ad, “We will cast out the communists, Marxists, and fascists. We will throw off the sick political class that hates our country, we will rout the fake news media, and we will liberate America from these villains once and for all.” It uses the same video and audio from an ad Trump used in 2024, with the main difference being that the 2024 version was paid for by Trump’s campaign committee instead of the US government. The White House claims its recent ads are public service announcements.
The updated version of the “Final Battle” ad aired during NBC’s Saturday Night Live, Fox News Sunday, and several college and NFL football games this past weekend, the Public Citizen letter said. Department of Homeland Security funds were used to pay for the ads, according to a Wall Street Journal report that said, “President Trump personally pushed for the ads and has been involved in their creation.” The other ads include one in which Trump says, “America will never be a communist country,” and another in which he says, “the United States of America is the most successful, most accomplished, most exceptional nation ever to exist in human history.”
Public Citizen, the nonprofit founded by Ralph Nader in 1971, said it “strongly urges the Federal Communications Commission and Federal Trade Commission to direct broadcasters to stop broadcasting these apparently illegal campaign advertisements.” We contacted the FCC and FTC about the Public Citizen letter today and will update this article if either provides any comment.
Broadcasters likely wary of FCC threats
It’s highly unlikely that the FCC or FTC would order broadcasters to stop running the Trump ads. FCC Chairman Brendan Carr’s history of attacking broadcasters that Trump dislikes suggests he is more likely to threaten the licenses of broadcasters that refuse to run the ads. Both Carr and FTC Chairman Andrew Ferguson also backed Trump’s expansion of power over nominally independent agencies like the FTC and FCC.
Senate Democrats said in a letter to Homeland Security Secretary Markwayne Mullin today that Trump is “illegally using Americans’ tax dollars to fund government propaganda and to put ads from his 2024 campaign back on the air. So far, it appears DHS has dedicated $20 million to this outrageous scheme, tapping funds provided to US Customs and Border Protection in Republicans’ ‘One Big Beautiful Bill Act’ for commemorative events relating to border security.”
The senators’ letter cited an estimate from analytics firm AdImpact “that over $1.7 million in federal funds have already been spent simply airing these ads.”
Broadcasters would likely be wary of rejecting the Trump ads. ABC owner Disney, which is fighting an FCC proceeding that could block renewal of its station broadcast licenses, said in a lawsuit that its editorial decision-making has been affected by FCC threats.
ABC’s The View hasn’t aired any interviews of political candidates since the FCC began the proceeding, and Jimmy Kimmel said he put an interview with a Democratic candidate on YouTube instead of broadcast TV so that local ABC stations wouldn’t have to deal with fallout from the FCC. The FCC historically exempted interviews on daytime and late-night talk shows from the equal-time rule, but the Trump FCC has upended that long-standing policy.
Equal-time rule doesn’t apply, group says
The equal-time rule prohibits broadcast stations from censoring or rejecting political ads paid for and sponsored by legally qualified candidates. However, Public Citizen said the rule doesn’t apply because Trump is not a candidate and that broadcasters should avoid airing ads that violate federal law:
Broadcasters may be held liable for airing any advertisement that is in violation of federal law. Federal law explicitly prohibits using appropriated public funds for government propaganda, self-aggrandizement, or partisan political advertising. Yet, that is exactly the nature of Trump’s recent campaign ads that began airing on September 23, 2026. Furthermore, Donald Trump is not a candidate for any elective office. The FCC and broadcasters are not constrained by the special rule protecting candidate advertisements.
Defending the use of government funds, the White House claimed the ads are “public service announcements.” The ads are “about reminding Americans to love their country and understand what makes it worth defending, at home, at our borders, and abroad,” the White House told media outlets.
But the ads do not qualify as public service announcements, Public Citizen said. “They do not refer to any specific governmental program and they do not serve the purpose of informing the public about any specific governmental program,” the group said. “Instead, the ads are designed to boost the image and character of Donald Trump right before the 2026 midterm elections, and by association to benefit the president’s political party in the upcoming election.”
Public Citizen separately filed a complaint with the US Government Accountability Office (GAO) and Office of Special Counsel (OSC). It alleged violations of “federal laws prohibiting the use of government expenditures for propaganda purposes (5 U.S.C. §3107 and related appropriations laws) and the Hatch Act.”
The Hatch Act prohibits federal employees from engaging in some kinds of political activities. It does not apply to the president, but Public Citizen said “the ad was conceived, produced and distributed by White House staff using government resources for partisan purposes.”
“You cannot use taxpayer dollars for political ads”
When contacted by Ars, the GAO did not offer any comment on the Public Citizen complaint. “Our work is done at the request of congressional committees or subcommittees or is statutorily required by public laws, per our Congressional Protocols,” the agency said. We also contacted the OSC and will update this article if we get a response.
US Rep. Jamie Raskin (D-Md.) last week said the Trump ads “may be felony criminal theft and conversion of government property for political campaign purposes,” and “a Hatch Act violation for any government employees who worked on it or used government resources to make it. It is blatantly unethical under numerous federal ethics provisions proscribing the use of government resources for political campaigns.”
Four Democratic Senate and House lawmakers told the White House in a letter, “The law is not complicated. You cannot use taxpayer dollars for political advertisements.” Senate Majority Leader John Thune (R-S.D.) reportedly said that although he likes the Trump ad’s message, “it shouldn’t be paid for with taxpayer dollars.”
Sen. John Kennedy (R-La.) similarly said he opposes using government money for Trump ads. He has also criticized Kristi Noem, the former secretary of Homeland Security, for a $220 million ad campaign.
“I don’t think any public official, including President Trump or Kristi Noem or John Kennedy, should spend public money on private ads for themselves,” Kennedy said in a CBS interview. “Now, the White House sees it differently, but I don’t… you can’t spend public money to promote yourself.”
Woman Faces Execution After Tennessee Governor Rejects Last-Minute Clemency Bid
Tennessee death row inmate Christa Pike is scheduled to be executed Wednesday after Gov. Bill Lee rejected her request for clemency, leaving the U.S. Supreme Court as one of the final possible avenues for stopping her execution.
Pike, 50, is scheduled to be executed on September 30 at Riverbend Maximum Security Institution in Nashville. If the execution goes forward, she would become the first woman put to death by Tennessee in more than 200 years.
The execution is scheduled for 10 a.m. Central Time.
Pike has spent roughly three decades on death row for the 1995 murder of 19-year-old Colleen Slemmer, a fellow student at the Knoxville Job Corps.
Lee announced Monday that he would not intervene after reviewing Pike’s request to have her death sentence commuted to life in prison without the possibility of parole.
The governor said he had deliberately considered the clemency request and conducted a thorough review of the case before deciding to allow the sentence imposed by the state to stand.
The decision was a major setback for Pike’s attorneys, who have spent the weeks leading up to her scheduled execution arguing that her age at the time of the murder, history of severe childhood abuse and later mental-health diagnoses warranted mercy.
But even after Lee’s decision, Pike’s legal battle was not completely over.
Her attorneys filed a petition with the U.S. Supreme Court on September 25 along with an application seeking a stay of execution. That request followed a September 23 decision by the Tennessee Supreme Court denying another attempt to stop the execution.
Unless a court intervenes or another extraordinary action occurs, Tennessee is continuing preparations for Wednesday morning’s execution.
A Murder That Shocked Tennessee
Pike was 18 years old when Slemmer was killed on January 12, 1995.
According to court records, Pike, her boyfriend Tadaryl Shipp and fellow Job Corps student Shadolla Peterson left the Knoxville Job Corps campus with Slemmer that evening.
Prosecutors said Pike had become angry with Slemmer and believed the teenager was interested in Shipp.
The group traveled to an isolated area near the University of Tennessee’s agricultural campus, where Slemmer was subjected to a prolonged and violent attack before she was killed.
The circumstances surrounding the murder were particularly disturbing and became a major part of prosecutors’ case when they sought the death penalty.
A Tennessee jury ultimately convicted Pike of premeditated first-degree murder and conspiracy to commit first-degree murder in 1996.
During the penalty phase, jurors concluded that aggravating factors justified a death sentence, including their finding that the killing involved torture or serious physical abuse beyond what was necessary to cause death.
The Tennessee Supreme Court later upheld Pike’s convictions and death sentence.
The other defendants received different sentences.
Shipp, who was 17 when Slemmer was killed and therefore could not receive the death penalty, was convicted of first-degree murder and conspiracy and received a life sentence. Peterson pleaded guilty to being an accessory after the fact and received a probationary sentence.
Clemency Appeal Focused on Pike’s Childhood
Pike’s attorneys submitted a lengthy clemency petition asking Lee to reduce her punishment to life without parole.
The defense did not dispute Pike’s responsibility for Slemmer’s murder.
Instead, attorneys argued that circumstances surrounding Pike’s childhood, mental health and age should be considered when deciding whether the state should carry out her execution more than 30 years after the crime.
The petition described an upbringing marked by sexual assault, abuse and neglect. Her attorneys have also said Pike was later diagnosed with bipolar disorder and post-traumatic stress disorder.
They argued that modern research into adolescent brain development provides a different understanding of the judgment and impulse control of an 18-year-old than was widely available when Pike went on trial in the 1990s.
Her lawyers also emphasized how unusual her case has become.
Pike is the only woman currently on Tennessee’s death row and would be the first person executed by the state in its modern death penalty era for a crime committed at age 18, 19 or 20.
Her attorneys have also pointed to what they view as a major disparity between Pike’s death sentence and the sentences received by the other young people connected to Slemmer’s murder.
Pike Expresses Remorse Decades Later
In material submitted as part of her clemency effort, Pike described the murder as the catastrophic act of a reckless and mentally ill teenager.
She said it took years for her to fully understand the enormous consequences of what she had done and acknowledged taking someone’s daughter, sister and friend.
Her legal team says Pike has changed substantially during her decades in prison.
Those arguments, however, did not persuade Lee to commute the sentence.
Victim’s Mother Has Waited More Than 30 Years
While much of the recent attention surrounding the case has focused on Pike’s childhood and mental health, Slemmer’s family has lived with the consequences of the murder for more than three decades.
Slemmer was just 19 when she was killed.
Her mother, May Slemmer Martinez, has publicly supported Pike’s execution and has spoken about the lifelong devastation caused by the loss of her daughter.
That divide has become central to the final days of the case.
Pike’s attorneys and anti-death-penalty advocates argue that her youth, background and mental-health history justify commuting her sentence.
Slemmer’s family, meanwhile, has waited more than 30 years for the punishment imposed by the jury to be carried out.
What Happens Next
Tennessee correction officials have been preparing for Pike’s September 30 execution at Riverbend Maximum Security Institution.
The state has already selected media witnesses and established security and reporting procedures surrounding the prison.
Pike is expected to be executed by lethal injection if no court intervenes.
The most significant remaining legal development is her U.S. Supreme Court case. Her attorneys have asked the justices to halt the execution while the court considers their petition.
That means the case could remain unsettled until very close to the scheduled execution time.
For now, however, Tennessee’s plans remain in place.
Barring a last-minute stay, Christa Pike is scheduled to be executed Wednesday morning — more than 31 years after the murder of Colleen Slemmer and roughly three decades after a Tennessee jury sentenced her to death.
Anthropic warns AI may pose ‘existential risks to humanity’
Anthropic plans to caution potential investors in its IPO that advanced AI could pose “catastrophic or existential risks to humanity,” an extraordinary warning by a company seeking to profit from the same technology.
The company’s IPO prospectus, reviewed by Reuters, highlights risks associated with its AI models, which it said could exhibit “self-preserving behaviors,” including attempts to “resist shutdown,” to “conceal or manipulate information” and behavior “resembling blackmail.”
“Our development of highly advanced models, platforms, and applications and expansion of use cases could further increase the risk that our models cause harm,” Anthropic said in the filing.
While public companies routinely outline product risks to investors, few, if any, have issued warnings suggesting their technology could cause potential human extinction. Anthropic emphasized both the transformative potential of AI on par with industrialization and electricity and the irreversible harm it could cause if mishandled.
Anthropic and other AI developers, including OpenAI, have faced scrutiny after incidents where experimental systems defied constraints, including a report of an OpenAI model breaching Australia’s health-system database.
Anthropic safety researcher Evan Hubinger estimated a greater than 10% probability that AI could kill humans within the next decade, echoing a sentiment by a former colleague, Jacob Coxon.
RISK-HEAVY DISCLOSURES
The company, which has positioned itself as a safety-first AI lab, devoted roughly 80 pages of the 261-page main body of its prospectus to laying out risk factors, nearly twice the 48 pages it used to describe its business.
For comparison, SpaceX, which owns xAI, dedicated just around 38 of the 277-page main body of its prospectus to risk factors.
“Potential model awareness of our evaluation efforts creates a significant limitation on our ability to assess model safety,” Anthropic said in the prospectus, adding that models sometimes develop unexpected capabilities during training that may not be discovered until they have been deployed and have resulted in significant safety incidents.
AI researchers have also warned that as models grow more capable, they increasingly recognize when they are being watched and adjust their behavior accordingly, which makes it harder to monitor model behavior.
Anthropic declined to comment in response to a request for comment on Monday.
UNCERTAIN RETURNS ON SAFETY INVESTMENT
Despite emphasizing AI safety, Anthropic said that returns on its safety investments are unclear.
It did not disclose in the filing how much the company was spending on such research. Earlier this month, Anthropic said about 6% of the computing power it used for AI research went to safety work in a sample week in July.
The company, creator of Claude AI models, described safety efforts as “resource-intensive” and said it must divide its limited funds between computing power, expensive AI talent and safety.
Anthropic said that its customer usage, and as a result revenue, is driven by new models and that a “continuous and overlapping cadence” of releases is “inherent to remaining at the frontier of AI development.”
The company last week released a new version of its Opus model, 10 days after CEO Dario Amodei published a nearly 4,000-word essay calling for pacing the frontier.
Some analysts and experts have said no leading AI lab would slow down when doing so risks handing rivals an advantage in an industry where valuations can change with each release.
Anthropic has pledged in recent weeks to disclose more data publicly about how it uses AI models to build future generations of the technology, as experts warn about recursive self-improvement — the point at which models can develop on their own without human help.
“We believe building reliable, trustworthy, and secure AI systems is a collective responsibility and that the market will reward it,” Anthropic said in the filing.
Inside the Business of Trump’s Third-Country Deportations
Tyler McBrien is the managing editor of Lawfare.
More than 25,000 people have had their lives upended and been forced into unfamiliar cultures, governments, and legal systems when the Department of Homeland Security deported them to so-called “third countries,” places where immigrants previously living in the United States have no connections and few rights — and where a deportation business booming under the second Trump administration finds its murky endpoints.
“The point is to scare people,” said one Trump administration official, according to a Senate Foreign Relations Committee minority report. “With countries like Palau or Eswatini, the point is that the Administration can threaten people that they will literally be dropped in the middle of nowhere.”
As is the case for many tools of its cruelty, the Trump administration did not invent third-country removals — it refashioned them to serve its own excesses. In the past, the U.S. used the obscure administrative compromise for people who could prove they would face persecution or harm if returned to their home country yet did not meet other requirements for asylum. The U.S. had to meet humanitarian and legal obligations, such as ensuring third countries would not turn around and deport migrants right back to their home countries. It was a last-resort tool that still left people displaced from their homes and communities in service of the country’s arcane immigration laws. But the current Trump administration’s policy, premised on agreements with at least 35 countries, has been anything but humane.
Nor has it been particularly legal. After the 1st U.S. Circuit Court of Appeals struck down the third-country deportation scheme, DHS scheduled a flight to deport migrants involved in a class-action lawsuit challenging the policy to Burundi, Rwanda, and the Central African Republic. Despite a successful emergency motion their lawyers filed last week, an Immigration and Customs Enforcement jet flew to two African countries that hold third-country removal agreements, according to the American Prospect. It is not yet clear who was on the plane or whether ICE carried out third-country removals in violation of a federal court order, and the Trump administration is now aiming to take the legal fight to the Supreme Court.
The saga illustrates the Trump administration’s proclivity for obfuscation, secret contracts, nondisclosure, and dealmaking of the backroom variety. Whether digitally disappearing people from ICE’s online detainee locator tool last week, or inking multimillion-dollar sole-source contracts with inexperienced companies of questionable provenance, the Trump administration has pursued its immigration goals under the cover of darkness, often justifying the secrecy with invocations of national security, all while attempting to shield itself from public scrutiny and accountability in the process.
“Deportation, Inc.: The Rise of the Immigration Enforcement Economy,” an ongoing investigative video series from Lawfare and SITU Research, in partnership with The Intercept, examines how U.S. immigration enforcement has evolved into a multibillion-dollar industry shaped by private profit and political power — where contracts, capital flows, and institutional incentives increasingly govern detention, deportation, and surveillance. The project joins other efforts to shed light on the often invisible economy of immigration enforcement, including the Berkeley Human Rights Center, Human Rights First, Third Country Deportation Watch, and others.
The first installment, released last December, focuses on the business of migrant detention and profit motive underpinning the brutality of privately run ICE detention facilities. The second chapter, out today, looks at the network of flights shuffling migrants between those facilities, as well as the shady deals brokered with the governments of deportation destinations.
Many details of these bilateral third-country removal agreements are not public, but the logic is transactional: Countries accept asylum-seekers and other immigrants from the United States in exchange for aid or other forms of financial compensation. Many of the countries involved in what The Intercept’s Nick Turse has called Trump’s “global gulag” are so dangerous that the State Department includes them on its “Level 4 Do Not Travel” list, “the highest advisory level due to life-threatening risks. And the very fact that these people are sent to a third country, rather than repatriated, means they have at least a viable asylum case.
Some but not all of the underlying texts of these agreements have come to light, often popping up in far-flung corners of the internet. This is not only a bad practice in government transparency, but a potentially illegal one as well. Lawfare, where I serve as managing editor, has argued as much in a federal district court in Washington, D.C., where the publication has sued the State Department over its failure to disclose international agreements into which the Trump administration has entered — in violation of the Case-Zablocki Act.
Though accelerated by the Trump administration, the secrecy that has enabled the growth of the vast, invisible economy of immigration enforcement traces back to at least 2001 with the creation of the Department of Homeland Security. Following the September 11 attacks, the new Cabinet-level department, along with other federal reforms, for the first time linked immigration with national security, which resulted in massive budget increases, new abilities to shield immigration information from the public, and general judicial deference to the executive branch that accompanies matters of foreign policy and national security.
Even the current Trump administration’s third-country deportation apparatus is beginning to build up its own record of cruel precedent. In the same class-action case last year, the Department of Homeland Security attempted to deport class members to Libya and South Sudan, in apparent violation of a temporary restraining order granted by a federal judge only weeks prior. During a hastily arranged hearing, a federal judge asked a Justice Department lawyer about the plane’s whereabouts, to which the government attorney responded, “I’m told that that information is classified, and I am told that the final destination is also classified.” Asked under what authority the government had classified the location, the Justice Department lawyer did not have an answer.
From the start, the Trump administration conceived of its third-country removal policy as another tool to inflict wanton abuse on irredeemable “illegal aliens.” As Secretary of State Marco Rubio explained last year during a Cabinet meeting, “We are working with other countries to say ‘We want to send you some of the most despicable human beings to your countries.’” What he failed to mention is that the people he’s demonizing have human rights, and that the countries he negotiated deals with have extensive records of violating them.
For the first time, “Deportation, Inc.” aims to render visible the dark web of agreements, transactions, and private contractors that make these grave abuses possible.
Correction: September 29, 2026, 12:50 p.m. ET This story has been updated to correct the premiere date of the first installment of Deportation, Inc.
Ethiopia is sliding back toward war – a look at what is driving the fighting and why it matters
Four years after a fragile ceasefire ended widespread conflict in Ethiopia, the region is again slipping back into full-out war.
On Sept. 23, 2026, advancing soldiers of the Tigray People’s Liberation Front (TPLF) seized government positions in northern Ethiopia.
It followed the announcement of a new alliance of seven armed movements in opposition to the Ethiopian government and deteriorating relations between the country and neighboring Eritrea, which it accuses of supporting the new alliance.
The Sept. 23 ground offensive by the TPLF drew global attention to the situation in Ethiopia. But in reality, the ratcheting up of tensions between the federal government in Addis Ababa and the TPLF has been going on for months.
Throughout August and September, the region of Tigray – Ethiopia’s northernmost constituent state, with an estimated population of roughly 7 million – has been subjected to federal air and drone attacks. Tigrayan organizations and human rights groupsreported that schools, residential areas, a farmer training center and a truck carrying food aid were all hit during the air campaign.
The strikes occurred amid mounting tensions over the TPLF’s restoration of the regional political administration that preceded the 2020-2022 civil war, an incomplete disarmament of the group and renewed localized clashes. In short, the federal government accused Tigrayan forces of remobilizing; the apparent purpose of the federal attacks was to degrade Tigrayan military capability.
The federal government had also been squeezing the region economically, by withholding funds needed to pay Tigray’s civil servants and limiting access to banks and fuel, according to the TPLF.
The situation escalated further on Sept. 20 when seven armed organizations announced the formation of the Ethiopian Peoples’ Forces Alliance for Survival in opposition to the federal government.
Three days later, Tigrayan forces began a coordinated ground operation against federal positions at airports in the regional capital Mekele, along with two other Tigrayan cities, Axum and Shire. Fighting then spread along Tigray’s borders with the Amhara and Afar regions.
Federal forces mounted a counteroffensive, and on Sept. 28 pro-government forces claimed to have entered Alamata – an important transport and supply center south of Mekelle.
Who are the main actors?
Prime Minister Abiy Ahmed’s federal government and the Ethiopian National Defense Forces retain formidable advantages in the current conflict: a national professional army, combat aircraft and drones, and federal revenues.
Its main opponent in the north is the TPLF, which was the dominant organization in Ethiopia’s ruling coalition from 1991 until Abiy became prime minister in 2018. It remains the strongest organized political and military force in Tigray.
The most striking partnership in the new alliance is between the TPLF and the Fano movement in Amhara, which fought alongside the federal army during the 2020–2022 civil war.
Some of the alliance’s members have limited military capacity, and there remains serious disagreements over territory and political objectives.
Nevertheless, it has adopted a workable minimum program: removing Abiy’s government, establishing an inclusive transitional administration and undertaking constitutional reform.
For its part, Eritrea is the most important potential external actor. Eritrean forces fought alongside Ethiopia’s federal military during the previous war and were accused of atrocities against Tigrayan civilians. Addis Ababa now accuses Eritrea and Sudan of backing the new alliance. Both countries deny the allegation.
What is the wider background to the tensions?
The conflict is rooted in an unresolved constitutional struggle over federalism, regional autonomy, territory and the distribution of political power.
Relations deteriorated further when Tigray held a regional election in 2020 after the federal government postponed national and regional elections. War began later that year.
Over the next two years, as many as 600,000 people may have died from violence, hunger and the collapse of health services, while more than 2 million were displaced.
The 2022 ceasefire agreement halted large-scale combat, but its political and security provisions were never fully implemented.
The TPLF began handing over heavy weapons in line with the accord, but further disarmament never happened as a result of a failure of Eritrean and other nonfederal forces to withdraw from parts of Tigray.
The agreement also required Tigray’s representation in federal institutions, including both houses of parliament. An inclusive interim regional administration was to be established until regional and federal elections could be held.
In 2025, the TPLF’s legal registration as a party about to compete in elections was revoked, and Tigray was excluded from Ethiopia’s June 2026 national vote. Tigray’s reintegration into Ethiopia’s federal institutions likewise remained incomplete.
So in part, the seeds to the conflict now were sown in the 2022 ceasefire. What was supposed to open the way to a comprehensive peace agreement instead hardened into a new and increasingly unstable status quo.
People walk in Mekele, the capital of Ethiopia’s Tigray region, on Sept 24, 2026, as fighting escalates with the central government.AP Photo
How could this affect the wider region?
The first danger is a direct confrontation between Ethiopia and Eritrea.
Such rhetoric has contributed to the conditions under which Eritrea might decide to support Ethiopian armed groups now.
A direct Ethiopian-Eritrean war would transform an internal Ethiopian conflict into an interstate war along the Red Sea’s western hinterland, with potentially devastating effects. A war between the two neighbors between 1998 and 2000 killed tens of thousands and displaced as many as 650,000 people on both sides of the Eritrea-Ethiopia border.
The second danger is convergence with the civil war in neighboring Sudan. Ethiopia has accused Sudanese forces of supporting the TPLF and other Ethiopian armed groups; Sudan has denied the charge.
The risk is that Ethiopia’s conflict and Sudan’s war become mutually enabling. For example, Eritrea could support the armed alliance to deter the Ethiopian government from pursuing access to the sea through the Eritrean port of Assab. Sudan’s army could support the same groups to pressure Ethiopia to end its alleged backing of the Rapid Support Forces. Such involvement could provide armed organizations with cross-border bases, weapons and supply routes, widening both conflicts.
Egypt and Gulf Arab countries already involved in Sudan could likewise be drawn more deeply into the Ethiopian conflict.
The third danger concerns Ethiopia’s trade corridor through Djibouti. Ethiopia depends on Djibouti for the overwhelming majority of its maritime commerce, including fuel imports. If fighting in Ethiopia’s Afar region threatens the highway and railway linking Addis Ababa to Djibouti, the conflict would have immediate national and international economic consequences.
For the United States, which has blamed the TPLF for the escalation, the stakes extend beyond humanitarian concerns. Ethiopia is one of Africa’s most populous countries, hosts the African Union and has long been an important U.S. security partner in the Horn.
Can both sides return to the ceasefire?
The agreement’s political framework has been overtaken by events, and Ethiopia’s conflicts are no longer confined to the relationship between Addis Ababa and Mekelle.
Preventing a wider war will, I believe, require an immediate halt to airstrikes and offensive ground operations, restoration of communications and humanitarian access, and independent investigations into attacks on civilians. It will also require negotiations toward the comprehensive political agreement that was supposed to follow the 2022 Pretoria agreement.
Those negotiations can no longer be limited to the federal government and the TPLF. They must address the wider alliance, other political constituencies, contested territories, regional representation and the future of Ethiopia’s federal system.
The question is no longer whether Addis Ababa and Mekelle can restore their former bilateral arrangement. It is whether Ethiopia can establish an inclusive constitutional process before its previously separate conflicts converge into a national and potentially regional war.
NASA has a Dragon dilemma, and there appear to be no good answers
For two decades, largely in service to the International Space Station, NASA has sought to foster an “economy” in low-Earth orbit.
Twenty years ago, with a program to develop private spacecraft for cargo delivery to the space station, NASA sought to “stimulate efforts within the private sector to develop and operate safe, reliable, and cost-effective commercial space transportation systems.” In recent years this has expanded to creating an entire commercial ecosystem in orbit, with transportation, space stations, manufacturing, tourism, and more, such that NASA is one of many customers in the market.
In April 2024, the space agency explicitly laid out its philosophy: “NASA supports a robust commercial space economy that advances American industry and promotes technological discovery through in-space work and research. NASA remains committed to fostering innovation and collaboration within the American space industry.”
But just two years later, there are growing questions about the viability of this. As the second space race heats up, NASA has become more interested in focusing on the lunar surface, with a robust Moon base. SpaceX has signaled it no longer wants to be in the business of flying astronauts into low-Earth orbit. Today, the grand plans for a low-Earth orbit economy, at least involving humans, appear to be going sideways.
So what happened, and why does it matter? Ars spoke with a number of industry sources, on background, to provide some answers.
Q. What precipitated this crisis?
A. In recent months SpaceX has made it clear to NASA that it no longer wishes to fly its Crew Dragon spacecraft, or the Falcon 9 rocket, on missions to low-Earth orbit. The company has agreed to support the International Space Station until 2030. But after that SpaceX intends to retire the spacecraft. SpaceX has told companies developing private space stations for low-Earth orbit, including Axiom Space, Voyager Space, and Vast Space, that they cannot order Crew Dragon missions for their habitats.
Q. Can NASA compel SpaceX to keep flying Dragon?
A. NASA invested $3.1 billion in the development and certification of Crew Dragon as part of the Commercial Crew Program. But SpaceX was only compelled to fly half a dozen missions. It has flown 13 missions for NASA to the space station, and will launch another one in a few days. The company had recently agreed to keep flying through the Crew-17 mission. SpaceX has therefore more than fulfilled its contract obligations to NASA.
Q. But isn’t NASA a really important customer for SpaceX?
A. It was in the past, yes. But SpaceX now derives a majority of its revenue from Starlink, and that proportion is likely to grow even more. Additionally, as part of the process of going public earlier this year, in financial filings, SpaceX made clear that it envisions a vast majority of its future revenue will come from Starlink and orbital data centers. The category of “space enabled solutions,” of which NASA is a fraction, represented approximately 1 percent of what SpaceX views as its “total addressable market.” In other words, NASA needs SpaceX more than SpaceX needs NASA. Going forward SpaceX wants to focus on launching its own payloads—on the Starship rocket. NASA Administrator Jared Isaacman recognized this reality during a news conference on Monday, saying, “I do not think it’s a secret that SpaceX intends to sunset older platforms like Falcon and Dragon as they concentrate on their next-generation capability, Starship.”
Q. What about Starship?
A. Four astronauts currently launch on Dragon. Starship could potentially bring dozens of astronauts into orbit at a time. That would be revolutionary for access to low-Earth orbit and an economy there. However SpaceX has told NASA it is not interested in developing Starship for human launches into Earth orbit at this time. (Again, they’re focused on their own payloads). Ascent and entry of Starship, carrying humans, would raise a tangle of safety and regulatory concerns and is just not a priority for the time being. NASA has no real way to compel SpaceX, and any political capital the space agency might expend on Starship is going to be focused on getting a variant of the vehicle for a “Human Landing System” as part of the Artemis Moon program rather than human launches from Earth.
Q. What’s happening with Boeing?
A. Boeing was NASA’s other partner in the Commercial Crew program. The agency invested $5.1 billion, to date, in Boeing to develop the Starliner spacecraft. For all of this, Boeing has yet to fly a single operational mission to the space station. The news this week is that, despite these struggles, NASA will invest $359 million more to support the company’s efforts to fix Starliner’s propulsion system and certify the Vulcan rocket for new missions. It is NASA’s hope that Starliner can supplement astronaut missions during the remainder of the International Space Station’s lifetime, and then be available for private space station operators.
Q. Is this a good plan?
A. A lot of people don’t like it. Some critics say NASA has basically handed Boeing (not a particularly benevolent monopolist) and Starliner a monopoly on Western human spaceflight to low-Earth orbit for the next 10 or 20 years. This may effectively end any hope of a low-Earth orbit economy that involves humans in space. However, others say NASA faced few good choices. And given NASA’s extraordinary investments in Boeing to date, it would have been fiscally irresponsible to abandon Starliner now. NASA funded two companies as part of the Commercial Crew program. If one of them is walking away, it makes sense to support the remaining one, even if there are legitimate concerns about Boeing’s past performance.
Q. What else might NASA have done?
A. Some people wanted to see NASA fund a new competition, a Commercial Crew 2.0 for the 2030s. This would have brought on a competitor, probably Blue Origin but maybe also someone like Sierra Nevada or The Exploration Company, to keep price pressure on Boeing for crew transportation services. However, a new competition would ultimately have cost NASA billions of dollars, and Isaacman seems reluctant to make such an investment given all of NASA’s other priorities. Isaacman believes Boeing can meet NASA’s needs, which are something like two seats every six to nine months, to orbit. The real unknown is whether a market beyond NASA—institutional customers from Europe, the Middle East, and beyond, in addition to privately funded astronauts—could exist at Starliner’s prices.
Q. How much does a seat cost?
A. This is an important question. SpaceX’s original price-per-seat for early Dragon flights was approximately $55 million. For more recent missions, the price has increased to $78.8 million. (And if SpaceX were to magically decide to keep flying Dragon longer, the price would only go up). By contrast, the Starliner price to NASA is $90 million per seat during the International Space Station era. So what happens after Dragon retires? Let’s just say no one expects prices to go down. I asked Boeing Vice President John Mulholland about Starliner seat prices in the 2030s yesterday, and he replied, in part, “Obviously we want to be as competitive as possible.” But competitive with whom?
Q. What about Blue Origin?
A. The space company founded by Jeff Bezos is developing a “Space Vehicle” for astronauts to launch on the New Glenn rocket. After some of my recent reporting, company officials reached out to let me know that design work is “well advanced” along with demonstration work such as cabin pressure-vessel manufacturing, extensive parachute testing, in-house thermal protection system testing, life support systems, and more. I’ve heard “no earlier than” dates of 2031 for a crew launch. But that’s probably optimistic, and if NASA and private space station operators need to book transport in the early 2030s, Starliner is probably the only option.
Russian Soyuz-FG rocket with the Soyuz TMA-12M spacecraft launches in March 2014.
Credit: VASILY MAXIMOV/AFP via Getty Images
Russian Soyuz-FG rocket with the Soyuz TMA-12M spacecraft launches in March 2014. Credit: VASILY MAXIMOV/AFP via Getty Images
Q. What other vehicles are out there?
A. NASA relied on Russian Soyuz vehicles in the 2010s after the Space Shuttle retired, and before Crew Dragon came online. With Russia’s invasion of Ukraine, Soyuz is off the table for private space stations. India is also developing a crewed spacecraft, Gaganyaan. But it was originally supposed to carry humans in late 2021, and the schedule has since slipped to at least 2027. And for a time Gaganyaan is likely to be used solely for Indian missions. Counting on this vehicle for private space stations seems like a stretch. NASA does have its Orion spacecraft, but the per seat cost for its missions is likely to be astronomical ($500 million per seat?), and Orion is needed for lunar missions. The Exploration Company, based in Europe, has ambitious plans for a crewed spacecraft, but it likely won’t be ready until 2035. Sierra Nevada’s Dream Chaser just does not seem like it’s ever going to happen, sorry.
Q. So what’s the answer?
A. You’re probably not going to like this, but the only real hope for a significantly lower sticker price for sending humans into low-Earth orbit is Starship. If incentivized, SpaceX probably could bring this capability online by 2030 and radically reshape the market. But from all publicly available evidence, and based on private conversations, SpaceX seems unlikely to prioritize crewed ascent and reentry on Starship any time soon. Could that change? Certainly. Will it? Probably not. SpaceX and its founder, Elon Musk, will do what they want.
Q. So is SpaceX just being selfish, or what?
A. SpaceX is a business, and like a lot of other businesses, especially publicly traded ones, the goal is to maximize revenue. From their perspective, it makes sense to remove distractions (such as Dragon and Falcon 9) and focus on the future of the company (Starship).
One way of looking at the last 20 years of spaceflight history, and NASA’s efforts to stimulate a low-Earth orbit economy, is to view SpaceX as the exception to the rule. In some sense, an economy based on astronauts in low-Earth orbit got lucky that SpaceX executed so successfully on Dragon. This allowed for the creation of a market around the idea of access at a price of $50 million a seat. At the same time, transportation competitors in cargo (Northrop) and crew (Boeing) struggled mightily. The best SpaceX’s competitors could do was nearly twice the price, and even then, not as reliably.
NASA seems to think Starliner, even at higher prices, will provide the guaranteed access it needs to low-Earth orbit in the 2030s for its astronauts. But in terms of a broader space economy in low-Earth orbit—which for decades the space agency has explicitly sought to foster—it is difficult to see Starliner providing a suitable solution. So yes, SpaceX pulling out of this market harms the industry. But should it be incumbent upon SpaceX to continue a line of business solely because it benefits its peers and competitors?
Education and Labor Reform Move to Center of Israel’s Postwar Economic Debate
Economists see stronger human capital, wider workforce integration and productivity gains as central to sustaining long-term growth
Israel’s next finance minister will inherit an economy with a tight labor market, a resilient shekel and a technology sector that continued to operate through nearly three years of war. At the same time, the government will face roughly 350 billion shekels (around $114 billion) in estimated war-related fiscal costs and a public-debt ratio approaching 70% of gross domestic product.
The next government will therefore confront a different economic environment from the one Israel faced before Oct. 7, 2023. The economy has absorbed repeated military shocks without entering a financial crisis, but the fiscal room available to respond to future shocks has narrowed.
Longstanding structural problems also remain unresolved, including weak productivity in parts of the economy, educational gaps and low labor-force participation in some communities.
The Bank of Israel estimates the fiscal cost of the war from 2023 through 2026 at about 350 billion shekels, with roughly half financed through borrowing. By the end of 2025, public debt had risen to 68.5% of GDP from 60.5% at the beginning of 2023, while the government deficit stood at 4.7%.
The central bank separately estimated that the economy produced about 177 billion shekels less between the start of the war and the end of 2025 than it otherwise would have.
Prof. Elise S. Brezis, a macroeconomist at Bar-Ilan University, cautions against treating the rise in debt as Israel’s most urgent economic problem.
“When you look at that compared to many OECD countries, our situation is not so bad,” Brezis told The Media Line.
She pointed to the exchange rate and Israel’s continuing ability to finance itself as indications that markets do not currently view the country as facing an imminent fiscal crisis.
“It is a problem,” she said, “but that’s not the first thing that a new government should take care of.”
Recent Bank of Israel indicators offer some support for that assessment. In September, the bank said Israel’s risk premium was around levels seen before Oct. 7, 2023, while the labor market remained tight.
Economic activity also rebounded sharply during the second quarter of 2026, although the bank cautioned that part of the increase reflected recovery from disruption caused by earlier fighting with Iran.
Adrian Filut, chief economic commentator at the Israeli financial newspaper Calcalist, takes a more cautious view.
Filut told The Media Line that Israel’s next government, regardless of its composition, will enter office with “very little fiscal room.”
In his view, comparisons with heavily indebted large economies can be misleading because Israel faces unusually high geopolitical risk and therefore requires larger fiscal reserves.
Israel is not Switzerland
“Israel is not Switzerland,” Filut said. “You need buffers. You need space and cushions of protection that other governments perhaps do not need.”
For Filut, a debt ratio of about 70% is therefore more consequential for Israel than the headline figure alone might suggest. The country entered the war with fiscal space accumulated over previous years, he said, and has since used much of it.
Defense spending will make rebuilding that fiscal space more difficult.
Before the war, military spending was about 4.5% of GDP. The International Monetary Fund estimated that it reached roughly 8% in 2025 and could remain near 6% in 2026. Even if wartime spending declines, defense would still consume a larger share of the economy than it did before Oct. 7.
The Bank of Israel’s July forecast projected a 2026 deficit of 4.9% of GDP and public debt of about 69%. The projection already incorporated additional defense spending but did not account for every possible increase.
The bank said that adding as much as 25 billion shekels to the defense budget, bringing it to roughly 183 billion shekels, could push the deficit to about 5.5%.
That would leave policymakers with a limited range of options.
“If you can’t touch defense, you don’t have that much room left in civilian and social spending,” Filut said.
The alternatives, he said, include increasing government revenue, broadening the tax base, reducing sectoral spending, cutting other expenditures or borrowing more.
Higher value-added or income taxes would raise substantial revenue but could also weigh on consumption and growth. Filut said he would first examine taxes associated with negative externalities, including pollution and products that generate future public-health costs.
Brezis places greater emphasis on education and human capital.
For her, the more serious long-term risk is whether a growing share of Israelis receives the education needed to participate productively in a modern economy.
She challenges the common framing of the Haredi economic problem as simply one of workforce participation. Some employment, she said, may not be fully reflected in official statistics. The deeper issue, in her view, is whether workers have the skills needed to enter productive sectors of the economy.
“They don’t know English, they don’t know math,” Brezis said, arguing that publicly funded schools across Israeli society should provide the same core academic curriculum while allowing communities to add religious, cultural or other subjects.
Asked what she would prioritize during the first year of a new government, Brezis proposed ending budget allocations based on family size or sectoral affiliation, redirecting funding toward education reform and beginning a major effort to improve public transportation.
She estimated that changes to sector-based spending could save roughly 20 billion shekels, although that figure was her own estimate.
Filut reaches a similar conclusion on human capital, but from a more restrictive fiscal starting point.
Israel, he said, cannot address its demographic challenge simply by increasing the number of people in employment if many of those jobs remain low-skilled and low-paid.
The game changer is the productive and skilled labor market
“The game changer is the productive and skilled labor market,” he said.
Israel’s economic advantage, Filut argued, rests overwhelmingly on its people rather than natural resources. Maintaining that advantage requires workers who can enter technology, artificial intelligence, cybersecurity and other high-productivity industries.
That makes education and labor-market integration among Haredi and Arab Israelis an immediate economic and budgetary issue rather than one that can be deferred to future governments.
Brezis also sees high tech as the sector capable of sustaining Israel’s relative economic strength.
The locomotive in the 21st-century economy will be high tech
“The locomotive in the 21st-century economy will be high tech,” she said, arguing that Israel’s capacity for entrepreneurship and technological innovation remains intact despite the war.
The two economists differ less on the need for structural reform than on the order in which the next government should act.
Brezis believes Israel has enough underlying economic strength to focus first on education, institutional quality and the allocation of public resources.
Filut wants the next government to begin with a medium-term fiscal plan that puts the debt ratio back on a downward path, followed by education and labor-market reform and measures addressing Israel’s high cost of living.
For Filut, delaying a decision on how to finance permanently higher defense spending could carry substantial long-term costs.
If the money comes from civilian budgets, he said, education, universities, research and public health will face greater pressure. If it comes from additional borrowing, a larger share of future budgets will go toward servicing debt.
Interest is the worst payment
“Interest is the worst payment,” Filut said. “You pay billions of shekels and receive nothing. You are only financing the debt.”
![The US Treasury Department is seen in Washington, DC, on August 20, 2026. [Mehmet Eser - Anadolu Agency]](https://i0.wp.com/www.middleeastmonitor.com/wp-content/uploads/2026/08/AA-20260821-42264105-42264093-US_PUBLIC_DEBT_SURPASSES_40_TRILLION_FOR_FIRST_TIME.jpg?fit=920%2C613&ssl=1)
