TheChek News Flow

US and Iran exchange fire for the first time in more than a month

Oil prices rise after American military hits island in Strait of Hormuz and Tehran retaliates with attack on Jordan

The US military has struck two Iranian rocket launchers on an island in the Strait of Hormuz, the first American attack on Tehran’s forces in more than a month.

Iran swiftly retaliated by firing missiles at Jordan.

“Islamic Revolutionary Guard Corps forces were observed preparing to launch rockets with sea mines into [the] Strait of Hormuz”, prompting American forces to strike two launchers on Larak Island, said Captain Tim Hawkins, spokesperson for US Central Command, on Sunday. Centcom oversees Washington’s military operations in the Middle East.

Tehran lobbed “a few missiles” at Jordan in retaliation but they were intercepted by American forces, according to a person familiar with the matter.

The attacks caused oil prices to rise above $90 a barrel, as traders considered whether the resumption of hostilities posed a renewed risk to flows from the Middle East.

“Tonight’s renewed strikes on Iran not only underscore the fragility of de-escalation, but the adamancy of the IRGC to complicate free transit in the Strait of Hormuz,” said Michael Alfaro, chief investment officer of Gallo Partners, an energy and industrials-focused hedge fund.

“This situation feels like an endless powder keg which continues to embed a premium in prices,” he said.

The US attack appeared smaller and more focused than its last known strikes on the Islamic republic on July 29, when “dozens” of targets were hit in a two-hour wave of bombings.

Last month featured about three weeks of tit-for-tat strikes between the US and Iran that threatened a return to full-blown war, but failed to fully reopen the Strait of Hormuz or bring down the Islamic regime.

Since then, the US has shifted to prioritising economic warfare to pressure Tehran into capitulation. On Monday, US Treasury secretary Scott Bessent said Washington was expanding secondary sanctions on Iran, threatening measures that could “blow up” the global financial system.

However, Bessent has stopped short of the “economic D-Day” he had threatened against Iran. So far, Washington has only revoked US financial access to an Egyptian bank’s branches in the United Arab Emirates, and has not provided a deadline for Tehran’s trading partners to comply with its pressure campaign.

The limited first move from Bessent showed Washington was hesitant to go after bigger financial targets that help finance the Iranian regime, including major banks and state-backed entities in China, out of fear of further destabilising the global economy and financial markets.

Bessent told the Associated Press on Sunday that the US would announce an action against another bank in the coming days.

The six-month war between the US and Iran has become a stalemate, and talks between Washington and Tehran have collapsed.

President Donald Trump has struggled to end the conflict, which has stoked inflation and is unpopular with American voters ahead of crucial midterm elections in November.

Iran’s reformist President Masoud Pezeshkian, who left Tehran early on Monday to attend the Shanghai Cooperation Organisation security forum in Kyrgyzstan, said his country was “not seeking war”.

He added: “However, in the face of any aggression, we will never remain passive and we will give a firm response to the aggressors.”

The IRGC said it used missiles and drones to carry out a retaliatory strike on Sunday against technical and maintenance infrastructure as well as fighter jet positions at two US air bases in Jordan. They also claimed to have shot down an American MQ-9 drone over the Strait of Hormuz.

The group said the US will not be able to “weaken Iran’s control over the Strait of Hormuz through aggression” and threatened to “forcefully respond” to every strike. The IRGC also said the US strikes killed military personnel and civilians.

Challenger golf league is negotiating a funding package with BC Partners for a slimmed-down circuit

LIV Golf may file for bankruptcy protection as soon as the week of September 7, as the challenger golf league races to find funding in the coming days for a slimmed-down version slated for 2027, according to people briefed on the negotiations. 

The league in recent days has sent settlement offers to the current LIV players who are owed millions in guaranteed payouts past 2026, with those initial offers at just a few cents on the dollar. 

People with knowledge of the situation said the reluctance of the Saudi sovereign wealth fund, which launched the league in 2022, to provide more funding had played into the meagre offer to the players. The LIV management, meanwhile, remains dependent on the Saudi Public Investment Fund for financial support needed to facilitate the transition.

For several weeks, LIV has been negotiating a funding package with the credit division of private capital firm BC Partners and BC executives have attended recent LIV tournaments in Indiana and New Jersey to make their pitch to players.  

The FT previously reported on LIV’s attempt to strike a “grand bargain” whereby its golfers would simultaneously settle claims with the current LIV regime and agree to equity awards and other financial terms for the “LIV 2.0”. Such a deal could be clinched in a pre-packaged bankruptcy. 

However, getting commitments from enough top players to entice BC Partners and other funders has proved challenging so far. The DP World Tour, which runs events in Europe and the Middle East, has said recently that players on a future LIV circuit may face fines or bans for playing in league events scheduled in the same weeks as the DP World Tour.

The new LIV circuit envisions a global, 10-tournament slate that would probably require other tours to accommodate its players for the remainder of the season.

“We shouldn’t live in a world where one group of golfers is uniquely restricted from playing elsewhere,” said LIV CEO Scott O’Neil at the 2026 LIV finale earlier this month in Indiana.

LIV was launched by the Saudi Public Investment Fund as a challenger to the US PGA Tour. PIF earlier this year said it would end its funding at the conclusion of the current season, having burned through billions on player signing bonuses and prize money.

Persons close to the bankruptcy said that a Chapter 11 filing may happen without a LIV 2.0 being finalised and described three camps of players: those who settle with LIV and join the reimagined league, those who settle and do not join and those who choose to “fight it out” for their contracted money as unsecured creditors in the bankruptcy.

PIF is expected to fund a bankruptcy loan known as a “debtor-in-possession” financing of under $100mn, though it is adamant about committing no other new funds, according to the people.

“PIF wants a clean baton toss,” said one person involved in the conversations.

News reports in recent weeks have revealed that multiple vendors to LIV golf tournaments have sued over payments they said had not yet been made.

Fees for LIV professional advisers and new independent directors have soared into the tens of millions of dollars, said people familiar with the situation.

The bankruptcy filing may be made in the federal district court of New Jersey, a venue that is known to be friendly to debtors and has overseen such recent high-profile Chapter 11 cases as WeWork and Rite Aid. 

Trade publication Sportico reported this summer that LIV Golf had set up a new subsidiary in New Jersey. A person close to the situation said that the New Jersey entity had been established as a possible way to allow for a New Jersey-based Chapter 11 filing.

BC Partners, while conducting due diligence, is itself waiting for LIV and PIF to resolve player payments and structure the bankruptcy terms. Multiple people said that the firm was examining an equity-like investment that would preserve LIV’s “net operating losses”, which total more than $5bn across the US and the UK.

NOLs can be used to reduce taxable income of acquirers, subject to limitations set by the tax code. BC Partners may lead an investment of as much as $300mn to acquire the LIV assets and create a vehicle to “roll-up” other sports assets which can together utilise the NOL, according to a person familiar with BC’s thinking.

Liberty Strategic Capital, the private equity firm founded by former US Treasury secretary Steven Mnuchin, is among the funds to have expressed preliminary interest in backing a BC Partners transaction. 

LIV, PIF and BC Partners each declined to comment.

High-End Credit Cardholders Find It’s Harder to Get Their Money’s Worth

Rising annual fees have intensified the cat-and-mouse game over rewards between card companies and customers

Higher annual fees are forcing premium credit cardholders to work harder to get their money’s worth. Tightened rules around rewards and perks aren’t making it any easier.

American Express AXP -0.29%decrease; down pointing triangle and Capital One are among the card issuers cracking down on bringing guests into airport lounges. Perks like DoorDash DASH 2.09%increase; up pointing triangle credits are coming with fine print, including minimum spending requirements. And loopholes that let people get around rules to milk more from their rewards are being tightened.

“It’s just a forever changing landscape,” said Ravi Wadan, a 39-year-old content creator in Seattle who posts on social media about credit cards. “They’re adding some benefits here and there but in general it’s getting more and more restrictive.”

Amex Platinum cardholders, for instance, aren’t supposed to use their quarterly $75 Lululemon credit on gift cards. But Wadan had been getting around the rule by buying gift cards online. He recently noticed that trick no longer works.

Premium cards have long fueled a world of rewards enthusiasts who hunt for ways to game the system. Now with annual fees on some cards climbing toward four figures, the cat-and-mouse game between banks and customers over rewards is intensifying. For card issuers, it means walking a fine line between stopping people from milking perks in a way that wasn’t intended, while also being careful not to upset some of their most valuable customers.

Marietta Williams, a 42-year-old education consultant and literacy coach for teachers in New Jersey, was recently notified that her annual fee was going up to $895, from $695. Amex hiked the price for new cardholders last year, but many existing cardholders weren’t hit with the increase right away.

The increase made Williams decide the card was no longer worth it. One reason: rule changes around bringing guests into airport lounges.

“I feel like the benefits are kind of going away,” said Williams, who got the card in 2022.

In July, Amex implemented a new rule saying guests must be traveling on the same flight as the cardholder. Not long after she got the card, Amex had also started tacking on a $50 fee for adult guests who tagged along into its Centurion Lounges.

Earlier this year, Capital One also stopped allowing free guests in lounges for its Venture X cardholders. Unless cardholders spend at least $75,000 a year, bringing along an adult guest now costs $45.

Card issuers say the crackdown on guests is intended to maintain the quality of perks for all their customers. With premium cards becoming more popular, including among Gen Z customers, many cardholders have complained about overcrowded lounges.

Issuers are also putting an end to workarounds cardholders use to circumvent their rules.

Some American Express Platinum cardholders had found ways to apply a $200 airline credit intended only for incidentals like checked bags, Wi-Fi and in-flight meals to flights. Earlier this year, Amex put an end to a loophole where cardholders deposited the credit into a United program, which they would then use to purchase flights.

“By closing this loophole, it’s likely that a lot of people are not going to remember to use the credit before the end of the quarter,” said Jake Norcross, a 31-year-old who works for a marketing agency in Chicago and posts about credit card points online.

Lisa Kalhans, Amex’s executive vice president of U.S. consumer proprietary products, said in a statement that the company has invested in making benefits easier to discover, enroll in and track.

In some cases, changes to rewards stem from the businesses card issuers are working with. In July, Amex got rid of its $100 Saks Fifth Avenue credit after the company went bankrupt. Chase Sapphire Reserve’s monthly $10 DoorDash credit, offered twice a month, now requires a $20 minimum purchase after DoorDash changed its requirements.

“The points game in itself has gotten more difficult in the course of the last three years,” said Jared Friedman, who works for a consulting firm in New York and posts on social media about how to maximize points. “At the end of the day there’s still value to be had, it just requires a lot more time and attention.”

The escalating vigilance around any changes to rewards can also cause confusion.

In August, language on Chase’s website stated that Sapphire Reserve cardholders couldn’t use both of their $250 luxury hotel credits for back-to-back stays at the same property within 24 hours of checkout. Customers swiftly noticed and took to social media to complain.

Chase quickly clarified that there had been an error and that Sapphire Reserve cardholders could use both $250 statement credits at any time and for back-to-back stays.

In a statement, Chase said that Sapphire’s popularity is a testament to the investments it has made in its rewards.

Ukraine’s ex-defence minister touts ‘new history’ in war against Russia with planned tech fund

Mykhailo Fedorov seeks western capital to build and manage weapons development companies he says will help Kyiv resist Moscow

Ukraine’s former defence minister Mykhailo Fedorov is seeking US investment for a defence technology fund he hopes will write a “new history” of the war against Russia by building and managing companies to develop new weapons.

In an interview with the FT in Washington, Fedorov, who was sacked by President Volodymyr Zelenskyy last month, also said he would be happy to help the US “build an army of drones” in exchange for American supplies of Patriot missile interceptors to help Ukraine in coming months.

Fedorov is seeking backing for the proposed fund from venture capitalists, private investors and US tech companies and said he was in discussion with “a large number” of prospective investors interested in Ukrainian innovation.

“This isn’t exactly a classic venture fund,” Fedorov said. “We’re writing a new history of the war,” he said.

The vehicle would go beyond a conventional venture capital fund, being able to create companies itself as well as invest in existing Ukrainian start-ups, he said. It would use western capital to develop technologies critical to Ukraine’s ability to resist Russia’s ongoing invasion.

Fedorov did not give details of his role overseeing the fund or whether he might personally benefit from it financially.

He identified three initial priorities for the fund: battlefield robotics to reduce Kyiv’s reliance on manpower and the number of troops on the front line; cheap high-speed interceptor drones to combat Russia’s new jet-powered attack drones; and low-cost AI-enabled missiles.

It was likely Russia would drag out the conflict, and if it did for another two or three years, the fund would develop state of the art “solutions for the battlefield” while also giving investors insights into the lessons Ukraine had learnt, he said.

Fedorov, 35, was Ukraine’s youngest-ever defence minister until he was sacked after just six months in office. The ousting sparked weeks of protest across the country and thrust Zelenskyy into the deepest political crisis of his presidency.

Ukrainian MPs, veterans and civil society leaders — including many who had supported Fedorov after his dismissal — were dismayed at news of his trip abroad. Many questioned his motives when he announced on Friday after a visit to Rome that he would serve as an unpaid adviser to the Italian defence minister.

Oleksiy Goncharenko, an opposition MP who chairs a parliamentary investigative committee, said in a post on social media site X that Fedorov’s Italian role could pose “a threat of leakage of knowledge about our cutting-edge military technologies”.

But Fedorov said helping Italy’s military innovate could open new doors for co-operation with a country that had air-defence capabilities Kyiv lacked. He said he would seek similar advisory positions with other partner nations, possibly including France and the US.

Fedorov declined to comment on whether he would meet anyone from the Trump administration while in Washington, but said he would be “very glad to help America build an army of drones — especially in exchange for Patriot [interceptors] so that we survive this winter”.

“I’m ready to make that trade,” he said.

Fedorov, who previously served as a deputy prime minister in charge of digital transformation of the government, built a reputation as the driving force behind Ukraine’s drone development. His start-up-inspired management style set him apart from the country’s political establishment but also unsettled parts of the country’s traditional military leadership.

Fedorov said he had run into resistance from the military establishment and political rivals as he set out to revamp a murky procurement system.

Zelenskyy fired him over what the president said was a failure to address problems with Ukraine’s military mobilisation and the minister’s unreconcilable differences with the military commander-in-chief Oleksandr Syrsky.

Syrsky was sacked days later after the public protests erupted. Fedorov has since broken the longstanding taboo against publicly criticising state power to maintain a united wartime front and has called for wartime national elections. He has further sparked controversy by saying his country was losing the war.

Fedorov told the FT that “unity” would be key to Ukraine’s survival, shrugging off criticism that he himself has put it at risk by speaking out. He declined to say whether his call for new elections was in effect a demand for Zelenskyy to be replaced.

“This isn’t against Zelenskyy, it’s not against some specific person,” he said. “It’s about the fact that we need to . . . update the state’s operating system.”

Fedorov this week triggered outrage in Ukraine when he said the country was “gradually, slowly losing” the war — after insisting in spring that it had turned the tide against Moscow.

The situation could “change from month to month”, he told the FT, noting that Russia’s recent air strikes and ever-intensifying use of missiles and advanced drones were critical challenges.

“Right now, if we don’t respond to these issues, they will smash us.”

Russia has pummeled Ukraine’s critical infrastructure, factories and warehouses with missiles and swarms of attack drones this summer. A critical shortage of Patriot interceptors caused by shrinking global stockpiles of the US-made weapon has left Ukraine struggling to shoot down Russia’s ballistic missiles.

Fedorov argued that his fund would in time offer one answer to that problem. He said high-speed drones capable of bringing down Russia’s jet-powered drones had already been tested in Ukraine and, with more support, could be produced at scale.

He declined to comment on whether he planned to meet Elon Musk while in the US but said he remained in close contact with the tech billionaire.

Zelenskyy has sought permission from Musk to use Starlink-equipped drones against ballistic missile launchers up to 200km inside Russia, something Fedorov said would probably have a big impact.

People familiar with the matter said Musk had previously told Fedorov he would be open to allowing Ukraine to use Starlink-equipped drones beyond its border to hit Russian targets. However, Musk, the people said, had told Ukraine that the decision was a political one that needed to be made in the White House.

Fedorov declined to comment on his conversations with Musk. Musk and SpaceX did not immediately respond to requests for comment.

Fedorov said Ukraine’s intensifying long-range strikes against Russia were the most viable path to forcing President Vladimir Putin to end the war, but more investment in the weapons used for them was needed.

“We need more jet drones, we need cheap missiles, we need missiles with artificial intelligence,” he said. “You can’t conduct any kind of dialogue with Putin any other way . . . He understands only force.”

Nearly every company that’s trying to deploy artificial intelligence runs into the same problem: it’s hard to integrate the tech into everyday operations. Industrial heavyweight Caterpillar has spent decades dealing with a version of that problem in the physical world, and now it’s using its experience to deploy AI.

Caterpillar’s push into the autonomous space started with mining, where labor shortages and hazardous conditions can make automation particularly useful. Today, it sells automated haul trucks, drilling, underground loaders, dozers, remote-controlled construction equipment, and more. It also offers a software command center, fleet management, and even remote terrain intelligence as part of its autonomous toolkit.

“Now we’re in this super exciting time where we can take all of that learning from mining and bring it into much more dynamic environments, jobsites, quarries, and construction sites,” the company’s CTO, Jaime Mineart, told TechCrunch on the sidelines of the Ai4 conference in Las Vegas earlier this month.

The industrial giant is now applying AI more broadly, including in tools used by technicians and its own employees. One example is the Cat AI Assistant, which lets field technicians standing next to a machine use voice commands to pull up repair procedures, troubleshoot potential problems, and identify parts that may be needed before beginning a repair. Mineart said the tool is now being used by customers, operators and technicians.

The assistant draws on Caterpillar’s proprietary data, which spans information generated by its connected machines. Mineart said Caterpillar has about 1.6 million connected assets globally and more than 16 petabytes of structured data.

The company is also using AI to power software for scanning sites and generating digital twins in manufacturing to analyze operations, she said. And like nearly every other company, Caterpillar is using AI across its enterprise operations, as well as for software development. “We use AI agents to modernize legacy code, generate and test new software, and identify defects earlier,” Mineart said.

But Mineart is quick to point out that building the technology is only part of the challenge, as deploying an autonomous machine is not the same as transforming a site to use AI. Companies also have to rethink how people work alongside the technology and how existing processes need to change.

“The hard part about autonomy and about physical AI is incorporating that technology into the customer jobsite and into the workflows,” she said.

Mineart said the company leans on experienced operators to help train AI systems, leveraging institutional knowledge built over decades. And as machines become more autonomous, some operators may shift from controlling a single machine to overseeing multiple machines from a remote command center.

That transition, however, is creating a new challenge for Caterpillar: training its 118,000 employees. Mineart said the company plans to spend $100 million over the next five years to train its workforce in AI, autonomy and robotics.

That investment is likely being put towards helping the company make the most of the broader boom in AI infrastructure, which is already helping its top-line. Caterpillar’s quarterly revenue reached an all-time high of $20.5 billion in the second quarter, helped by strong demand for power-generation equipment used in data centers. Its power-generation division saw sales spike 72% to $3.10 billion, and CEO Joe Creed said that “no one is slowing down” when it comes to demand for cloud computing and generative AI infrastructure.

Elon Musk says he’s found a way to solve one of AI’s biggest bottlenecks by making a hard-to-manufacture turbine part himself.

On Saturday, Musk confirmed what a secret foundry SpaceX has been building in Bastrop, Texas, is for — an apparent response to a story that was already closing in on the details. Earlier in the day, The Information published a report citing job listings that explicitly mention a “blades and vanes foundry,” plus findings from Corey Trinetti, a due diligence specialist who authors detailed reviews of AI infrastructure sites in his newsletter and who’d reported that SpaceX had bought roughly 830 acres near its existing Starlink factory in Bastrop between March and June.

“SpaceX and Tesla are each building 100GW/year of solar production capacity as fast as possible,” Musk wrote on X on Saturday, “but natural gas will still be needed to supplement and bootstrap solar for several years. The limiting factor for nat gas turbine production is casting the blades & vanes. By doing in-house casting at SpaceX, we can accelerate nat gas turbines coming online by up to 18 months, which is a profound game-changer.”

The “why” of all this goes back to one of the biggest challenges facing the AI industry right now. GPU shortages are still an issue — Nvidia’s newest Blackwell chips are still running lead times of several months, for example — but a second constraint has emerged alongside it, which is the physical power grid. The International Energy Agency projects global data center electricity use will roughly double by 2030, and gas turbine maker GE Vernova says it’s essentially sold out of production capacity through 2030 due largely to AI infrastructure demand.

That shortage is why building private gas-fired plants next to data centers, instead of waiting on the grid, has become a ubiquitous strategy for so-called hyperscalers, including Amazon, Google, Meta, OpenAI, and Microsoft. After years of prioritizing wind and solar, they’re all now betting on natural gas to get data centers online faster.

As for the casting bottleneck specifically, according to The Information, the blades inside a gas turbine’s hottest section run at temperatures around 3,000 to 3,600 degrees Fahrenheit, which is roughly 800 degrees hotter than the melting point of the very metal alloy they’re made from. That’s only possible because of the blades’ internal cooling channels and thermal-barrier coatings, plus the specific way each blade is cast. Just four companies worldwide have mastered the casting process well enough to produce them at industrial scale, and all of them are tapped out right now.

What makes the whole thing especially difficult is that each blade has to be cast as a single, unbroken crystal, grown slowly inside a vacuum furnace, without the microscopic seams that let ordinary cast metal crack under stress. It’s a tricky process even for the smaller blades used in jet engines; the blades in power-plant turbines are considerably larger, which makes producing them at that scale and without defects even harder.

If SpaceX pulls this off — and it’s easier said than done, of course — it would mean a Musk-controlled entity holds a manufacturing capability that every other AI infrastructure builder currently depends on a tiny oligopoly for, giving SpaceXAI an edge that’s difficult for any well-funded but non-manufacturing competitor to copy quickly.

But it would also mean more gas turbines coming on fast, and turbines in the ground are already drawing federal lawsuits and peer-reviewed health research over the pollution they emit.

In Memphis, where SpaceXAI has run gas turbines to power its Colossus data centers since 2024, the NAACP has repeatedly accused the company of operating turbines without the permits or pollution controls required by federal law. The organization’s concern is that turbines like these emit smog-forming compounds and hazardous chemicals like formaldehyde, pollutants linked to asthma, respiratory disease, and certain cancers. (The site sits near neighborhoods that already face heavy industrial pollution, and University of Memphis researchers said that in their own admittedly limited analysis, air pollution grew “slightly worse” because of the data center.)

But Memphis just happens to be the most visible case. The same fight is playing out anywhere gas turbines have become the default fix for data center power shortages. In Virginia’s “Data Center Alley,” a study commissioned by the Piedmont Environmental Council, using the EPA’s own COBRA health-impact model, found that emissions from a single facility’s eight full-time gas turbines could reach more than 2.5 million people across multiple counties — with the heaviest impact landing on already-marginalized communities — and cause an estimated 3.4 to 6.5 additional premature deaths a year, translating to $53 million to $99 million in annual health-related damages.

The list, and complaints, go on.

Elon Musk intends to bypass the power supply chain for AI data centers in a way others assumed was impossible, by making highly complex components himself.

It’s a move no one else has been bold enough to attempt.

SpaceX has taken steps to develop a foundry in Bastrop, Texas, about 30 miles from Austin, to make vanes and blades for large gas turbines, according to job listings and other indicators.

Update: After this story was published, Musk confirmed the reporting by tweeting that despite large investments in solar energy, “natural gas will still be needed to supplement and bootstrap solar for several years. By doing in-house casting [of blades and vanes] at SpaceX, we can accelerate nat gas turbines coming online by up to 18 months, which is a profound game-changer.”

In recent months, SpaceX advertised that it is seeking a senior operations engineer “to serve as the extreme owner of the construction, buildout, and operational ramp-up of a new blades and vanes foundry.” It has sought a senior materials engineer, an automation engineer, a senior tooling engineer to design castings.

SpaceX also says it is looking for a programmer to assist in the milling of parts in high-temperature superalloys at “SpaceX’s new industrial gas turbine (IGT) blades and vanes manufacturing line in Bastrop,” the Austin-area location.

Morgan Stanley analyst Adam Jonas says his research also shows Musk is developing such a plant.

Until now, a global oligopoly of four casting companies has tightly controlled production of these vanes and blades, engineered for ultrahigh heat of 3,000 to 3,600 degrees Fahrenheit, 800 degrees above their own melting point. The maxing out of their foundries is a key reason large turbine makers like GE Vernova are sold out until 2030 and why the AI building boom is short on power.

‘The Limiting Factor’

Musk previously stated in a February podcast that blades and vanes are “the limiting factor” to faster provisioning of turbines and ultimately to standing up data centers. (Blades are also known to crack under stress running volatile AI workloads, so having a few extras handy might be helpful.)

At the current production rate, Musk said, “it’s not clear to me that there’s enough usable electricity to turn on all the AI chips that are being made.”

His prognostication continues to prove accurate, as we have been chronicling in this column week after week.

If he succeeds in making the parts and using them to build or refurbish more turbines, Musk could potentially build some of the 10 gigawatts or more of terrestrial data centers he has told SpaceX investors he could put up by the end of 2027.

Musk said earlier this month that he had secured Nvidia chips for these future sites, and AI cloud providers I spoke with argued that he has the capital and gall to bootstrap his way to power.

SpaceX shares plunged after the call on fears about what it would all cost. Debates continued to rage between traditional energy experts who asserted that powering hyperscale data centers off grid is far harder than it looks and AI bulls like Dylan Patel of research and investment firm SemiAnalysis, who stated his belief that Musk could pull it off.

“This is because you have no idea how to build infrastructure,” Jigar Shah, a clean energy and power entrepreneur who ran President Biden’s energy loan office, shot back at Patel on X.

Musk’s potential to unlock even more fossil fuel power for AI data centers would represent a setback to low-carbon energy advocates who had hoped the supply crunch for turbines would encourage renewables. BloombergNEF has estimated that if planned off-grid gas plants for data centers come to fruition, they would boost U.S. carbon dioxide emissions 20%. That figure wouldn’t include Musk’s latest efforts, but it would count 2 GW of off-grid gas deployments that SpaceX’s xAI unit has said will be running by the end of the year at its data centers in the Memphis, Tenn., area.

Dual Use

It couldn’t be learned where he is seeking to locate his new data centers, but Grace reported last month that Musk was laying the groundwork for a major facility in Texas, where he also has plans in various stages for a massive chip foundry and solar panel manufacturing complex. He has also said he would expand facilities in the Memphis area.

Morgan Stanley’s Jonas believes SpaceX is standing up a Texas turbine-component foundry for both its data center needs and its Raptor turbopumps, vital internal components of the company’s rocket engines.

“Raptor turbopump castings and data center turbine airfoils draw on the same alloys, furnaces and trained workforce, letting a single foundry spread fixed cost across the space and AI businesses,” Jonas said in a report this month.

To be sure, nailing the casting process to replicate what leaders like Precision Castparts have done for decades isn’t easy and could take years. But it’s critical to SpaceX’s relatively recently adopted quest to become a cloud provider on par with today’s incumbents.

Getting Turbines

In the interim, Musk is working to access several gigawatts of bridge power in the also-scarce market for smaller turbines. Jonas said Musk could access about 3 to 4 GW of power by the end of 2027 through a mix of purchase agreements.

In recent months, Musk has worked quietly to source turbines with Wall Street hedge fund Fortress Investment Group, which has invested in at least two separate suppliers.

Fortress sold a supplier of mobile industrial turbines, APR Energy, to Musk personally, according to a Federal Trade Commission filing in May. APR said in January it had about 1 GW of turbines in its fleet, suggesting they would become available to Musk.

SpaceX may also be purchasing turbines from FTAI Aviation, which modifies and refurbishes the machines for commercial aircraft engines. FTAI, spun out from another Fortress portfolio company, announced a $1.465 billion turbine sale in July via its Chinese joint venture partner to “a leading international cloud services provider.” Jonas has identified FTAI as a possible supplier to SpaceX.

The Bastrop foundry plans are relatively new. SpaceXAI already operates a factory there, related to its Starlink satellite business, but between March and June it bought about 830 nearby acres, according to Corey Trinetti, a due diligence specialist who publishes detailed reviews of AI infrastructure sites in his newsletter, Measured AI. Two newly developed buildings adjacent to the Starlink factory that were not there a year earlier appeared in satellite imagery in July.

Trinetti was unable to find foundry permits or construction filings in Texas environmental or building records, however, so it isn’t clear whether SpaceX has started a permitting process.

Musk already stunned the data center industry once by completing the Colossus data center in Memphis in record time two years ago. It wasn’t the highest-quality facility, to be sure, but SpaceX is now making a lot of money from renting it out to Anthropic.

He’s trying to do it again with power, and betting against him is usually a risky strategy.

AI labs and neoclouds are buying them by the truckloads. It’s Apple’s fastest-growing business. Macs are on fire, thanks to agents and AI.

The hottest products at Apple right now are not the iPhone, the iPad or a buzzy new show on the company’s streaming service. They are two of the lowliest members of Apple’s venerable Mac product line: its boxy Mac mini and Mac Studio computers, which come without monitors, keyboards or mice and are designed for professional users.

The products are currently flying off the shelves, because it turns out that the machines are well suited to running agents—AI software that handles multistep tasks, from editing and testing code to automating email inbox organization and summarizing documents. And they’re also a hit with AI developers who want to save on cloud computing bills by training and running their models locally. In the most recent June quarter, Mac sales grew nearly 29% annually to $10.4 billion, faster than any other product segment at the company.

One sign of the unlikely star status for the products occurred on June 23, when Apple held an unusual event on its Cupertino, Calif., campus, known as Apple Park. Normally, Apple is hyperfocused on selling its hardware and services to consumers, but it devoted the June event—called Business at the Park—to enterprise customers. Executives from Walt Disney Co. and Ford Motor were there, as was Anthropic co-founder Jared Kaplan. Tim Cook and John Ternus—Apple’s outgoing and incoming CEOs, respectively—showed up too.

One of Apple’s loudest messages to the attendees was that its hardware is ideal for handling AI chores locally rather than in the data centers of pricey cloud services. And according to one attendee, Apple’s Mac mini was the “darling” of the event.

For AI developers, Mac minis and Mac Studios have become favorites because Apple loads them with more powerful versions of its M-series computer chips and more memory. Unlike a thin MacBook, which can quickly heat up and throttle chip performance, the boxy Macs have cooling systems that can handle long, complex AI tasks. And while Apple’s in-house silicon chips aren’t as powerful overall as Nvidia’s graphics processing units—which dominate the AI chip market—their single shared pool of memory gives them performance advantages for AI workloads.

Some AI labs like OpenAI have bought tens of thousands of Mac minis and Mac Studios for the task of reinforcement learning, a technique where the AI learns through trial and error, said people familiar with OpenAI. OpenAI is using these Macs for training computer-use agents and is desperate to acquire more of them, the people said. Anthropic also rents Mac minis from Amazon’s cloud service, Amazon Web Services, people familiar with Anthropic’s efforts said.

Some startups are even building out Mac-only cloud services, in anticipation of a future where demand for Apple hardware grows so much that some AI developers will want to run on Macs in data centers rather than locally. One such startup is a new Apple hardware–based neocloud company, Mount Thor, founded by Peter Voell, a former OpenAI compute infrastructure employee. Though it is still mostly operating in stealth mode, Mount Thor’s minimal website describes its offerings as “AI execution environments on Apple hardware.”

Apple promotes open-source software from EXO Labs that enables people to cluster Macs together to run trillion-parameter AI models locally.

The Mac’s popularity for local AI has attracted the attention of Nvidia. While most AI training and inference chores are likely to remain in the cloud, demand for local AI has been on the rise, and Nvidia sees Apple as its biggest competitor in local AI, according to a person who has discussed the competitive threat with Nvidia executives.

Nvidia has even come up with new products to get a bigger piece of that market. Late last year, the chipmaker released DGX Spark, an AI desktop computer with a compact squared-off design similar to that of the Mac mini.

Unfortunately for Apple, the exploding popularity of Macs has come at a time when it and other device makers are suffering from severe supply-chain woes. They are wrestling with a historic shortage of memory chips caused by AI data centers’ voracious need for them. The most advanced models of Mac minis and Mac Studios, which have the chips and memory configurations most appealing to AI developers, have been out of stock for months.

“The memory shortage is impacting Apple, and if enterprises are unable to get their Macs, they’ll migrate to different hardware choices,” said David Stout, CEO of webAI, a startup selling AI tools to enterprises that use Apple hardware.

That’s already started to happen. Some businesses have started to look at alternatives to Apple’s hardware because of the supply limitations for Macs over the past year, said Todd Dailey, a former enterprise marketing manager for Apple’s AI products. Companies and AI developers consistently bring up Nvidia’s DGX Spark as an option and is readily available now, said Dailey, who left Apple in April and is now an independent AI consultant.

Because Mac demand has been so hot, Apple earlier this week announced new Mac mini and Mac Studio models with more powerful chips. It was an unusual time for the company to come out with the products: Apple typically does its end-of-the-year Mac refresh in October or November. In its announcement, it bragged about how the new Mac Studios can be clustered together to make a more powerful system for running giant frontier models.

According to Dailey, the soaring demand for Macs among business customers has been accidental rather than something the company had planned for: Apple doesn’t have a dedicated engineering team for business customers, nor does it have any employees focused on developer relations. “The idea that any team at Apple has an actual plan for embracing enterprise AI is a joke,” Dailey said.

Over the years, Apple has tried selling various enterprise server products, but support for these products has been inconsistent and spotty. The company sold Xserve, an Apple-made server product with Intel processing chips, between 2002 and 2011. It also released a server operating system based on Mac but discontinued that product in 2022.

Apple has recently begun building its own servers using Mac chips. However, they’re strictly for internal use with Private Cloud Compute, the company’s server system for processing heavy-duty AI tasks that exceed the processing capabilities of the iPhone or Mac. Some business customers have asked if Apple can sell them access to these servers, but it has turned them down so far. Instead, the company is hoping partners like webAI and Mount Thor can help push its hardware deeper into the business world, people familiar with Apple’s enterprise strategy said.

The Trump administration had proposed canceling the project, named after astronomer Nancy Grace Roman, before Congress opted to continue funding it

NASA launched the Nancy Grace Roman Space Telescope aboard a SpaceX rocket to study dark matter and search for other planets.

The telescope’s main camera has 300 megapixels, giving it a much larger field of view than NASA’s Hubble observatory.

The total cost of the telescope, including development and operations, is expected to be roughly $4.3 billion.

The National Aeronautics and Space Administration launched a new flagship telescope, a powerful observatory device that will study so-called dark matter and hunt for other planets.

A SpaceX rocket took off with the device, called the Nancy Grace Roman Space Telescope, just before 7:30 a.m. ET Sunday from a launchpad in Florida. The launch went smoothly, setting up the telescope to fly out to an orbit about one million miles from the planet.

During the course of a mission designed to last at least five years, the observatory will take sweeping surveys of the cosmos. The telescope’s main scientific sensor is essentially a giant, infrared camera dubbed the Wide Field Instrument.

The WFI camera has 300 megapixels, giving it a much bigger field of view than NASA’s Hubble observatory, according to the agency, and will gather data much faster than that device. The camera should allow the telescope to capture extremely high-resolution images in space.

“It’s the instrument that allows us the panoramic imagery that make Roman so iconic,” Dominic Benford, a NASA scientist working on the telescope program, said Sunday during a NASA livestream of the launch.

The telescope also includes what is called a coronagraph, a system that includes mirrors and optical sensors to try to image other planets.

It will study dark matter, a mysterious form of matter that scientists believe will help them understand the nature of the universe. The telescope will also track hundreds of millions of stars, searching for light that could indicate the presence of so-called exoplanets.

Named after a pioneering female astronomer at NASA, the telescope nearly didn’t get to space. The Trump administration proposed canceling the device in budget pitches, but Congress continued to support funding the device and having it launched. The administration has sought significant cuts to many NASA science programs.

During a post-launch media briefing Sunday, President Trump called in to tout the flight, giving NASA kudos for the mission. “I just want to thank everybody and congratulate you. Boy, it looks beautiful on television,” Trump said of the launch.

NASA has said the total cost of the device, including development and operations, is expected to be roughly $4.3 billion.

The telescope uses mirror technology originally developed for the National Reconnaissance Office, a U.S. intelligence agency that operates powerful spy satellites that monitor Earth. The NRO donated the gear to NASA more than a decade ago, allowing teams at NASA to repurpose the technology.

SpaceX launched the device using a Falcon Heavy rocket, a more powerful variant of its workhorse Falcon 9 vehicle. The telescope weighs about 18,000 pounds, and measures more than 42 feet long and over 14 feet wide.