Microsoft Outlook outage blamed after 15 injured in hotel pool chlorine incident
Most victims of last week's Microsoft outage lost access to their email, but 15 unlucky guests at a German hotel got a dose of chlorine when a warning about dangerous levels never reached staff.
The incident, first reported by German broadcaster WDR, happened on Friday at Hotel Ling Bao, part of the Phantasialand theme park in Brühl, Germany, where a malfunctioning automatic dosing system pumped too much chlorine into the swimming pool.
The hotel's monitoring system spotted the problem and emailed staff a warning about the excessive chlorine levels. Unfortunately, Microsoft's Outlook service was caught up in a global outage, meaning the warning never arrived.
Brühl Fire Department said the problem started when a technical fault caused the pool's dosing equipment to pump in more chlorine than it should have. Chlorine gas then escaped into the swimming area, leaving guests needing medical attention.
Paramedics checked over 15 people at the hotel, and five were taken to hospital. Four suffered minor injuries, while one was moderately injured and needed further treatment. The fire department sent about 40 personnel to the scene, and crews spent roughly two hours dealing with the incident.
The incident came as Microsoft was battling a major outage that began on Thursday, November 12, and dragged well into Friday. Outlook, Teams, SharePoint, and other Microsoft 365 services were knocked offline, leaving businesses unable to access email and documents. The disruption wasn't confined to office workers unable to check their inboxes, with hospitals, police forces, and government departments reporting problems accessing the Microsoft services they use for day-to-day operations.
Redmond has since blamed the outage on a bug introduced during an earlier network capacity upgrade. It sat there until a later network change set it off, and the failures then spread across the company's infrastructure. Microsoft ruled out both cybercriminals and AI as culprits, and has issued the customary apology.
At Phantasialand, though, a broken inbox was the least of anyone's worries.
The chlorine dosing equipment went haywire independently of Microsoft's troubles, but the email warning staff about the problem apparently never arrived. Whether the hotel had a backup alarm, or anything capable of shutting off the chlorine supply automatically, remains unanswered.
Why a hotel would rely on Outlook to deliver warnings about dangerous chlorine levels is anyone's guess. Email can be delayed, filtered into spam, or simply disappear when Microsoft's infrastructure throws a tantrum, none of which is particularly helpful when you've got a pool with guests breathing in noxious gas. ®
THE FLASH’S Jesse L. Martin Joins PERCY JACKSON Series
- The Flash‘s Jesse L. Martin portrays Annabeth’s father in season 3 of Percy Jackson and the Olympians.
- Walker Scobell can’t believe he gets to share the screen with someone from The Flash.
- Nerdist Take: The Flash should be everyone’s favorite show, and it looks like Scobell’s dream of playing Kid Flash is one step closer to coming true.
It looks like Walker Scobell’s Flash manifestation is coming to life! Percy Jackson and The Olympians season 3 is on its way, and it looks like we’re in for a crossover. Well, not exactly! The Flash‘s Jesse L. Martin, aka Joe West, is guest starring in Percy Jackson season 3 as Annabeth Chase’s (Leah Jeffries) father. Being cast for this role proves he’s one of the greatest TV fathers.
Disney +/CW
While Military historian Frederick Chase loves his daughter, he struggles to find the right words. Conceived through thought and not action, Annabeth arrived on her Dad’s doorstep in a golden cradle. Because Annabeth was a gift of love and shared wisdom, Athena, goddess of wisdom, insists on Fredrick raising their daughter in the mortal world. In both the Percy Jackson books and series, Annabeth and Fredrick have a complicated relationship, but it looks like this season they will find their footing.
RELATED ARTICLE
PERCY JACKSON Takes on His Biggest Challenge Yet in Season 3 TrailerFlash fans know Martin’s Joe West as a loving and protective father in that world as well. I look forward to seeing how the best Dad in the Arrowverse will portray the other side of fatherhood in Percy Jackson. And it looks like I’m not the only one.
Scobell grew up watching The Flash, with the rest of the world, using the show’s opening monologue for auditions. Scobell calls out The Flash as his “favorite show” during an interview with ScreenRant. The young actor is happy to share the screen with Martin, and still can’t get over how “crazy” it is.
CW
Scobell’s love for The Flash doesn’t end there; earlier this year, he revealed Wally West as his dream role. During an interview for The Supes Show, Scobell shared that the two DC characters he would like to portray are Kid Flash and Green Lantern. And while the DCU has yet to introduce Kid Flash, the character played a major role in the show.
Given how well he captures the heroism of Percy, I think he’s more than proven himself. And given the recent shift in the DCU, leaning in a younger direction is something DC should consider. If you ask me, it’s time for another “Arrowverse,” not a reboot, but instead another generation of heroes for audiences to fall in love with.
Disney+
Ironically, Joe’s relationship with Wally might mirror what we see between Fredrick and Annabeth in Percy Jackson. In The Flash season two, Joe discovers his grown-up son Wally; while excited initially, he didn’t know how to parent Wally. Martin’s performance portraying the layers of fatherhood, and the on-screen juxtaposition of how he parents his children differently yet effectively, was phenomenal. Casting Martin as Fredrick Chase wasn’t just needed; it was necessary!
RELATED ARTICLE
SHOCKER: James Gunn and Peter Safran Out at DC Studios Kendall Gayle Washington writes about movies, TV, games, comics, and all things nerdy for Nerdist.The post THE FLASH’S Jesse L. Martin Joins PERCY JACKSON Series appeared first on Nerdist.
Trump finally gets around to talking about the FAA catastrophe
President Donald Trump remains disconnected from reality while the government crumbles around him. This was made distressingly clear after the Federal Aviation Administration issued a lengthy nationwide ground stop on Friday, one that seems to have occurred because of an overloaded communications system in desperate need of updating. Trump didn’t have one word to say about this…
Trump Demands More Money to Fix Flight Issues
Artificial intelligence was not to blame for the nationwide grounding of airplanes, says the Federal Aviation Administration.
Instead, government officials say that a router issue was at fault: Problems during a routine router upgrade in Arkansas caused it to ping the FAA’s communications network hundreds of times per second, overwhelming the system and causing difficulties with air traffic communications.
However, not everyone buys the FAA’s explanation. Robert W. Mann Jr., the founder of an airline analysis and consulting firm, told The New York Times, “The idea that you would have a single router take down a national network is just not plausible.”
The timing of the ground stop also came conveniently close to the FAA’s rollout of a new AI system, SMART, an $875 million setup that officials say centralizes data on weather, flight paths, and staffing to help specialists find and fix problems in the sky. SMART was launched nationwide on November 10, three days before flights were grounded across the country.
Homeland Security Secretary Markwayne Mullin said Friday that there’s no indication that the new AI system was related to the ground stop, as did the startup that designed the system, Air Space Intelligence.
House Democrats are raising concerns about safety, and asked to audit the SMART system last month.
“Concerningly, the nation’s front line aviation safety workforce apparently had little to no input into SMART’s development and have received limited training and information during the current deployment process,” Democrats on the House Committee on Transportation and Infrastructure said in an October 13 letter. “The use of artificial intelligence within SMART has rightfully raised serious questions about its reliability and safety benefits, particularly on its potential to interfere with decision-making by experienced air traffic controllers or to hallucinate safety-critical information.”
Meanwhile, the Trump administration is asking for more funding for the FAA, blaming this failure on old equipment.
Transportation Secretary Sean Duffy said that $50 billion is needed to upgrade the FAA’s network, but billions have already been spent: President Joe Biden awarded $25 billion over five years as part of a larger infrastructure bill, and $12.5 billion was approved in 2025 to modernize the FAA’s air traffic system.
President Donald Trump finally spoke up about the situation on Sunday morning, two days after the planes were grounded, advocating for further funding.
“If Congress approves the desperately needed money for Air Traffic Control, we can get it built quickly and efficiently,” Trump wrote on Truth Social. “It will be the BEST SYSTEM anywhere in the World. Much of our current air control system is 50 years old - RIDICULOUS!”
Trump Still Crashing Out Over Midterm Elections
President Trump is still whining about his poor midterm election results, posting a clip of a withered Bill O’Reilly threatening Democrats with legal action Monday morning.
“The Democratic Party in Alaska cheated, and in a big way,” conservative commentator O’Reilly said in the clip, which Trump posted without a caption. “And that’s gonna lead to criminal prosecutions.”
O’Reilly is referring to the Alaska Senate race between incumbent Republican Senator Dan Sullivan and Democrat Mary Peltola in which Republicans accuse Democrats of purposefully attempting to sow confusion on the ballot by running a “Dan J. Sullivan” as an independent alongside their own Dan Sullivan.
Trump preemptively protested the outcome of the race for weeks, calling for a criminal prosecution of Peltola—and for the Justice Department to sue Alaska and block the certification of the vote—if the incumbent Sullivan lost. Trump even went so far as to call for votes for Dan J. Sullivan to go directly to the senator.
While the race has yet to be called, Peltola has pulled away from Sullivan to establish a comfortable lead with 97 percent of the ballots counted.
Newly empowered Democrats ready to overhaul gun laws in red states
By Chip Brownlee and Jennifer Mascia for Stateline The blue tsunami triggered by the Nov. 3 midterm elections saw Democrats win not just Congress but also several governors’ offices. Some—like Ohio—were expected, while others—like Texas—came as a surprise. The new map has some implications for state-level gun law overhauls. In the absence of comprehensive federal legislation…
Google: AI agent turned a security patch into a working exploit with no human help
A suspected Russian-speaking operator used an AI agent to turn security patches into working exploits with no human in the loop – a first such occurrence, according to the Google Threat Intelligence Group (GTIG) – before ultimately stealing data from victim organizations.
Google documented this attack in its fourth-quarter 2026 AI threat tracker, and said it marks a significant advance in how attackers use LLMs for fully autonomous vulnerability discovery and exploitation compared to Q3.
In this case, the threat hunters found an exposed staging server hosting a multi-agent vulnerability research and weaponization framework that integrates with multiple frontier and local models. The framework, called exploit_agent, used Russian-language tooling, and was configured to conduct continuous differential analysis of “popular web application software,” according to the report.
This includes comparing code before and after security updates, identifying potential vulnerabilities, developing exploits, and then automatically executing them without any human involvement.
The server also contained data stolen from victim organizations, and “multiple” post-compromise scripts that allowed attackers to perform nefarious activities including internal reconnaissance, brute-forcing, installing webshells, and removing other attackers’ webshells.
“Victim telemetry on the server suggests the operator was specifically interested in targeting French websites associated with the media and entertainment industries, as well as the education sector,” according to the report.
The report also documents other instances of attackers moving beyond manual LLM prompting during cyberattacks and instead deploying autonomous, multi-agent offensive security frameworks to carry out multi-stage intrusions. These include both financially motivated cybercriminals and cyber-espionage groups.
In addition to the recent attacks involving the open source penetration-testing tool ARTEX, attackers have also used other AI-powered offensive frameworks, including CAIRN, CyberStrikeAI, HexStrike AI, and PentAGI. In another instance, GTIG observed attackers using PentAGI alongside proxy services to target the aerospace sector.
“We have observed several threat actors separately experimenting with use of these frameworks and harnesses, including financially motivated, as well as suspected espionage threat actors with a PRC nexus,” the Googlers wrote. “Given the low barrier to entry, ease of integrating new exploits, and massive scaling potential, we believe it is highly likely that threat actors will continue to use these frameworks in campaigns targeting organizations globally.” ®
Mazama Energy has a solution for geothermal’s biggest drilling problem
Monday morning traffic: E Cliff Dr closed; Hwy 1 & Hwy 9 lane closures ongoing
This post is updated throughout the day to reflect the latest incidents. It was last updated at 6:31 a.m..
Here’s what’s happening on the roads this morning…
▼︎ new incidents
Road incidents as of 6:30 a.m. on November 16- CHP closed the road at E Cliff Dr and 26th Ave. in Eastside / Live Oak on Nov. 15 at 7:35 a.m. because sand was still on the road. There was no flooding or standing water. Signs were put up to keep the road closed for safety.
- South Highway 1 at Park Avenue in Capitola / Soquel is facing closures for paving work. The closure is expected to end at 7:01 a.m. on Nov. 19.
- Highway 9 at Pool Drive in San Lorenzo Valley has alternating lane closures because of bridge work. This will continue until April 30, 2027, at 6:59 a.m.
Disclosure: Traffic incidents are partially generated by artificial intelligence. We are constantly working to improve the accuracy and quality of our AI-generated content. However, there may still be errors or inaccuracies. If you have any questions or concerns, please contact us.
The post Monday morning traffic: E Cliff Dr closed; Hwy 1 & Hwy 9 lane closures ongoing appeared first on Lookout Santa Cruz.
A period of transition
As always, if you find value in this work I do, please consider helping me keep it sustainable by joining my weekly newsletter, Sparky’s List! You can get it in your inbox or read it on Patreon, the content is the same. Also, if you missed the big Kickstarter, there are still some copies of the latest book, Our Long National Nightmare, available on the Tom Tomorrow popup store! (Don’t delay…
The Jensen Huang saga could be the beginning of the end of the AI bubble – or just a boon for competitors
KETTLE Jensen Huang is out at Nvidia after a trio of affairs that include an alleged Chinese spy and two of his subordinates, turning the week into one that's shaken the tech sector.
You can listen to the latest episode of The Kettle right here on this page, as well as on Spotify, Apple Music, or YouTube. Those platforms also let you subscribe to The Kettle, so you are always notified when the latest episode goes live.
What will become of Nvidia now that its only leader in 33 years, and one whose vision shaped the company into one of the largest in the world, is gone? Who will take charge? Will other chipmakers take advantage of the chaos to surpass the Geforce maker? Will Jensen be the one to put a pin to the AI bubble?
Join host Brandon Vigliarolo, systems editor Tobias Mann, and reporter Dan Robinson to talk about the Jensen Huang affair(s) and how the misdeeds of one of AI's biggest names could reshape the entire industry on this week's episode.
A lightly edited transcript is included below:
Brandon (00:01)
Hey everyone and welcome to another episode of The Register’s Kettle podcast. I'm your host, Reg Reporter Brandon Vigliarolo, and the big news on the tech industry's mind this week is the ouster of Nvidia founder, CEO, president, and board member Jensen Huang over allegations that he not only had an affair with a suspected Chinese spy, but also at least two Nvidia employees who were his subordinates.
Nvidia's been shaped into one of the most valuable companies in the world over the past thirty-three years, and Huang has played a pivotal, if not the pivotal, role in that ascension, meaning his controversial departure is going to rock the industry, if not the greater world. With me to discuss this is our systems editor Tobias Mann and Reg reporter Dan Robinson. Thanks to you both for coming on.
Dan Robinson (00:42)
Hi.
Tobias Mann (00:43)
Good to be here.
Brandon (00:43)
Quite the week, huh? It seems like we went from Jensen’s just going to survive this spy kerfuffle to the board basically seeming to demand his resignation in the course of a few days, right? So Dan, bring us up to speed on the timeline here, what happened?
Dan Robinson (01:01)
Well, it all happened over the last couple of weeks. That was when the story first surfaced, or it was first reported in the media, of an affair with an alleged Chinese spy who happened to work for this company in Singapore that bought a lot of Nvidia's products.
Brandon (01:30)
Right, company was Megaspeed, right? And her name's Alice Huang, no relation to Jensen, coincidence.
Dan Robinson (01:36)
So it's all happened fairly quickly, but like you say, at the start of this week, it all seemed like Jensen could weather the storm. We were discussing this in our editorial chat and we thought that basically there isn't really anybody who could replace him, or so we thought. He practically is Nvidia, and it's been his vision that's driven the company for so long. But then he came out and he confessed and said he was sorry to all the Nvidia employees. Then the second affair came to light and very quickly, on November 11th actually, Wednesday, he handed in his resignation and it was accepted by the board. So we got to hear about it yesterday [Thursday] when they announced it and it was filed with the SEC.
Brandon (03:03)
And since then a third affair has come out, right? The New York Times published a finding.
Tobias Mann (03:06)
Just a few hours later, yeah.
Dan Robinson (03:08)
It seems so. Makes you wonder how many others may be about to come out in the following days.
Brandon (03:13)
We're not going to speculate, but my gosh, it seems like that leather jacket really made him irresistible. It’s kind of crazy, right? It seems like the internal affairs with Nvidia employees are the ones that sort of broke the board's back, right? The spy affair seemed like it was kind of like…. Apparently American authorities have basically said they've found no direct evidence to say she is a spy.
Dan Robinson (03:44)
Exactly.
Brandon (03:45)
It could just be that she runs a chip company too, or at least a company that is dealing in servers and datacenters. So she has a business reason beyond politics to potentially want to get involved with Jensen. What do you guys think? Was it the internal elements that finally kind of broke the camel's back here and forced the board to act?
Dan Robinson (04:08)
I think that was the final straw. I think quite possibly he could have weathered the storm as we were saying, if it had just been the one affair outside the business.
Brandon (04:20)
Outside the business, right? Because now like you were saying, Tobias, not only do you have potentially an international spy affair, but you also have an HR issue to contend with.
Tobias Mann (04:29)
Right. And I do think that the board could wave their hands around an alleged honeypot scenario. “He was targeted, this was a deliberate act, they preyed on his insecurity.”
Brandon (04:44)
“It wasn't as big as it's made out to be in the media.”
Tobias Mann (04:47)
They could make up any number of excuses that they wanted in order to wave their hands over it and keep him CEO or if their arms were twisted by US intelligence services they could have taken him out of an executive role and pushed him into a chairmanship. They had options.
And then you had the HR kerfuffle of him allegedly sleeping with another employee. And at that point now there's a pattern. I do sort of question once the news came out that he had an extramarital affair with an employee, whether or not that had influenced him being targeted by Megaspeed and or the Chinese intelligence service, assuming those allegations prove out.
Brandon (05:42)
We don't know the timeline on one or the other, do we?
Tobias Mann (05:46)
Right, no. And so it could be that the Chinese intelligence service became aware that this was an avenue that could be exploited and they took it. It's not lost on anybody that he spent most of the last year and a half lobbying to get his chips back into China. And largely successfully on the US side of things. It's been held up by Beijing, which kind of undermines that theory a little bit.
Brandon (06:22)
It seems like if she was a spy doing China's bidding and he managed to get the Trump administration to allow those older chips in, then Beijing wouldn't have been like, “No, screw you, we're making our own….”
Tobias Mann (06:34)
But there's the other allegation here, which was that Alice Huang and Megaspeed were smuggling GPUs. Their operation in Singapore was set up such to bring more modern GPUs into the market. And so it could have been that Beijing all of this is alleged didn't really care for its model labs getting old hardware. So they could appear to be blocking us and promoting their own chips while smuggling in more advanced silicon from Nvidia through Singapore.
Brandon (07:10)
Through Alice Huang's company, essentially, right? So I know Jensen's going to be appearing before Congress to talk about all this, right? This is not over by a long shot, right?
Tobias Mann (07:26)
So, I don't think that we should play up the gray market and the black market side of the GPU smuggling operation too much. I highly doubt Jensen knew anything about that, other than he was aware that it happens. But gray market and black market smuggling, as anybody who's worked in the electronics and the technology field will know is nothing new.
When I was a young kid in college, gray market cameras, if you needed a new camera, was a great way to save a couple hundred bucks on a camera so long as you weren't expecting a warranty to come with it. So I wouldn't read too much into the smuggling side of things. I think that's happening, but I don't think Jensen's involved at all.
It's hard to ignore the fact that he spent the last year and a half lobbying Washington to allow chips into China, more powerful chips into China. And even prior to that, he had instructed his engineers to develop China-specific accelerators and cut down owned fuse off cores, whatever it took in order to limbo under export controls. And it was not particularly popular with the last administration. Gina Raimondo openly criticized Nvidia about how they were taking the spirit of the law out of context and Nvidia pushed back a little bit about that.
Dan Robinson (08:56)
That's right. She did say at one point that whatever Nvidia did to get around the rules, they were going to change them to block him.
Tobias Mann (09:08)
And she eventually walked that back a little bit, because the rules are there, and so you can't really get too upset with somebody for trying to abide by the rules. Was Nvidia following the spirit of it? No.
Brandon (09:18)
Sure. That's a huge potential market for Nvidia, right? China is a massive market and they have a huge appetite for chips, right?
Tobias Mann (09:29)
In one quarter They wrote off ten and a half billion dollars of potential profits because they could not ship H20s to China because the Trump administration kneejerked.
Brandon (09:38)
I remember when that story came out. That was a huge amount of money, right? How much of their market is Singapore in that area? It's some massive chunk of change.
Tobias Mann (09:50)
And Dan, you probably are aware of this as well. So there's a lot of companies that operate out of Singapore but don't actually have data centers there because it's not a particularly power-dense area for them. And so there's a lot of companies that, if you look at it at face value, looks like they're importing a lot of GPUs to Singapore, but the GPUs don't dor the most part don't actually make it there.
Brandon (10:21)
They're ending up in Malaysia.
Tobias Mann (10:22)
Except perhaps in the case of Megaspeed.
Brandon (10:24)
I think Simon [Sharwood] wrote a piece recently, I think it was actually just yesterday, where he mentioned that in one case, Singaporean buyers sometimes account for over twenty percent of the quarterly revenue for Nvidia. So that's a huge market, right? Whether or not those chips are ending up in Singapore or China or some other country that's operating as a territorial expansion of Singapore for serving data centers to them, that's a lot of money. And that market could grow immensely. So, there's a fair argument to be made there that in Jensen's just being a capitalist, right? He sees a market opportunity. He's going to try to take it, it's not a conspiracy. It's not necessarily that the company's acting in collusion with China to circumvent imports. It's just, well, we got this opportunity, so we're going to follow the letter of the law, not the spirit. I don't think it's necessarily fair to fault Nvidia for that.
Dan Robinson (11:20)
And Simon wrote [the] story which was about the Singaporean authorities saying that they were having a very hard time getting cooperation from the US authorities when they were investigating Megaspeed. So it does make you wonder if this really is a spying thing or if they're really worried about the smuggling, why were they not cooperating with the Singaporean authorities over this?
Brandon (11:54)
Right. Right.
Dan Robinson (11:55)
There's something else going on there, isn't there?
Brandon (11:58)
Very possibly. So OK, so more broadly, stepping back from just Nvidia and its trials and tribulations as one of the world's most largely successful companies, Tobias, you watch the chip industry. It's your thing here at the Reg. And watching chips obviously nowadays means watching Nvidia. I'm sure this is going to have an impact on the AI industry, the tech industry. And possibly beyond, right? So what have you been hearing from people you talk to, from other companies, from Nvidia itself, that is going to point to how this is going to impact the space?
Tobias Mann (12:41)
I just saw this morning, I think it was Wired had a piece out, that there's some disbelief and some upset within Nvidia's own employees right now. Bring the boss back.
Brandon (12:54)
I think some are petitioning to bring him back.
Tobias Mann (12:57)
And I think one of the things that can't be overstated is there's no real replacement for Jensen. He is the face of the company and they're now in the middle of an identity crisis. Ian Buck, who runs their HPC division and is the father of CUDA, is favored by many as a possible replacement. But I don't know that that's something Ian really wants to do. And while he's taken a much more public role in the last couple years, I don't know that he has the pull and the same degree of charisma that Jensen did in order to kind of reshape reality. The joke used to be that Steve Jobs had a reality distortion field. And that was certainly true of Jensen. Right now, Colette Kress, the acting CEO, who is the bean counter keeping Nvidia's operations and Jensen's ambitions grounded for the last 13 years, is acting CEO but I don't think she's going to be long term in that position.
Brandon (14:12)
Sure. It's interesting you say like keeping Jensen's ambitions grounded, right? But I know a lot of people have been talking about how a large part of the reason that Nvidia turned into what it is is because of Jensen Huang's bets that he's made over the years. Betting that GPU architecture was going to be a good model for AI was one of the reasons the company exploded to the way it is. So I wonder if she could go on to keeping Nvidia on pace, right? Because she doesn't maybe have the risk-taking nature that Jensen did or does.
Tobias Mann (14:44)
Right. So this is the give and take of that relationship as far as I've seen it over the years. Things didn't always go well for Nvidia. There was a lot of hard pivots that they had to do in order to keep the business growing. You may recall back to when the pandemic shut down the economy. There was a hard pivot to selling GPUs for crypto mining because at least that's where the money was coming from. And while it was not at all popular with PC enthusiasts, it did keep revenues coming in. And so there were certainly risks and rewards that were measured. And I have no doubt that Kress was heavily involved in that. With that said, I don't think that she is the personality, how to say this, caricature. I don't think that she is the person to replace Huang unless the culture is going to change dramatically.
Brandon (15:48)
You can't imagine a CFO having this sort of vision thing that Jensen Huang has had to push all these various projects and everything that he does on stage with the dancing robots and what have you.
Brandon (16:10)
And Jensen was an engineer, right? The company was headed by an engineer for 33 years. That creates a certain culture, right? You don't want this to turn to another Boeing situation, right? Where the bean counters take over and the company suddenly becomes so risk averse that they're just doing anything they can to keep the profits coming in, which leads to an inferior product.
Tobias Mann (16:29)
And to draw some comparisons, this is not the first time we've seen something like that happen. Bob Swan's tenure at Intel as CEO, he was forced out because he could not take, he was running it like a CFO. He had been the CFO. He was forced out and replaced by Gelsinger, whose unchecked ambition proved too much for the board in the end. So there is some grounding that is required in companies this large in order to keep personality and ambition from getting in the way of the business, and vice versa. In this chaos, however, there's great opportunity for other players right now. And it's not going to be the clean fight, I don't think, that any of us would necessarily want to see.
Brandon (17:22)
Any ideas who comes out on top or too early to tell?
Tobias Mann (17:25)
You can't ignore the fact that Nvidia is still the biggest player in this space and that's going to remain true. The execution over the roadmap, the existing roadmap, is going to tell us whether or not the company is going to survive long term. Having said that, from a technological standpoint, hardware is not Nvidia's advantage right now.
AMD has caught up and in some ways surpassed Nvidia in terms of accelerators and chips. So Lisa Su is in a great position to capitalize on this right now. There is an emerging push from hyperscalers and increasingly model devs towards semi-custom hardware developed in collaboration with companies like Broadcom and Marvell. And so if...in this chaos, hyperscalers go, I'm going to insulate myself from the potential that the roadmap may fall apart by investing more in our in-house silicon, you could see Nvidia's claim over the market decline over this time. And at the same time, you could also see chip startups asking inconvenient questions about how far does the rot go? Can Nvidia's chips be trusted? And it's not like chipmakers haven't been calling that into question for years now. Three years ago, Cerebras's CEO called out Jensen as un-American and an AI arms dealer for insisting on selling chips to China. I fully expect there's going to be quite a bit of underhanded dealing and playing off of all of this conspiracy going forward.
Brandon (19:15)
Like I was saying earlier, it was there before, those claims of Nvidia being a risk or there being conspiracies surrounding their Chinese market. I can't imagine that slows down now.
Tobias Mann (19:32)
That said, there are an awful lot of firearms manufacturers accused of being arms dealers during prior wars that are still kicking, so…
Brandon (19:41)
Absolutely. Absolutely. I guess the continued success of any of these companies again sort of hinges on whether frontier models continue to be the de facto norm in AI computing, right? Because that could change as well.
Tobias Mann (19:56)
There's a lot of things that could happen. The reality is the bubble is here because there is a disconnect between the actual value of these goods and services and the prescribed value of these goods and services. If that gets brought back to reality, the bubble pops. If it happens quickly, it's going to be tremendously damaging to economies. If it happens slowly, things could play out much cleaner. But, there's a lot of people who are concerned right now that Huang may be the one that pops the bubble.
Brandon (20:36)
Pops the bubble. That's what I was going to ask next was basically, is this a bubble-popping moment for AI? Does Jensen's ouster, is Jensen's ouster the pin that pops this bubble? Do either of you have any thoughts then?
Tobias Mann (20:48)
I'm going to let Dan answer that question first.
Dan Robinson (20:52)
Well, people have been asking questions about Nvidia's financing of a lot of the AI industry.
Brandon (21:02)
They're kind of key to a lot of the circular investing that's been going on.
Dan Robinson (21:06)
So you do start to wonder how many of these AI companies, especially like the rent-a-GPU cloud companies, how many of those are being kept going by investment from Nvidia and whether that's going to continue with the CFO in charge. And also whoever happens to be the follow-on chief executive, will they continue to do that kind of thing? If so, maybe that could start to let some of the air out of the AI bubble.
Brandon (21:50)
Those neoclouds are always kind of the ones that are canaries in the coal mine, so to speak, in this industry.
Tobias Mann (21:58)
That's what we think. I would agree with what Dan had to add there. I would note there's two elements that I'm sort of watching. Whoever ends up being the new CEO is going to have an awful lot of say in terms of how additional funding and investment takes place. Nvidia invests in neoclouds in two ways. One of them is direct funding. They'll invest money and it's tied to a procurement agreement. So if they're going to use that money to buy our chips, we're going to help facilitate that. The other side of it is that GPUs are so valuable and in such demand. It's particularly Nvidia's GPUs, which are seen by the market as being the most valuable and most potent, which isn't necessarily true as I mentioned earlier. If that valuation of the prescribed value of those GPUs declined, getting debt backed by those GPUs, kind of like collateral on an auto loan, might get much, much more difficult because the reason that neoclouds are able to go get billions of dollars in loans is because the lenders believe that the assets are still going to be worth something if the company isn't successful.
Brandon (23:27)
And llke you said, there could just be a first domino to get pushed. What do you guys think about the potential of Tim Cook being Jensen's replacement as Donald Trump has suggested. Tim Apple.
Dan Robinson (23:44)
Tim Apple.
Brandon (23:46)
Becomes Tim Nvidia? Tim Vidia? No.
Dan Robinson (23:52)
I don't know.
Tobias Mann (23:53)
My bet would be on Jensen making a surprise return before Tim.
Brandon (23:58)
I don't think that Tim Cook wants to take over at Nvidia.
Tobias Mann (24:01)
Well not just that, but Tim's success was driven in large part by driving a logistics monster at Apple, ensuring that they always had the materials that they needed and anticipating market changes before they happened so that things like when the memory crisis hit, Apple's prices were insulated for several months and they could gain sales because people wouldn't buy others' products because they were getting more expensive faster. I don't think that that is something that Nvidia needs right now.
Brandon (24:44)
That's totally fair. Also I feel like Tim Cook is a bit I mean, we've called him bland on several occasions and I feel like the risk-taking nature may not be there in the same way that Jensen's kind of jet setting super hip tech man vibe gave Nvidia. I don't know. Dan, what are your thoughts?
Dan Robinson (25:07)
I don't know. I don't know about Tim Cook. Like you say. I'm not going to say he's bland because that's unkind.
Brandon (25:17)
You're more diplomatic than we are, I guess.
Dan Robinson (25:19)
About some of the other names that have been put forward by people taking odds on this have been Lisa Su. But I couldn't see her leaving AMD really.
Brandon (25:32)
Tobias shaking his head over here, yeah. No.
Dan Robinson (25:35)
Or Rene Haas from Arm. Who does share a lot of the same views on things as Jensen Huang did, especially on AI and on the physical AI on robots being the next big thing. So. I don't know, that could position him as a potential candidate. Who knows, would you want to step into the potential mess that's been left by Jensen's departure?
Tobias Mann (26:14)
I think it's pretty telling that most of the candidates outside of Nvidia are already the CEO of major technology companies. There's like Lip-Bu Tan at Intel, I could see, if he wasn't already at Intel, making that jump. And maybe he will. Intel is still kind of a mess. But it is somewhat telling that there's just not many good candidates to fill that role that aren't already in that position.
Brandon (26:48)
Jensen Huang is, I feel like it's akin to Apple hunting for a successor to Steve Jobs, right? Apple was Steve Jobs for many years, in the same way that Jensen lived 33 years at the helm of this company, from it being a tiny startup to being the behemoth that it is now. I don't want to call it a cult of personality, but Jensen Huang, like you said before, was in many ways Nvidia. This is a huge shift for this company, especially with the amount of influence that it has. So any thoughts on what might be next for Jensen? Does he come out with a new startup? Is this an OpenAI sort of thing where the storm ebbs and he returns? Any thoughts?
Tobias Mann (27:34)
Dan, you want to take that one?
Dan Robinson (27:35)
That's difficult to say. He might go for a startup. According to the company's proxy statement that was filed some time ago, he could be in for a $264 million payout for his termination which... He may just want to go to Hawaii and retire.
Brandon (28:09)
Maybe, I know Tobias, I think you mentioned in your story that you don't think that this is the end of his career in the space. He's just got too much get up and go and too much ambition to just walk away.
Tobias Mann (28:26)
There's the ambition and the cult of personality. So I would imagine he's probably going to disappear from the spotlight for a little while. Might be months, it might be a couple years. Without knowing the terms of his contract and his mobility in terms of being able to stand up a competing startup. Assuming that's what he even wants to do. But his cult of personality is such that, if he did start a new company, there would be very little stopping folks from jumping ship very quickly from Nvidia to come work for him. They may have to go take a two-year hiatus, but right now would not be the worst time to cash out some Nvidia stock to take a break for two years and go back to work for the old boss either.
Dan Robinson (29:20)
He might be on dangerous territory if he tried to go into a startup making AI accelerators though.
Tobias Mann (29:31)
Who knows? Maybe he will trade spaces with Lip-Bu and go run Intel.
Brandon (29:39)
Just a good old shuffling of leadership here.
Dan Robinson (29:43)
He was actually joking with the press and analysts at GTC that Intel has had 10 CEOs in the last 30 years and Nvidia has only had one. Famous last words, I guess.
Brandon (30:05)
Well, we'll see. It's going to be an interesting few weeks to kind of watch this unfold. I know he's testifying before Congress soon on the whole matter. The investigations into the Alice Huang matter aren't done. And I'm sure there's more revelations and shakeups to come in Nvidia. And we will be right here to write about it at the register and talk about it on The Kettle. Thanks for joining us and we will talk to you soon. ®
Marvell and Penguin Solutions aim to end the RAMpocalypse with photonic memory fabric
Celestial AI’s dream of high-bandwidth memory (HBM) at a distance will soon be a reality as part of a new collaboration between Marvell Technology and system integrator Penguin Solutions to build a photonic memory fabric.
Marvell acquired the underlying technology for the fabric from Celestial AI for $3.5 billion in February. It comprises a memory expansion module with two stacks of HBM3e totaling 72 GB along with support for eight channels of DDR5 stacked atop a silicon photonic interposer with 7.2 Tbps (900 GB/s) of bidirectional bandwidth.
“What we provide in this appliance is what we call a Photonic Fabric memory module… We provide these modules to Penguin, and they build the box, do all the power management, do the software, and all the management layer,” Preet Virk, senior VP of Marvell’s photonic fabric biz, tells El Reg.
Penguin’s first appliances based on the tech will support 16 of these memory modules for up to 16 TB of DDR5 and 1 TB of HBM, with future systems capable of scaling to twice the DDR5 capacity.
With 57.6 Tbps of bandwidth distributed via a Marvell-designed optical fiber shuffle, each appliance supports up to 16 servers either via a network interface card, or via co-packaged (CPO) or near-packaged optical (NPO) modules.
If you’re not familiar, CPO modules are typically integrated directly into the compute or switch ASIC, while NPO modules are a sort of half measure. They connect electrically to the host, whether that’s a CPU, GPU, or some other ASIC, and the rest of the fabric optically. NPO is expected to play a major role in the next generation of rack-scale systems from Nvidia and AMD, as it provides better serviceability and lower blast radius than baking photonic chiplets directly into CPUs or GPUs.
Through a software layer, which we’re told was developed by Penguin, each system can access the entire shared memory space, similar to how systems connecting to network-attached storage would have access to the appliance's full capacity.
As we understand it, a lot of this functionality is made possible today through CXL 3.1 compatible NICs and host CPUs, like AMD’s shiny new Venice CPUs. However, for NPO- and CPO-based memory, Virk tells us that CXL can be bypassed entirely.
The implications of this tech should be obvious. If memory can be shared and pooled across multiple servers, each of those servers doesn't need as much of its own DRAM, possibly saving some money.
In addition to facilitating communications, Virk says, the software also allows for several memory tiering modes by which customers can expose the resources to connected systems. For the most part, these follow typical CXL memory tiering conventions, with the exception of an HBM cache-through mode that Virk says will enable users to extract the performance of HBM with the capacity of DDR5.
This mode essentially treats the HBM as a read/write cache. Frequently accessed files are written to and read from HBM first before being pushed into standard DDR5 as they get stale.
Marvell and Penguin are positioning the new memory appliance as an alternative to storage-based key value (KV) cache offload. AI inference generates large KV caches, which are used to track model state across chat sessions, and long-running agents. Depending on the context length, these caches can be both computationally expensive to recompute and large to keep in memory.
Because of this, inactive KV caches are often offloaded to DRAM or flash memory, since it’s faster to store and copy them than to recompute them if and when a session is restored. Marvell contends that by offloading KV caches from HBM directly to these memory appliances, customers can achieve better performance and higher memory utilization.
Beyond AI-adjacent workloads, this technology also has potential to reduce the amount of DRAM required for individual systems, a concept we explored in greater depth earlier this year, and one that Meta has already embraced in its own datacenters.
Marvell expects Penguin’s first generation of Photonic Fabric memory expansion systems to arrive sometime in the second half of next year. ®
Los Alamos boffins to put NextSilicon's Maverick-2 through natural selection in Darwin testbed
The US Department of Energy (DoE) is deepening its relationship with HPC-first chip startup NextSilicon.
During Supercomputing 2026 in Chicago on Monday, the Israeli startup revealed that the agency had deployed its second Maverick-2 cluster in the Darwin testbed at Los Alamos National Laboratory (LANL).
The deployment comes just six months after Sandia National Laboratories powered on NextSilicon’s first supercomputer, which features 64 nodes and 128 of the company's “runtime-configurable” accelerators.
NextSilicon’s latest DoE system is comparatively smaller, comprising just four accelerators, yet for LANL, the system is arguably more important despite its size.
Darwin, LANL’s chief scientist Galen Shipman, tells El Reg is a “long-lived testbed environment where we bring in newer technologies for evaluation.”
“It gives us the ability to work with the technology much more quickly, and integrate it with other components of our architecture in our environment,” he explained.
Smaller systems like the Maverick-2 cluster now deployed in Darwin allow LANL to validate and port a wider array of codebases to the chips, identify problems, and work with NextSilicon to not only address them, but directly influence the capabilities of future chips, Shipman explained.
In other words, Los Alamos’ interest in Maverick-2 is the first step toward larger scale deployments of NextSilicon silicon in future DoE supers. For this, hands-on access is kind of important as many of LANL’s most important workloads can’t exactly be run on public systems.
A compute chameleonLANL is one of several labs tasked with ensuring that the US’ nuclear deterrent remains effective decades after underground testing was banned. The lab is also responsible for running simulations on energy sources like fuel candidates for future fusion reactors. Both of these workloads, Shipman tells us, fall into a category called multi-physics simulations.
These workloads aren’t really one workload, but multiple simulations that can have wildly different compute requirements.
A “good example is inertial confinement fusion. At the National Ignition Facility at Lawrence Livermore National Laboratory, you're using lasers to implode a fuel capsule, and the goal is to generate more energy through yield than you do from depositing energy to the capsule,” Shipman explained. “That's inherently multiphysics.”
There are simulations for the lasers, the hydrodynamics within the fuel capsule, and the chemistry and material interactions taking place during the fusion process. The problem, Shipman explains, is that a system that’s ideal for one of those simulations may not be well suited to others. “Identifying one specific platform that is optimal for that type of simulation is challenging.”
CPUs, he says, are well suited to workloads with large serial sections, irregular access patterns, and jobs that benefit from deep cache hierarchies. Meanwhile, GPUs are fantastic for dense, matrix-multiplication heavy workloads.
"Where NextSilicon fits in is being able to reconfigure to match the workload, on demand,” Shipman said. “The hope is that we won't have to make these rigid determinations upfront like, ‘we need this much CPU or we need this much GPU.’ We can use a much more adaptable, reconfigurable architecture to meet the workflow where it is at.”
A closer look at Maverick 2Despite some visual similarities to Nvidia’s current crop of GPUs, Maverick-2 is not a GPU. Instead it employs a dataflow architecture which overlaps communication and computation. Rather than waiting for load-store operations to shuffle data around as you would with a conventional CPU or GPU, the chips themselves comprise a grid of arithmetic logic units (ALUs) which are configured to perform an operation like addition or multiplication in the order that they’re needed.
Data is computed as it moves from one ALU to the next, a bit like how a ball moves through a Rube Goldberg contraption.
This can make dataflow architectures extremely efficient in bandwidth-bound scenarios, which has implications not only for HPC and scientific computing but also for AI. It’s why we’re seeing dataflow accelerators, like Nvidia’s Groq, Cerebras’ WSE-3, and SambaNova’s SN50, being deployed as decode accelerators in AI inference clusters. Where NextSilicon stands out is that its designs, so far, have been specifically tailored toward sparse double-precision FP64 workloads.
Real FLOPS are rarely advertisedProgramming dataflow architectures, historically, has been a challenging endeavor. To overcome this, NextSilicon claims to have developed a compiler that can decompose any existing C, Python, Fortran, or CUDA codebases into a graph that can run directly on its chips.
So far, NextSilicon has managed to validate its systems against the High Performance Conjugate Gradients (HPCG) benchmark, the LAMMPS molecular dynamics test suite, and the Sparta Monte Carlo simulation suite.
But as NextSilicon CEO Elad Raz notes, the roofline performance of Maverick-2 is different from one workload to another, and so the company has opted not to advertise peak theoretical performance the same way GPU vendors like Nvidia or AMD have traditionally done.
That’s not to say NextSilicon is shy about Maverick’s performance characteristics. Rather than advertising peak performance that the chip may not be able to hit in the workloads its customers actually care most about, it published performance figures across a swath of HPC benchmarks. For example, in HPCG, one of the closest watched HPC benchmarks after High Performance Linpack (HPL), NextSilicon says a single Maverick achieves 600 gigaFLOPS of performance, which it says is on par with contemporary GPUs but at half the power.
That’s just fine and dandy as far as Shipman is concerned. He tells us peak advertised performance is rarely a useful proxy for real-world performance.
“The thing that my team spends the most time on is how do we generate representative proxies, representative kernels; how do we generate specifications that communicate our needs to the broader ecosystem so that vendors like NextSilicon can meet us and co-design with us.” ®
How many times are you paying to store the same file?
Pick any busy project folder and count the copies. There's one on the file server, one in a cloud bucket and one on each laptop that needed it at a usable speed. If a team got fed up with sync and built its own workaround, there's another in there too. The storage bill pays for every one of them.
The Register and LucidLink are hosting senior technology, infrastructure and data leaders at the Covent Garden Hotel on Tuesday 24 November to compare how they're bringing that number down. Joe Fay of The Register is in the chair.
Nobody keeps duplicates for fun. They're what people do when files have to move between sites faster than sync allows. Every copy IT can't see into is a security question as well as a cost, so the duplicates you don't know about matter most.
What the table wants to hear is how others got the copies down without ripping out storage they've already paid for, or the governance built around it. The evening runs under the Chatham House Rule, so people say what worked and what didn't. Alan Marshall, Strategic Director, EMEA at LucidLink, joins a small group of peers. Nobody presents.
Drinks are at 6.30pm and dinner around one table runs until 9pm. Places are limited.
Request a place at the 24 November dinner
Places go by invitation and each one is confirmed personally.
Sponsored by LucidLink.
After dodging death, battery startup Our Next Energy is now going public
Bedrock Energy raises $24M to build more AI-enabled geothermal drill rigs for enterprise buildings
Transcript: MAGA Fractures on Hegseth as Trump Iran Humiliation Grows
The following is a lightly edited transcript of the November 16 episode of The Daily Blast podcast. Listen to it here.
Greg Sargent: This is The Daily Blast from The New Republic, produced and presented by the DSR Network. I’m your host, Greg Sargent.
Donald Trump has finally figured out how to look as if he’s won a world-historical triumph over Iran: by doing nothing. Trump has decided not to restart military action, and not to make a deal either. Instead, he’ll wait until the regime collapses.
This comes as officials are suggesting in leaks that Trump is siding with Hegseth in this regard, and against others in the administration who want a deal, including JD Vance. Meanwhile, numerous MAGA figures are rallying behind Hegseth.
The situation is deeply humiliating for Trump. He’s still gotten nothing from Iran, yet he and much of MAGA are basically following Hegseth over a cliff. Is there any way out? We’re making sense out of all of it with former national security official Emily Horne. Thanks for coming on, Emily.
Emily Horne: Great to be with you, Greg.
Sargent: So let’s just quickly situate where we are. Trump now appears to be ruling out new military action against Iran, but his new strategy on Iran is just to wait until economic sanctions topple the Iranian regime. Can you quickly recap where we are, kind of?
Horne: Yeah, unfortunately, I don’t have a lot to add to what you just laid down, because we are in a very strange holding pattern, where there are ships that are moving through the Strait of Hormuz, but they are reliant on U.S. military support in order to do so. So it’s not accurate to say that the strait is open, as President Trump claims.
I don’t think open would include a personal escort by a U.S. aircraft carrier in order to move a shipload of goods or oil through what was previously an open commercial maritime pathway.
That said it’s still better than having open conflict, which is what the Trump administration had been rumor-mongering in the run-up to the midterms, that they were prepared to start massive strikes on both Iranian military and civilian infrastructure.
And so I don’t want to give anyone credit for avoiding yet another illegal war, but it’s worth noting that at least that scenario has not come to pass. The one thing that we know for sure is that the Iranians and the hard-liners and the IRGC really have all the cards right now, and Trump has really nothing to play.
Sargent: In fact, the New York Times’ David Sanger had an absolutely withering piece pointing out how absurd this all is. He noted that it’s not clear that Iran will crack under the sanctions all that soon.
And while oil is coming out of the Strait of Hormuz right now, as you pointed out, we’re spending enormous amounts of money on the U.S. military to enable that to happen. And it’s hard to know how sustainable that is.
And then on top of that, the enriched uranium is still there underground. This is the whole reason Trump went to war. He then claimed he had obliterated the nuclear threat entirely.
Now Trump and his advisers have flipped to saying, it’s no biggie that the uranium is still there, we got this. Trump is winning because he says so. Is that basically the situation we’re in?
Horne: You know, the analogy that I like to use sometimes when you’re talking about Trump and Iran is, do you ever watch South Park? You remember the underpants gnomes? This is Trump’s whole approach to Iran. The underpants gnomes were: Step one, collect underpants. Step two, question mark. Step three, profit. And that’s always been Trump’s approach to Iran.
And to be fair, it’s not just Trump. This has been a lot of hard-liner hawk positions toward Iran, where if you just take unspecified military action toward Iran, then question mark, question mark, question mark, and Iran will never have a nuclear weapon. Trump is just the latest flavor of that particular approach to Iran that we’ve seen for a long time.
Look, the fact remains, at the end of the Obama administration, the United States had a deal with the Iranian regime that would have kept Iran’s nuclear ambitions in check. We were on that path, and Donald Trump walked away from all of that hard work in the first administration that he won.
And everything that we are living through now is the logical outcome of that very foolhardy and shortsighted and, frankly, selfish decision that he made in his first administration to abandon diplomacy.
Sargent: There was even some leaking a while back that Trump was desperately urging his advisers to find some way, any way at all, for him to say that he had gotten more out of Iran than Obama did. And of course nobody could find anything. And so there you have it.
Horne: Yeah, and it turns out you can’t yada-yada your way out of a war. There’s been a lot that Trump has been able to escape consequences on, because he’s a master at manipulating media, and he’s a master at getting people to do his bidding if it makes him feel better about himself.
But when you are talking about deploying aircraft carriers and tens of thousands of troops to adjust our force posture across the Middle East, and trying to negotiate with allies for how they are going to protect their own assets from Iranian missile and drone attacks, you can’t just close your eyes and will it all to go away, or make someone bring you a better talking point that lets you feel bigger and more badass than the president who preceded you. It just doesn’t work like that.
Sargent: No, it doesn’t. And the Post actually shed some light with some reporting on the situation. One official briefed on ongoing discussions leaked word that JD Vance, special envoy Steve Witkoff, and Trump’s son-in-law Jared Kushner privately pushed Trump to make a deal, but Hegseth appeared opposed to one.
Emily, if I’m reading that right, Trump, by not making a deal and sticking to this waiting game of his, is essentially siding with Hegseth, right?
Horne: So I think we have to be really careful when we’re looking at what is essentially a Washington political media that is just a cesspool of people leaking against each other in the Trump administration right now, especially in the aftermath of the midterms.
JD Vance took almost a solid 24 hours after the shellacking of the blue wave on midterms eve to come out, and the first thing that he did was say the Iran war is responsible for Republicans’ losses on election night. And by doing so, he’s trying to differentiate himself from the very administration of which he is literally the vice president.
And so I think what we can know is that, one, people are going to keep leaking on each other, and they’re going to keep doing so with an eye less toward trying to curry favor with Trump, which was their goal for the first two years.
And now that Trump is a lame duck, it’s going to be much more with an eye toward what are their own future paths in Republican politics.
Sargent: Yes. I think we can probably conclude, based on Hegseth’s public posture, which is just nothing but malevolent fury and aggressiveness in basically every guise—we could probably conclude from that that Hegseth really doesn’t want to make a deal. He doesn’t want an easy way out of this thing that looks like any kind of capitulation.
Horne: As long as Trump insists the Iran war is going great, he’s got to stand by the guy he has running it. And to abandon Hegseth now, or to set Hegseth aside, as was rumored immediately after the midterms that Trump was considering doing, would be to admit that he made a mistake with the illegal war against Iran, that his attacks on Iran are not working, that there is no diplomatic or military end in sight in the foreseeable future, and that this is a catastrophe entirely of his own making.
And of course, Trump can’t admit that. He has to insist that the war is going great, because he’s never admitted in his life that he’s made a mistake. And so why would he start now?
Sargent: Well, that’s exactly why MAGA figures are rallying aggressively around Hegseth right now, and they’re in a rage at Hegseth’s critics. Media Matters flagged a bunch of examples. Here’s Fox’s Mark Levin ripping Republican critics of Hegseth.
Mark Levin (voiceover): What a bunch of cowards. We see this with all the isolationists. Pete Hegseth is a patriot.
Here’s influencer Jack Posobiec.
Jack Posobiec (voiceover): And I see a lot of people trying to blame all this on Secretary Hegseth. Don’t think I don’t see it out there. Don’t think I don’t see what’s going on. The blame game’s coming around. No, no, no, no. No, no, no, no. We’re not playing that game. We’re not playing that game. Pete Hegseth is an absolute patriot, and we’re going to leave it at that.
And here’s Hugh Hewitt.
Hugh Hewitt (voiceover): Pete Hegseth has done a great job for two years of executing the president’s orders. The president’s commander in chief. President Trump has done this well. And I think he ought to stick with his national security team, which means Rubio and Hegseth.
Sargent: Emily, these guys are really just following Hegseth over a cliff here, aren’t they?
Horne: I think a lot of them are. And you’re seeing a lot of defense of Hegseth come, unsurprisingly, from his former colleagues at Fox News—a lot of people with whom he has a longstanding personal and professional relationship, but also people who want to protect their access to the Trump administration. They’re defending one of their own, but they’re also defending their turf.
I think, though, it’s important to watch very closely where we’re starting to see some cracks emerging in this coalition. That’s something I’m keeping a really close eye on, especially as we are seeing more and more voices recognize the mandate that Democrats were given in the midterm elections, a lot of which came out of frustration with the Iran war.
And the ensuing rising cost of living—cost of gas, cost of groceries, et cetera—that have come from, yes, tariffs and other foolhardy economic choices, but also this conflict that is now entering its eighth month.
Sargent: Well, let’s just bear down on how humiliating this situation actually is for Trump. One expert at the Carnegie Endowment for International Peace told the New York Times this: “Trump began the war thinking he would be a hero. He’s now trying to figure out how to end the war without looking like a loser.”
Another expert, at the Middle East Institute in Washington, said that if Trump leaves the situation as it is, which is what he’s doing, that’s, “a victory for Iran.” Emily, is that basically the situation, the way those two experts described it?
Horne: It’s interesting. I hear that first quote and I think, who the heck was he listening to that was telling him that this would be a quick and easy victory? Because no Iran expert I know thought it would be.
Look, I think there were people in Trump’s inner circle that were telling him Venezuela was a big success: We got in, we got Maduro, we got out, no muss, no fuss. You hear Trump echoing that line of reasoning in his own defense of Iran.
Anyone who knows the tiniest thing about Iran specifically, or the Middle East broadly, about the history of the U.S.-Iran-Israel relationship—you know, take your pick of framing—but anyone who knows anything about that region would have told Trump, had they been asked, this is not going to be the quick and easy win that your inner circle is telling you. And a lot of people, I think, did try to tell him that in various ways.
Iran has all of the leverage right now. They control access to the Strait of Hormuz. They control the pace and the tempo of proxy strikes across the region. This is terrible to say, but that they do not care about the suffering of their own people is a strategic asset for them in this moment.
Because they are perfectly willing to let Iranian civilians bear the brunt of the humanitarian catastrophe that is coming to them via not just strikes, but the effects of sanctions and other forms of economic punishment that don’t impact the elites of the regime, but very much harm the life of the average, everyday Iranian.
So yeah, they have all of the leverage right now. And, you know, we don’t know what their demands have been exactly in the diplomatic negotiations that have happened to date, but I would be willing to venture they’re probably pretty maximalist.
They probably run along some line of, let us maintain control of the Strait of Hormuz, or at the very least impose tolls, which they have never done before. And now it’s hard to imagine a world in which they would not derive tremendous economic benefit from controlling access to the strait.
That bell cannot be unrung, and that’s entirely thanks to Donald Trump. But more broadly, that they would tell Trump pound sand on anything that we want to do vis-à-vis our nuclear program, and we’re just going to keep doing what we’re doing, and you’re just going to have to deal with it.
Sargent: Just to refer back to what one of those experts said, the fact that Trump is looking for a way to appear not to be a loser just is the real story in a nutshell, and everybody knows it.
And the entire national security apparatus around Trump, and the entire GOP, they’re all required to pretend this isn’t what’s happening. They must maintain the illusion that Trump is currently winning a glorious triumph. It’s almost like the North Korean dictator has no clothes, right?
Horne: Well, and it’s really interesting, because we are seeing a split that is emerging around Iran. We’ve talked earlier about the Fox News crowd, who need to protect their access to and relationship with Trump.
But there’s a different audience, a podcast and independent content creator and radio audience—they have their own audiences. They are young. They are more anti-war. They were people who voted for Trump in 2024 because of his promise of no new foreign wars and lower costs. They are not dyed-in-the-wool MAGA. They do not maintain the same loyalties. And a lot of them turned out for Democrats in the midterms this year.
And so you’re starting to see the influencers and content creators and hosts who reach primarily those audiences start to sing a very different tune than the Fox News crowd. Because this is not an ideological split for them. This is a business model split.
Sargent: Well, what’s the way out? Is it just essentially waiting for Democrats to take over and hoping, really, really hoping, that they make good on their threat so far to defund the Iran war? Is that the only way this gets resolved?
Horne: Look, Democrats were elected last week to do two things: affordability and accountability. And Iran sits right in the middle of those two priorities. So Democrats defunding Trump’s illegal war in Iran are going to be able to help bring down gas prices, the prices of goods and services.
But they’re also going to get America out of a war that it has no business being in. They’re going to cut back on the war profiteers who have figured out, whether through the prediction markets or through their investments in oil futures, how to personally profit from this war, while the average person can’t afford a pack of hamburger at the grocery store the way they used to.
I think people are really tired of the sense that a small group of elites are figuring out how to profit from the very kind of wars that Trump said he was not going to get us into anymore. And people are impatient for change. And defunding the Iran war is a great place to start delivering on the request for those changes.
Sargent: OK. When Democrats take over in January, do you think Pete Hegseth will still be there? Yes or no?
Horne: Yeah, probably. Look, I think a lot of things about Massie’s attempt to impeach Hegseth, and we could have a whole separate conversation on that. But the top line is that it actually wound up really securing Hegseth’s job, at least for the next couple of months.
Because here’s what: This is an administration where being under attack is their comfort zone, and something like an impeachment attempt actually becomes a credential. In Trump world, getting impeached proves that you’re on the right side. And before the midterms, Trump was rumored to be asking his associates, I don’t know about this Hegseth guy, what do you think about him? Which is usually a precursor to someone getting fired in a very public and embarrassing way.
Instead, Thomas Massie impeached Pete Hegseth, or tried to, and he gave him a lifeline and a storyline—a storyline where he was able to hold a news conference at the Pentagon and say, I’m under attack. There he was able to stand behind the troops and use our military as a shield.
And he also probably overreached by introducing eight articles of impeachment, some of which Republicans may have agreed with, but some of which they would have said, I don’t know about that. And so if and when there is a Democratic impeachment movement against Pete Hegseth, I would expect that it would be more limited in scope, and it would be far more competently executed.
Sargent: So I guess that really means we’re going to have to absolutely hold Democrats to their pledge to defund this thing. Otherwise, we’re really screwed. But I think they probably will do it. Emily Horne, awesome to have you on.
Folks, if you want more of Emily, you can check out her Substack, which is called Spin Class. Emily, thanks so much for coming on.
Horne: Thanks, Greg. Great to be with you.
SpaceX catches Starship on the first try
SpaceX caught its Starship upper stage returning from orbit on its first attempt, but the Super Heavy booster ended its flight in an explosion after recovery was aborted.
The catch was a remarkable achievement: Starship returned to the same launch tower it had left almost a day earlier, settling into its giant mechanical arms.
Flight 15 began well, with the 33 Raptor engines of the Super Heavy booster launching the stack from Pad 2 at Starbase, Texas. Hot staging followed, and the booster performed its directional flip maneuver.
Thirty-two Raptor engines relit for the boostback burn, intended to return Super Heavy to the tower for another catch. The booster had managed the maneuver on flight tests 5, 7, and 8, but this time something went wrong. "The boostback burn ended early due to low fuel mass," SpaceX said.
The booster automatically aborted its return to the tower and diverted offshore. It completed a landing burn, briefly hovering above the Gulf before its engines shut down and the flight termination system delivered an explosive finale.
Starship reached orbit under the power of its six Raptor engines and deployed 26 Starlink V3 satellites. Unlike the previous flight's brief orbital stint, this mission lasted almost a day.
Following a deorbit burn, Starship returned to the launch site 22 and a half hours after it left.
Many more development flights remain before SpaceX can deliver its promised fully reusable launch system. NASA's Space Shuttle came close, but the large external tank was expended with every launch, and the orbiters never matched the cadence SpaceX reckons it can achieve with Starship.
SpaceX boss Elon Musk admitted that "many small fixes are needed to achieve immediate reusability (like an airplane)."
Cadence will be key if Starship is going to meet NASA's needs for the Artemis missions. Somewhere between 6 and 22 launches (depending on who you ask) will be required for a lunar mission, and SpaceX has yet to demonstrate in-space propellant transfer between vehicles.
Next year's Artemis III mission also calls for docking tests between NASA's Orion spacecraft and a test version of SpaceX's Starship Human Landing System (HLS) in low Earth orbit.
The booster fumble shouldn't obscure what the engineering teams achieved. Now comes the small matter of doing it again – preferably with both stages intact. ®
Freshworks alums raise $53M to take on Salesforce, HubSpot with AI-native CRM
Michiganders Voted to Get Corporate Money Out of Politics. Now Utilities Are Suing.
This story was originally published by WIRED and is reproduced here as part of the Climate Desk collaboration.
Michigan residents overwhelmingly passed a midterm ballot measure intended to restrict the ability of utilities and state contractors to influence politics. Now, the utilities are fighting back.
In a lawsuit filed on November 5, the two major utilities in Michigan, Consumers Energy and DTE, argue that the measure—dubbed Proposal 2—threatens their First and 14th Amendment rights.
The ballot measure, which passed with 81 percent of the vote, amends the state’s campaign finance laws to put restrictions on donations for both utilities and companies with more than $250,000 annually in state and local government contracts. The new language explicitly bans political donations from utilities, contractors, and high-level employees, lobbyists, and their immediate family members. The ballot measure also tightens the state’s regulations around disclosures for political ads, mailers, and other forms of communication.
Michigan was part of the blue tsunami election where energy prices and government corruption—and data centers—played a huge role in getting Democrats elected and getting ballot measures like this passed. The state is a major hub for data center development, with at least 26 projects planned across the state as of July.
“The language of Proposal 2 was carefully crafted with constitutionality in mind.”
The utilities have supported that development: the industry has lobbied in the past for tax breaks for data centers, and DTE has proposed building two new massive gas plants to handle potential data center load, at a cost of $6.4 billion, with consumers potentially on the hook. Michiganders have seen energy bills increase by 10 percent over the past 12 months, the fourth-largest increase in the US, according to Energy Bill Lab.
“Michigan really is ground zero right now for a utility accountability wave that is quite national,” says David Pomerantz, the director of the Energy & Policy Institute, a nonprofit utility watchdog.
The ballot initiative, the utilities say in the lawsuit, was drawn up as “a means of excluding Plaintiffs from critical aspects of the political debate.”
Consumers spokesperson Katie Carey writes in an email that the proposal “unfairly targets a Michigan business, its Michigan employees, and their family members from exercising their constitutionally protected rights to participate in the political process, while allowing out-of-state special interest groups and billionaires to continue flooding Michigan with political spending. Michiganders should not lose their First Amendment rights because of where they work.”
DTE did not respond to WIRED’s request for comment.
Utilities are regulated by state-level bodies, either elected or appointed by the governor, which Pomerantz says means that there’s much more incentive to participate in pay-to-play schemes. Corruption scandals involving utilities have rocked neighboring states like Ohio and Illinois. In the lawsuit, DTE and Consumers Energy say there is “no scandal, no exchange of official acts for money, no examples whatsoever.”
Daniel Weiner, the director of the Elections and Government program at the Brennan Center for Justice, says that the new language set up by the ballot proposal “moves Michigan toward being one of the states that more effectively regulates this kind of political giving, where there’s a particularly heightened risk of corruption.”
The lawsuit states that the new provisions “goes well beyond the laws of any other state.” Weiner points out that it’s unusual, but not unprecedented, for states to restrict or prohibit political donations from the family members of people affiliated with government contractors. New Jersey, Connecticut, and New Mexico have similar language in their laws. However, these laws make no specific mention of utilities. (Other states, including Georgia and Alabama, do have restrictions on utilities making political donations.)
“If you’re going to have an effective restriction on major state contractors pouring tons of money into campaigns, it cannot just be limited to the entity that holds the contract—that just creates an enormous loophole,” Weiner says, pointing to a recent ProPublica story about a government contractor who used a shell company to funnel donations to a super PAC affiliated with Senator Susan Collins of Maine.
A 2023 analysis from local outlet Bridge Michigan found organizations linked to the two utilities had given more than $2 million to state politicians via their PACs. But the companies do much of their political spending via dark money distributed by affiliated outside groups, Pomerantz says.
Michigan Energy First, a 501(c)4 linked to DTE, has received more than $51 million in donations since 2014, according to the Energy & Policy Institute; DTE gave $8 million to Michigan Energy First from September 2025 through August 2026. (DTE has repeatedly denied any connections between the company and this group, despite the fact that it has been staffed by company executives.)
The Detroit News reported last year that Michigan Energy First had given $550,000 to a group associated with Gov. Gretchen Whitmer and $750,000 to state House Speaker Matt Hall in 2024, amid an influx of data center proposals in the state. The donations also arrived right before the Public Utilities Commission approved a rate hike for customers. In total, Michigan Energy First had given $1.8 million to the Whitmer-affiliated group since its founding in 2020. (DTE told the outlet that any insinuation that it expected preferential treatment for its contributions was “blatantly false.”)
Michigan utilities spent a little over $200,000 in an attempt to defeat the measure.
“We are fully confident in the legality of Proposal 2 and look forward to seeing this out in court,” says Samuel Bagenstos, legal counsel for the groups behind the measure. “The language of Proposal 2 was carefully crafted with constitutionality in mind, and we are well prepared for this expected legal challenge. We intend to vigorously defend the will of Michiganders who overwhelmingly approved this measure just one week ago.”
Even if they had spent a massive amount of money to defeat the initiative, “they probably still would have lost, because they’re so unbelievably unpopular,” Pomerantz says.