>>> TradeGate Pre-Market Indications

DAX:
  • Scout24 (G24 TH) +1.9%
  • Siemens Energy (ENR TH) +1.1%
  • Daimler Truck (DTG TH) +0.9%
    • Daimler Truck Raised at Barclays on Truck Market Upturn (1)
  • MTU Aero (MTX TH) -0.8%
    • Airbus, MTU Aero Engines Form Hydrogen Fuel Cell Joint Venture
  • Fresenius Medical Care (FME TH) -0.9%
  • Vonovia (VNA TH) -1%
MDAX:
  • TKMS (TKMS TH) +3%
    • Carney and Merz Lay Ground for a New Transatlantic Alliance (1)
  • TUI (TUI1 TH) -1.4%
  • Talanx (TLX TH) -3.1%
    • Talanx Holder Meiji Yasuda Life Insurance Offers Shares: Terms
  • Lufthansa (LHA TH) -4.4%
    • Lufthansa Cut to Sell at Citi; PT 8.70 euros
SDAX:
  • Verbio SE (VBK TH) +3.9%
  • HelloFresh (HFG TH) +1.3%
  • Hamborner REIT (HABA TH) -0.8%
  • LPKF (LPK TH) -0.9%

>>> Europe Must Spend $3 Trillion to Cut Cloud, AI and Chip Reliance Europe may

Europe Must Spend $3 Trillion to Cut Cloud, AI and Chip Reliance

Europe may need about $3 trillion of investment through 2035 across cloud infrastructure, AI data centers, large-language-model training and other technologies to reduce reliance on US and Asian suppliers. The region will remain indispensable in select chipmaking tools, led by ASML in lithography and ASM International in equipment that builds chip layers, but tools alone aren't enough. Even if EU policy pairs chip funding with reliable power, common digital and security rules, faster procurement, private capital, talent and public-sector demand, selective resilience rather than full autonomy is likely. Europe can strengthen its power-and-analog chip, engineered wafer, low-power specialty logic, advanced packaging, quantum computing and cybersecurity positions, but will stay reliant on AI accelerator chips and LLM training.

The Information : Nvidia’s New Hedge Against Chip Competitors? Partner with Them

Nvidia’s New Hedge Against Chip Competitors? Partner with Them
The AI chip leader’s new deal with d-Matrix shows how Nvidia CEO Jensen Huang is adapting to a world with more competitors.

The Takeaway
  • Nvidia partners with AI chip rivals like d-Matrix to combine hardware.
  • Strategy helps Nvidia adapt to competition and generate new revenue.
  • Using chips made by different designers to handle the same AI task is a fairly new concept.

Nvidia has a plan to deal with the growing list of AI server chip competitors: partner with them.

In the latest example, Nvidia and AI chip startup d-Matrix are combining their respective hardware in a new system to power AI models, the companies told The Information exclusively.

The move comes a month after another AI server chip designer, SambaNova, said it worked with Nvidia to make its chips for powering AI models work in concert with Nvidia graphics processing units.

Nvidia is now hinting that more such deals could be in the works.

“I won’t pre-announce the others,” said Dion Harris, an Nvidia senior director of high-performance computing.

Companies such as d-Matrix and SambaNova don’t require partnerships for their chips to work with Nvidia’s—they can simply plug directly into an Nvidia GPU using readily available Ethernet cables. But with the recent partnerships, Nvidia engineers work with those other chip companies’ engineers to tweak software controlling the GPUs so they work better with the other chips. And a formal collaboration increases the chances a potential customer will want the combined system.

The strategy shows how CEO Jensen Huang has adapted to a growing array of potential threats to Nvidia’s dominant share of the AI chip market. By working with would-be rivals, Nvidia is putting itself in a position to generate revenue alongside them if the newbie chips succeed.

Nvidia “definitely had a change in strategy and their view of the world,” said Thomas Sohmers, chief technology officer at AI chip startup Positron, which isn’t currently working with Nvidia but is open to it. “Nvidia is playing nice and building out so that they’re actually part of that heterogeneous ecosystem rather than fighting it tooth and nail.”

The partnership strategy may also help it beat back past allegations that Nvidia pressured customers to stick with its hardware. The Department of Justice two years ago initiated an investigation into such allegations, The Information reported at the time, but no case has publicly materialized.

Harris said Nvidia has always wanted to be a broader AI infrastructure company. “We are not just a chip company,” he said.

The effort to broaden Nvidia’s approach accelerated last year when Nvidia said other chip developers would be able to run their chips in conjunction with NVLink, the company’s networking gear that connects AI servers to each other so they operate more efficiently. That could help Nvidia sell more of its networking gear, if not its chips.

“We would always rather sell something than nothing,” Harris said.

In December, Nvidia paid $20 billion to license technology and hire key staff from Groq, an AI inference chip designer that was trying to compete with Nvidia. The move was akin to an acquisition, though Groq has continued to operate as a standalone firm while Nvidia develops a server rack that combines Nvidia GPUs with Groq’s chips. It isn’t clear how much demand Nvidia is fielding for the product.

Then came the SambaNova partnership. Nvidia sees the startup as “more partners than people might expect,” SambaNova CEO Rodrigo Liang said.

To be sure, Nvidia in recent years has actually increased its share of the market for AI inference chips despite rising competition from the likes of Google and Amazon, according to The Information’s estimates. And Nvidia CEO Jensen Huang insists the company’s GPUs can do all inference more effectively than its competitors, though he isn’t leaving it to chance: Nvidia is using its powerful balance sheet to help more companies buy its expensive AI chips, including backstopping young cloud providers that want to rent out the chips.

But numerous other firms are also getting into the inference market or considering how to do so, including Microsoft, Meta, OpenAI and, most recently, Anthropic. That could change Nvidia’s share of the market down the line. (See The Information’s AI Chip Database.)


An OpenAI spokesperson said the company hasn’t yet decided whether it will work with Nvidia to operate its chips with OpenAI’s simultaneously. Nvidia has invested heavily in OpenAI and is in talks to backstop a giant data center project for OpenAI in Ohio.

“I think the world is going to a place where [Nvidia] is perfectly comfortable” with a mix of chips powering different pieces of the same AI task, said d-Matrix CEO Sid Sheth.
Upstart chip designers have historically positioned themselves as Nvidia killers. But given Nvidia’s substantial lead, they’re increasingly seeking to work in conjunction with Nvidia’s GPUs rather than replace them entirely, just as Nvidia is trying harder to work with those newer firms.

In fact, it was Nvidia that approached d-Matrix about partnering, according to a person who spoke to the startup about it. (Both companies are based in Santa Clara, Calif.)

AI developers such as OpenAI have long used different Nvidia chips to power the same AI task, after realizing less powerful GPUs were more efficient at handling certain workloads. But using chips made by different designers is a fairly new concept.

Nvidia isn’t the only chip developer to embrace that concept. Amazon, which develops Trainium chips used by Anthropic and soon OpenAI, said in March it would develop a combined AI inference server system with AI chip designer Cerebras.

Splitting up the process of running AI models between completely different types of chips is sometimes known as disaggregated inference. In the version involving Nvidia and SambaNova or Groq, Nvidia GPUs handle prefill, the most compute-intensive part of running an AI model, while the upstarts’ chip handle the rest, known as decode.

In d-Matrix’s case, its chips as well as Nvidia’s each handle some prefill and decode. D-Matrix focuses on powering a process known as speculative decoding that speeds up an AI model’s performance. When an AI customer asks a model to complete a task, the d-Matrix chip runs a small AI model that guesses what a bigger model’s answer to the customer should be, and the bigger model running on an Nvidia GPU then verifies and accepts the guesses.

‘Perfectly Comfortable’

“I think the world is going to a place where [Nvidia] is perfectly comfortable” with a mix of chips powering different pieces of the same AI task, said d-Matrix CEO Sid Sheth. “It’s not going to be a GPU-only world we live in moving forward.”

D-Matrix, founded in 2019, last raised $275 million at a $2 billion valuation in November but is in talks to fundraise again, according to three people with direct knowledge of the discussions.

Sheth said d-Matrix sets itself apart by putting compute and memory on the same chip and by not using the kind of high-bandwidth memory that Nvidia uses for its chips, and which is in short supply these days.

Taiwan Semiconductor Manufacturing Co., which produces AI chips for Nvidia and many others, started producing d-Matrix chips earlier this summer and plans to produce thousands of them per month by the end of this year, Sheth said.

The startup is currently generating single digit millions in revenue, he said, adding that he hopes the company’s chips will consume 30 to 40 megawatts of power in data centers next year—a relatively small sum—and run AI coding as well as voice and video.

Parasail, a young AI cloud provider based in San Mateo, Calif., will be the first company to purchase the combined Nvidia/D-Matrix server system. Parasail plans to rent out the system to its customers starting later this year. Parasail CEO Mike Henry said the combined server system was attractive because it helps his company avoid being too dependent on buying new Nvidia hardware.

The Information : SpaceXAI Plans to Launch New Model with Cursor as Soon as Wedn

SpaceXAI Plans to Launch New Model with Cursor as Soon as Wednesday
Elon Musk said last month that SpaceX and Tesla staff were testing the jointly developed model.

SpaceXAI and Cursor plan to launch their first jointly developed AI model as soon as Wednesday, according to a memo sent to staff.

The companies pushed back plans to launch the new model earlier this week in order to improve its efficiency, according to the memo. The model is expected to process information quickly, making it competitive in some respects with Anthropic’s Opus 4.8 and OpenAI’s GPT 5.5, the memo said.

SpaceX is acquiring Cursor for $60 billion in stock. SpaceX CEO Elon Musk said last month that SpaceX and Tesla staff were testing the jointly developed model. Cursor CEO Michael Truell said during an event for the coding app’s customers that Cursor was working on an AI model that had been trained from scratch on xAI’s Colossus data center and would compete directly with Anthropic and OpenAI models.

Cursor declined to comment. SpaceXAI did not respond to a request for comment.

The Information : China’s AI Lab Zhipu Weighs Custom Chip As Demand for its GLM

China’s AI Lab Zhipu Weighs Custom Chip As Demand for its GLM Model Soars

The Takeaway
  • Zhipu AI explores custom chip design due to soaring demand and U.S. export controls.
  • U.S. export controls and blacklist force Zhipu to seek Chinese chip partners.
  • GLM-5.2 model sees 27x token usage surge, straining Zhipu’s compute resources.

Zhipu AI, the Chinese AI lab behind the highly regarded GLM series of open-source AI models, is weighing designing its own AI chip as surging demand and U.S. export controls make computing resources a growing constraint, according to three people with direct knowledge of the plan.

The Beijing-based company, one of China’s leading AI labs, recently made preliminary inquiries with some Chinese chip design houses about the possibility of working on a bespoke AI processor optimized for running its models, according to two of the three people. The discussions are still at an early stage, and Zhipu has not selected a designer, according to the two people.

Zhipu currently relies on a combination of chips from Chinese tech giant Huawei, other locally made chips and some Nvidia chips. The company in January released an image-generation model trained entirely on Huawei chips, the first major image model to use only Chinese chips for training.

Still, Huawei’s chips have their own constraints. U.S. export controls continue to limit access to the advanced equipment Chinese foundries need to produce Huawei chips, while Zhipu must do extra software engineering work to run its models efficiently on Huawei’s hardware.

Zhipu, also known as Z.ai, is focused solely on finding a design partner in China because of its designation on a U.S. blacklist, two of the three people said. The U.S. Commerce Department has put Ziphu on a list of foreign companies prohibited from procuring U.S. technologies. If the project moves ahead, Zhipu is also likely to manufacture the chip at Chinese foundries, the two people added.

That blacklist doesn’t prevent U.S. companies from using Zhipu’s models, which are increasingly popular. Coinbase CEO Brian Armstrong, for instance, recently said in an X post that Coinbase was experimenting with GLM-5.2 as part of a broader effort to save money on AI costs. GLM-5.2 is also available on Oracle Cloud Infrastructure, allowing enterprise customers to deploy the open-weight model on Oracle computing clusters.

While anyone can download an open-source model for free and run it at their own cost, for AI application developers, the benefit of procuring access from the model maker directly is that they can keep the expense of running the models low as the model maker shoulders the bulk of the compute costs.

But the rising demand has strained Zhipu’s computing resources. On U.S. startup Vercel’s model aggregator platform, for instance, GLM-5.2 has been the fastest-growing model since its release last month, with daily token usage surging as much as 27 times during the first week of launch.

Zhipu, founded by researchers at the prestigious Tsinghua University in 2019, has been trying to generate more revenue from selling enterprise customers cloud-based access to its AI models, while reducing its reliance on deploying models for customers at their own data centers. This is because such on-site deployment usually generates just one-off revenue, because once a model is installed and up and running at a customer’s data center, little additional service will be required from the model maker. By comparison, selling cloud-based access lets the model maker charge for revenue on a continuous basis,

A custom AI processor would be a longer-term solution rather than a quick fix for Zhipu’s computing crunch. The company would need to build or expand a semiconductor team, choose a design partner, test the processor and adapt its software before its models could run on the chip. The whole endeavor could take more than two years.

Zhipu’s chip plan mirrors similar pushes by leading AI developers in the U.S. and China. These custom AI chips, known in the industry as application-specific integrated circuits, are processors designed to perform specific tasks tailored to specific models, as opposed to general-purpose AI chips like those Nvidia produces.

Major AI developers, including Google, OpenAI, ByteDance and Alibaba, have come up with their own custom chips to wean themselves off outside suppliers and reduce the costs of running their own models. The latest example is OpenAI’s Jalapeño, which it developed together with Broadcom and will deploy to run its GPT models by the end of this year. Other popular ASICs include Google’s tensor processing units and Amazon’s Trainium.

For Zhipu and other Chinese AI developers, U.S. export restrictions on advanced chips are another incentive to build their own bespoke chips.

Zhipu was the world’s first large language model developer to go public, and its stock price has soared more than 100 times since its debut on the Hong Kong stock exchange in January. With a market capitalization of around $100 billion, the company is planning a dual listing in the Shanghai Stock Exchange’s tech-heavy Star Market.

Other custom chip efforts by Chinese companies include Alibaba, whose T-Head unit has developed processors for running and training AI models, and ByteDance, which has pursued several custom AI chip projects to support its AI infrastructure. Baidu’s Kunlunxin chip arm, which grew out of the company’s decade-old internal AI hardware work, is targeting a $50 billion valuation for a Hong Kong listing, The Information reported.

The Information : When Andreessen Horowitz Has a Sticky Situation to Solve, It C

When Andreessen Horowitz Has a Sticky Situation to Solve, It Calls ‘The Wolf’
For some of Silicon Valley’s biggest names, Mark Dyne is the person they turn to when they need help dealing with thorny business and personal problems.

The Takeaway
  • Mark Dyne provides legal strategy, crisis management and complex deal negotiation for top tech investors and entrepreneurs.
  • His clients include Andreessen Horowitz and Silver Lake.
  • Dyne helped settle disputes with Solano County land owners who were in a legal battle over California Forever, a new city backed by the tech elite.

Last year, Alex Bouaziz, CEO of $17 billion payroll startup Deel, was several months into an escalating legal battle with an archrival, Rippling, when he texted one of his investors, Ben Horowitz, about an issue. He alerted Horowitz to a recent podcast in which Rippling CEO Parker Conrad had criticized Horowitz’s venture capital firm, Andreessen Horowitz, for standing by Deel amid a corporate spying scandal. (Rippling had sued Deel in March 2025, accusing it of paying a Rippling employee to snoop on Deel’s behalf.)

Bouaziz assured Horowitz someone was handling the matter. “Mark (the wolf) is on this,” he wrote in a text message that emerged last month as part of the litigation between Deel and Rippling.

Bouaziz was referring to Mark Dyne, a South Africa–born jack-of-all-trades—part investor, part deal negotiator, part legal strategist—who caters to influential tech executives and investors with thorny business and personal problems to solve. The nickname some of his clients use for him, The Wolf, speaks to Dyne’s skills in cleaning up an array of legal and personal problems on their behalf. According to two of Dyne’s clients, it’s also a joking reference to the tuxedoed character in “Pulp Fiction” played by Harvey Keitel, who is called in to get rid of a dead body and clean up a blood-splattered car.

Dyne’s work is far less grisly. He gets called in to organize surveillance of rivals for his clients during high-stakes litigation and to serve as an intermediary in sensitive personal conflicts, according to people familiar with his services, including past clients. He also helps distressed companies shut down and organize rescue financing, and he reincorporates businesses offshore to protect clients from litigation, those people said. Dyne typically gets paid a retainer, sometimes with extra compensation if he successfully negotiates a deal, according to court records and a person who has heard Dyne describe his arrangement with a client.

After Dyne began advising Deel, the company hired a law firm that he has had a long association with, Skadden, Arps, Slate, Meagher & Flom, which countersued Rippling for defamation. And according to a person familiar with the matter, Dyne advised Deel to hire crisis communicator Michael Sitrick, who has represented wrestling entrepreneur Vince McMahon, Kobe Bryant and Jeffrey Epstein’s lawyer. (Deel later withdrew its defamation suit and filed a separate countersuit against Rippling; the litigation between the two companies is ongoing.)

Over the years, Dyne has gotten close to Egon Durban, founder of private equity firm Silver Lake, who entrusted Dyne with acting as a litigation strategist and mediator for a housing analytics startup, HouseCanary, where Durban was an investor and board director, according to people familiar with the situation. In 2016, he advised Silver Lake during a crisis involving Nikesh Arora, now CEO of Palo Alto Networks, who was under fire at the time over potential conflicts based on his dual roles at Silver Lake and SoftBank.

Dyne has worked for Flannery Associates, a group backed by Marc Andreessen, Reid Hoffman and other tech billionaires, which is seeking to build a tech-friendly city called California Forever in rural Solano County, Calif. Recently, Dyne has been involved in hammering out agreements with property owners in the area to settle litigation between Flannery and the property owners and to buy their land.

In a recent settlement meeting with a lawyer who represented some of those property owners, Dyne introduced himself as having “a lot of rich friends,” according to contemporaneous notes taken by the lawyer. “I have worked on a lot of matters for a lot of different people, for a lot of different companies,” he said.

“I am the fixer,” Dyne added, according to the lawyer’s notes.

An attorney representing Dyne denied in a letter to The Information that his client had ever made such a statement. However, the About page for Morpheus Ventures, a Los Angeles–based investment firm Dyne founded, uses the term to describe the services it provides clients.

“We are…professional fixers for complex situations,” it reads.

“The way we see the world, gone are the days when startups simply have a clear and linear path from Series A to unicorn,” Morpheus’ website continues. “Nowadays, something or someone will try to block their path—a competitor, regulator, customer, rogue employee, or attorney general. That’s where we come in.”

When I called Dyne’s phone recently, the person who answered denied that he was Dyne, despite the fact that a subsequent WhatsApp message sent to the same phone number displayed Dyne’s photo. “You have an old number of his,” said the person who answered the call, speaking with a noticeable South African accent. He then promptly hung up. The next day, Dyne’s LinkedIn account was deactivated.

Dyne’s early career in business took him on a winding journey through entertainment, videogames and file sharing. Starting in 1992, he was a co-CEO of the Australian arm of Sega, the Japanese videogame publisher, where in 1997 he launched a $70 million theme park, Sega World, in Sydney. It quickly closed due to low turnout.

Around the same time, he co-founded Brilliant Digital Entertainment, which made interactive CD-ROM movies. People who interacted with Dyne around this time recalled him as an irrepressible dealmaker. In the late ’90s, Jordan Mechner, the creator of the Prince of Persia videogame series, had a lunch meeting with him, in which Dyne “spun off financing ideas like a Roman candle throwing off sparks.”

“My kind of guy,” Mechner wrote in a diary entry at the time, which he published on his personal website.

A big break for Dyne came after he founded a Los Angeles–based advisory firm, Europlay Capital Advisors, in 2001. He got close to a couple of European entrepreneurs, Janus Friis and Niklas Zennström, who were developing a file-sharing service, Kazaa, that jumped into the void left behind by the shutdown of Napster.

Dyne helped Friis and Zennström move Kazaa’s intellectual property into corporate entities in Vanuatu and the British Virgin Islands, which would make it harder for record companies, movie studios and others to pursue litigation against the company, according to court filings. Then Dyne allegedly helped them cripple its rival, StreamCast Networks, by cutting off access to Kazaa’s underlying peer-to-peer file-sharing technology, on which StreamCast had relied, according to a lawsuit Streamcast filed against Europlay, Zennström and Friis that a judge later dismissed.

Even before Dyne worked for Kazaa, he already was familiar with StreamCast: It, too, had been his client. The company had previously hired Europlay to help it raise cash, according to court filings.

In 2006, Dyne countersued StreamCast for allegedly breaching an earlier agreement that barred it from suing Europlay. StreamCast settled the case and soon filed for bankruptcy.

Despite the litigation, there seems to be no hard feelings over the episode on the part of StreamCast founder Steve Griffin. He said Dyne was at turns hard-nosed and generous (Dyne once loaned him $200 for a rental car).

“One moment he’s talking about his children, and the next moment he’s a corporate shark,” Griffin said in an interview. “He’s a chameleon.”

In 2003, Dyne helped Zennström and Friis launch a new company: Skype, which used Kazaa’s underlying software to power its voice-call technology. Dyne became a seed investor and Skype board member.

After Skype sold itself to eBay for $2.6 billion in 2005, its new parent company struggled to reap the benefits it had originally anticipated from the acquisition. Four years later, eBay began taking steps to spin Skype out as a separate company.

That created an opportunity for Zennström, Friis and Dyne to try to spin more gold out of Skype by buying it back. But eBay had announced a sale to a consortium of investors, including Silver Lake, Index Ventures and Andreessen Horowitz.

‘Mark (the wolf) is on this.’
Friis and Zennström had leverage, though. They still controlled an entity, Joltid, that owned the underlying peer-to-peer technology Skype relied on, which eBay had not acquired as part of the deal.

Just prior to eBay’s announcement of its plans to sell Skype, Dyne helped the Skype founders sue eBay for allegedly breaching copyright laws by using Joltid’s technology without a license, according to a person familiar with the matter. Then he helped the founders sue members of the would-be investing consortium, accusing them of stealing information in their bid for Skype.

These hardball tactics eventually brought the consortium to the negotiating table with Dyne, who hammered out an agreement with them to secure a 14% stake in Skype in exchange for Joltid’s intellectual property. (EBay and the Skype founders settled Joltid’s suit against the auction site as well.)

Andreessen Horowitz and Silver Lake’s Durban admired Dyne’s tenacity, according to people familiar with the situation. “He did such a good job on winning the battle for the Skype founders that ultimately, Silver Lake and Marc and Ben thought they should hire this guy,” said a person who has worked closely with Dyne.

Dyne rejoined the board of directors of Skype once it separated from eBay. Two years later, he helped engineer a second sale of Skype—this time, for $8.5 billion to Microsoft—Dyne said in his biography on Morpheus Ventures’ website.

Not all of Dyne’s work has involved corporate dealmaking.

In 2013, when Friis’ fiancée, Maria Louise Joensen (better known as Danish pop star Aura Dione), wanted to move to Los Angeles to break into the U.S. music scene, Dyne rented her a house, bought her a new car and paid her bills, according to court filings. When Joensen’s former manager claimed the rights to her music, Dyne hired Lance Etcheverry, a Skadden Arps attorney he worked with often, to sue the agent, the filings said.

Joensen prevailed. In a celebratory email later filed in court, Etcheverry credited the victory to a legal maneuver he and Dyne deployed to turn the judge against the former manager and his legal team. “The Strategy worked to perfection,” he wrote.

The celebration didn’t last. In 2015, Friis broke off his engagement with Joensen, alleging that she had cheated on him. Then Friis launched a series of lawsuits against her—including one that sought to get back an engagement ring and other gifts.

Separately, Dyne sued Joensen for allegedly breaking a handshake deal to repay more than $2 million of Skadden Arp’s legal fees. Joensen alleged in a court filing that Dyne had kicked her out of the Hollywood Hills home. Dyne in an affidavit said the lease on the home was about to end.

After Joensen’s lawyer offered half of her future royalties as a settlement, Dyne lost patience with what he viewed as an unsatisfactory proposal to resolve the conflict. “I am done wasting my time,” Dyne wrote to Joensen’s lawyer. “I don’t think that anyone can control Aura in her dealings. She does what she wants when she wants and thinks she’s a Diva. Perhaps only the courts can control her.”

A court ultimately ordered Joensen to pay Dyne’s firm $3 million.

Eventually, Dyne’s relationship with Friis frayed. In 2021, the two had a blow-up over an investment partnership they both directed that had stakes in Wire, a file-sharing startup, and Starship Technologies, a delivery robot startup.

Dyne sued Friis, alleging that he had a plan to reduce Dyne’s stake in the partnership. In his legal complaint, he called Friis “a serial litigator” who “believes he is always entitled to have what he wants.” In his own legal filing, Friis accused Dyne of extracting millions of dollars from his association with Friis while pretending to act in his best interests. Dyne eventually withdrew the case.

More recently, Dyne has been using his negotiating skills to help keep an effort to build a new, walkable city in Solano County on track. In 2023, The New York Times revealed that Flannery Associates had spent around $900 million to quietly snap up large plots of land in the area for the California Forever project, which is about an hour’s drive northeast of San Francisco.

Dyne and his firms, Europlay and Morpheus, have popped up in a variety of ways in the area. Dyne and another partner at Morpheus each donated $5,000 to the Solano County Deputy Sheriff’s Association Political Action Committee in May this year, according to public campaign filings.

In late 2023, a local resident, Aiden Mayhood, applied to join a community advisory board that California Forever had advertised to get feedback on the project. Mayhood, then 22 years old, soon got a call from a junior partner at Dyne’s Europlay Capital and Morpheus Ventures who wanted him to fly to Los Angeles to discuss his application, Mayhood wrote in a Facebook group dedicated to opposing California Forever last month. Mayhood declined the invitation and learned of Dyne’s connection to the firms after some online research.


In 2023, the tech-backed group behind California Forever, Flannery Associates, filed a lawsuit against a group of landowners in the area that sought $510 million in damages, accusing them of conspiring to inflate the value of their properties. A lawyer who represents some of those landowners said he recently got a call from an attorney for Skadden, the law firm for Flannery Associates, telling him to expect a call from a settlement negotiator: Dyne.

When they connected, the lawyer asked Dyne who he was. Dyne said he wasn’t himself a lawyer and that he was “just here to get the deal done,” according to the lawyer’s notes of the conversation.

Dyne told the lawyer he didn’t report to Jan Sramek, the public face of California Forever. He was coy about the identity of his ultimate employer.

“I talk to the money,” he told the lawyer.

FT : Saudi Arabia blocking payments to UAE accounts, businesses say Transfers fr

Saudi Arabia blocking payments to UAE accounts, businesses say
Transfers from kingdom to companies and individuals in Dubai repeatedly returned or delayed in a sign of economic tension

Saudi Arabia has been blocking or delaying financial transfers from banks in the kingdom to accounts based in the United Arab Emirates, raising concerns among businesses that friction between the Gulf powerhouses is spilling over into trade.

Multiple people told the FT that payments from Saudi entities to the UAE accounts of companies and individuals based in Dubai had since May either been returned or delayed by the kingdom’s banks, without any clear explanation.

A western executive with a Dubai-based healthcare firm said three payments from a Saudi client it has been trading with for several years had been blocked and returned by their Saudi banks since mid-May. The executive said the funds were “typically held for about a week with no query raised to sender or beneficiary, then simply returned”.

“They’ve even tried to make a transfer personally but that gets blocked too,” the executive said. “They spoke to the bank guys and got very sheepish answers. They are saying there’s a block from the [Saudi] central bank and they can’t really give more detail than that.”

Saudi Arabia’s central bank said the financial sector “operates within a robust regulatory framework” and there were “no direct restrictions on specific countries; banks apply risk-based measures consistently across all transactions to safeguard the integrity of the financial system”.

Another Dubai-based executive said his payments from a Saudi business now “take a long time”.

“We have a customer waiting for goods, but the payment is not happening,” that executive said. “A lot of suppliers [for Saudi Arabia] come from the UAE, by doing this they are shooting themselves in the foot.”

Saudi Arabia and the UAE boast the Arab world’s biggest economies and are major trading partners with annual bilateral trade of more than $20bn. Many companies use Dubai as a hub to serve the kingdom and flights to the kingdom are still packed, executives say. 

But tensions between the two have escalated since Riyadh accused Abu Dhabi of backing a secessionist Yemeni faction that launched an offensive against Saudi-aligned forces in Yemen in December. 

The kingdom accused the UAE — which had been its main partner in a Saudi-led coalition that intervened in Yemen’s civil war in 2015 to fight Iranian-backed Houthis — of threatening its national security interests by backing the offensive.

That crisis triggered the biggest rupture between the Gulf neighbours in decades and exacerbated long-simmering tensions over economic competition and differences over their approaches to conflicts in the region.

The US and Israel’s war against Iran initially appeared to ease the spat as Gulf states sought to put on a unified front after they became the main target of the Islamic republic’s retaliation.

But analysts say it merely papered over the cracks, and in late April, the UAE shocked its neighbours by announcing it was leaving Opec, the oil cartel of which Saudi Arabia is the de facto leader. 

The UAE — which was Opec’s third-largest producer — insisted the move was related to its “economic vision and evolving energy profile”. But Abu Dhabi had long been frustrated with production quotas imposed by Opec and the decision was viewed as dealing a blow to Saudi Arabia. 

Saudi officials have previously insisted the political spat with the UAE would not affect trade or economic relations.

An Emirati official said its economy ministry had not received any complaints or reports from private-sector companies regarding difficulties or unusual delays in completing bank transfers between the UAE and Saudi Arabia.

“The UAE and Saudi Arabia maintain deep and longstanding economic and commercial ties,” the official said. “We remain in regular engagement with the private sector and relevant stakeholders, and would review any specific concerns brought to our attention through the appropriate channels.”

But an Emirati businessman said after his company had faced unexplained blocks on bank transfers from Saudi Arabia to the UAE, he had to reroute the money via Bahrain.

Another Dubai-based services provider had a bank payment from a Saudi client cancelled last month, forcing them to use the more expensive option of PayPal.

“I got no info from the bank, nor did the Saudi guys — it just said ‘transaction failed’,” they said.

One multinational company also faced troubles transferring money from the kingdom to another firm in the UAE, according to a senior banking executive.

“You have some transfers that were stopped, it was very bizarre . . . some very small ones, we’re not talking about millions,” the executive said.

Another person familiar with the matter said people transferring money between their own accounts or to family and friends had also been running into issues in recent weeks.

FT : Apple interest thrusts China’s CXMT into memory chip spotlight Sharp turnar

Apple interest thrusts China’s CXMT into memory chip spotlight
Sharp turnaround for state-backed company central to Beijing’s AI supply chain efforts

On a sprawling campus in the eastern Chinese city of Hefei, construction workers are racing to complete a new production line for ChangXin Memory Technologies, a state-backed company that until recently was little known outside China’s semiconductor industry.

CXMT has been thrust into the global spotlight by the race for memory chips. Apple has begun testing the company’s DRam chips for devices sold in China, according to two people familiar with the matter, as the iPhone maker leads a lobbying effort among US tech companies to get the US government to allow broader use of the company’s products.

The interest in CXMT marks a sharp turnaround for a company that spent nearly a decade burning through billions of dollars but has now become central to Beijing’s efforts to build a domestic AI supply chain — and is poised to become one of the most profitable technology companies to be listed on China’s domestic stock market.

“China has very high expectations of CXMT,” said Emory Tsai-Yi Wang, a fellow at the Taipei-based Research Institute for Democracy, Society and Emerging Technology. “It emerged as the leader in the race to find a local memory chip champion and is critical to the nationwide project to build a self-sufficient AI supply chain.”

The memory shortage has transformed CXMT’s finances. Its net profit soared to Rmb33bn ($4.8bn) in the first quarter of this year, according to its IPO prospectus — a striking reversal from the Rmb37bn ($5.4bn) in losses it has accumulated over the past decade.


CXMT is now the world’s fourth-largest producer of DRam — the chips used in everything from smartphones to servers — behind SK Hynix, Samsung Electronics and Micron.

It accounted for about 11 per cent of global DRam wafer capacity last year, according to SemiAnalysis, a figure expected to rise to 15 per cent by 2028 as new production lines come online in Hefei, Shanghai and Beijing.

For US tech groups competing over a finite global supply of DRam wafers, the prospect of a fourth global supplier in China is appealing but politically sensitive.

Apple has previously faced public pushback from US policymakers when it last explored using Chinese memory suppliers, including then Republican senator Marco Rubio, who flagged security risks in 2022.

Apple declined to comment on whether it was testing CXMT’s chips for devices sold in China.

Despite CXMT’s rapid growth and plans to increase production, analysts say additional Chinese supply is unlikely to ease memory chip prices soon, as virtually all of its output is already committed and demand continues to grow.

“There’s a misconception that Chinese memory is dramatically cheaper and will flood the market,” said Ray Wang, memory analyst at SemiAnalysis. “Capacity is extremely constrained. Even as CXMT expands, it will remain supply constrained for at least the next two years.”

Over the longer term, however, competitors fear a repeat of the pattern seen in Chinese industries from solar panels to electric vehicles: years of state-backed investment followed by rapid capacity expansion and falling prices that squeeze foreign rivals.

CXMT declined to comment.

Although CXMT is celebrated as a national tech champion, its origins are rooted in foreign technology. In 2016, Zhu Yiming, a Silicon Valley returnee who had founded GigaDevice, a fabless flash memory maker, began a Hefei government-backed project to build domestic fabs and cut the country’s reliance on foreign imports.

In 2019, it acquired critical patents from a bankrupt German company, called Qimonda, which according to SemiAnalysis “became the basis” of CXMT’s DRam business. It also hired key personnel from the German company and recruited heavily from Korea and Taiwan. This year, Korean prosecutors sentenced a former Samsung researcher for leaking chip technology to CXMT.

CXMT also developed a close relationship with Dutch lithography group ASML, whose deep ultraviolet (DUV) machines remain essential to its manufacturing process despite Beijing’s push to localise semiconductor equipment. CXMT has not been added to the US Entity List so can still buy China-approved equipment from ASML.

Outside the company’s headquarters, a “friendship garden” founded in 2024 commemorates its partnership with ASML with a sign depicting the two companies shaking hands.

Much of CXMT’s rise reflects its close relationship with the local government, which has spent years cultivating the company as the centrepiece of a local semiconductor ecosystem.

Analysts estimate that the government supplied cheap land for CXMT’s vast campus, provided financing and subsidies, and helped attract suppliers and customers to the city, creating an industrial cluster around CXMT. Between 2023 and 2025, CXMT received at least Rmb6bn ($880mn) in government subsidies, according to its financial disclosure.

CXMT has at least 15 state-owned shareholders, which collectively own 36 per cent of the company. Many of its privately owned funds also have backing from state-owned limited partners.

“Much of the support is intangible,” DSET’s Wang said. “The Hefei government likely helped connect CXMT with suppliers and customers, including local automakers. Those relationships were critical.”

The IPO will provide a boon for Hefei’s investors. Analysts expect that CXMT’s valuation could soar to as much as Rmb3tn after listing. If achieved, this would imply gains of more than Rmb1tn for Hefei — roughly equivalent to the city’s annual GDP.

CXMT’s standing among Hefei’s other tech groups, from AI group iFlytek to carmaker Nio, has risen sharply in the past year. “Everyone wants to join CXMT now. In the past, it wasn’t considered as prestigious as iFlytek,” said one Hefei-based engineer. Last year, it added about 6,000 people, taking its headcount to 20,000 employees.

CXMT disclosed in its IPO prospectus that it plans to raise Rmb29.5bn for upgrading its manufacturing processes, R&D for “next generation” DRam technology and hiring top engineering talent, including from overseas.

The company’s campus reflects both its strategic importance and its secrecy. Many employees live in apartments on or beside the site. Personal mobile phones are banned inside offices, a common practice in tech companies with sensitive technology. During a visit by the FT, guards stopped reporters photographing the heavily fenced complex, describing it as a “confidential compound”.


While its DRam business has made CXMT profitable, a bigger test will be whether it can crack the high-bandwidth memory (HBM) market.

CXMT did not mention its nascent HBM business in its IPO prospectus, which it recently entered and remains a tightly guarded secret. HBM is the most valuable form of memory used in AI accelerators, an area in which it remains far behind the three market leaders.

The Chinese group has allocated some of its new production lines, including in Shanghai, to HBM, but usage in China remains limited, analysts say. Chinese AI groups stockpiled large amounts of Korean HBM ahead of a ban in 2024, while supplies have continued to enter the country through smuggling routes, according to multiple people familiar with the matter.

CXMT also faces a higher cost of production for HBM than its foreign rivals because it has been barred from buying ASML’s cutting-edge EUV machines, which simplify the patterning of extremely small features on the silicon wafer.

“The restrictions on CXMT buying EUV have impacted its HBM business. Yield rates remain low, and it creates barriers to Chinese companies adopting the technology when CXMT is running trial-and-error runs to improve this,” said DSET’s Wang. 

Yet few industry experts doubt Beijing’s commitment to developing the technology, despite the higher costs. Profits from conventional DRam, together with proceeds from the forthcoming IPO, are expected to fund a multiyear push into HBM.

“Chinese AI companies will adopt CXMT’s memory because they share the same strategic objective,” added DSET’s Wang. “China has concluded that self-sufficiency is ultimately more important than relying on foreign suppliers. CXMT has become central to that ambition.”

FT : US dealmaking boom fuels payday for merger arbitrage Uncertainty around meg

US dealmaking boom fuels payday for merger arbitrage
Uncertainty around megadeals has created opportunity for hedge funds that wager correctly

The Trump administration’s permissive stance on dealmaking has created a “complexity premium” for merger arbitrage investors betting they can navigate clashing US federal and state rules as well as overseas regulation.

Deals facing antitrust, national security or political scrutiny have offered investors a premium of about 60 per cent compared to simpler transactions over the past five years, up from 25 per cent in the previous five-year period, according to new research by hedge fund Davidson Kempner.

The widening spread has pushed potential returns 2.5 percentage points higher since 2021 — which can be worth hundreds of millions of dollars in extra profits for savvy investors who bet correctly on blockbuster deals.

“You have this interesting paradox where the US regulatory environment from an antitrust standpoint has become more favourable, and that’s driving this boom in megadeals, at the same time global regulatory complexity has increased,” said Suzanne Gibbons, head of research at Davidson Kempner, which has more than $38bn of assets under management.


As transactions valued at more than $10bn comprise a greater share of deals than they have historically, a “complexity premium” had emerged for traders who can predict whether and when controversial deals will close, Gibbons said.

Merger arbitrage funds typically purchase shares of a company after a deal is struck when the target’s stock is trading below the offer price to account for the risk that a deal is blocked or delayed.

For example, hedge fund Pentwater Capital earned a large windfall betting that Nippon Steel’s $15bn acquisition of US Steel would be completed.

The fund, which held nearly a tenth of US Steel, according to security filings, benefited from a volatile share price that at one point fell nearly 40 per cent below the deal price during a lengthy review by both the Biden and Trump administrations.

High-profile deals that currently await approval include the $55bn leveraged buyout of video game maker Electronic Arts and Paramount’s $111bn takeover of Warner Bros Discovery. Whether they are completed depends on decisions from authorities including the European Commission and US state attorneys-general.

“If EA closes and Warner Brothers closes, yes, it will be a fantastic year” for the strategy, said a portfolio manager at an arbitrage-focused hedge fund.

While deal closure rates have been relatively high in the past 18 months, according to Felix Lo, a portfolio manager at Trium Capital, deals face more approvals than ever.

“It used to be mostly antitrust and maybe some national interest, but now there’s politics, now there’s consideration of industrial policy,” Lo said. “There’s definitely a lot more complexity in the world, and that’s great for the strategy.”

And the US president’s outsized interest in certain deals introduces heightened uncertainty.

US President Donald Trump suggested in December he would be personally involved in reviewing the takeover of Warner Bros before saying the process would be left to the justice department. Billionaire Oracle founder Larry Ellison, an ally of the president, is the father of Paramount chief executive David Ellison and helped finance the takeover.

“A lot of people in our space rely heavily on lawyers and contacts, whether in Brussels or DC,” said Matthew Osowiecki, co-chief investment officer at Water Island Capital. “We felt like with the change of administrations that wasn’t working anymore . . . Trump would swoop in at the last minute and approve the deal.”

As the Trump administration relaxes antitrust enforcement some Democratic state attorneys-general have attempted to block deals, creating more uncertainty around the approval process. A group of states led by California is reviewing the Paramount-Warner Bros deal and could still sue to attempt to stop it.

But much of the new regulatory scrutiny has come from non-US regulators, particularly with deals involving foreign investment in companies that have connections to national security or critical industries.

“You have to have a view on the process in Brazil, in the European Union, in different countries in Europe, in China . . . Twenty years ago it would’ve been the US and Europe,” said Michael Chiaramonte, partner at Davidson Kempner.

The proportion of megadeals expecting to face foreign investment review increased from 11 per cent between 2005 and 2020 to 40 per cent over the past five years, according to Davidson Kempner’s research.

Osowiecki said that rather than relying so heavily on navigating the global web of regulators that must approve deals, the primary factor his fund looks at was “buyer commitment”, or whether acquirers are willing to make remedies and wait out long approval processes.

Getting past multiple regulators in different jurisdictions can require politicking, especially as jurisdictions such as the EU are concerned about how mergers might affect workers and consumers. 

Microsoft closed its $75bn acquisition of Activision Blizzard after 21 months — and only after revising the deal to win the approval of the UK competition regulator. Many investors and analysts had believed the deal would not be completed.

But Osowiecki said Activision was “hitting the cover off the ball” throughout the process, and Microsoft knew it was getting a “steal” on the deal.

“We knew their buyer commitment was there,” he said. “We knew they had strings to pull in regards to offering up remedies and they could get the deal over the line.”

FT : Meta tests ‘super sensing’ AI glasses that can record every moment Mark Zuc

Meta tests ‘super sensing’ AI glasses that can record every moment
Mark Zuckerberg’s hardware ambitions are edging into a new privacy fight over who gets recorded

Meta is testing a prototype of “super sensing” AI glasses that would use cameras and audio recordings to capture a wearer’s every moment, as it pushes into the contentious market for all-seeing, all-hearing devices.

The $1.5tn social media platform has been advancing a new hardware line of smart glasses that would continuously record audio while taking photos every few seconds, according to multiple people familiar with the matter. A user could then use AI to help query what they saw or heard, or recall their day.

The glasses have prompted internal debates over how to handle novel privacy challenges, including non-wearers finding the technology invasive.

With Meta’s current AI smart glasses, an LED in the corner of the frame lights up to signal to others when a wearer is taking photographs or filming.

However, executives are currently planning not to activate the LED when the super-sensing features are being used, according to multiple people familiar with the matter. That would make it harder for bystanders to know when they were being recorded, potentially intensifying the privacy concerns surrounding the technology. Those plans could still change, however, several people said.

The super-sensing features could also be activated on Meta’s existing glasses via a software update, the people said.

The move comes as chief executive Mark Zuckerberg has argued that AI glasses could one day replace the smartphone as the main device people use to access AI tools including translation or chatbots.

In one proposed system, raw footage and audio would not be stored by Meta or made available to the user, several people said. Instead, the metadata from that audio and images would be extracted and uploaded to the server for Meta’s AI to query, which proponents argue would have fewer privacy implications.

The company is also discussing whether data collected through the glasses and their features could be used to train its own AI models, as it pours billions into rivalling competition such as OpenAI, Google and Anthropic in the AI race.

Meta declined to comment on “internal prototypes” but said its approach focused on “privacy built in from the ground up”. The Information has previously reported some details about the super-sensing glasses project.

Zuckerberg hinted at the technology during Meta’s first-quarter earnings, saying he wanted glasses to evolve from “being able to answer questions to being able to be a personal agent that’s with you all day long, helping you remember things and achieve your goals”.

Meta is also exploring other kinds of always-on devices beyond glasses. In December, it bought Limitless, a maker of AI-powered pendants, which can record and transcribe conversations in real time and allow users to search that information via the company’s app.

Meta has curbed its costly push to build an avatar-filled “metaverse” that needs headsets to access, after the concept failed to gain consumer traction.

Instead, it has shifted its virtual and augmented reality strategy towards AI glasses following the success of its Meta Ray-Bans, smart glasses with inbuilt speakers and cameras, sold in partnership with eyewear group EssilorLuxottica. Last month, Meta launched an even cheaper version of those glasses, including one product with frames designed by influencer Kylie Jenner.

A more advanced version of the Ray-Ban glasses, released last year, includes a display that can overlay text messages or video calls on to one of the lenses.

With the super-sensing glasses, Meta is likely to face novel legal and regulatory challenges as well as fierce resistance from privacy-conscious consumers.

Privacy experts argue that always-on devices could violate data privacy or biometric data laws, for example. It is also unclear whether the company or the wearer will be responsible for any breach of wiretapping laws, as it remains illegal in multiple US states to record audio of a third party without their consent.

Some are calling for fresh regulation to tackle the nascent space. “There’s no one law that addresses all the different dangerous ways these tools have been designed and built,” said Woodrow Hartzog, professor of law at the Boston University School of Law. “Lawmakers have to take this seriously and update to accommodate this reality of always-on, always-seeing devices.”

Meta’s current range of Ray-Ban smart glasses already do not light up when a wearer is using AI to ask questions about the world around them. In a 2025 policy paper, Meta said: “If the LED blinked for extended periods at a time (like whenever there was an AI interaction), people could stop noticing it — reducing the awareness of when photos or videos are being captured by users to have for later.”

However it added that when glasses were used for AI features, it would “take steps to protect people’s privacy, like removing key identifiable information”.

Meta has faced privacy snafus related to the existing Ray-Ban glasses. In February, contractors in Kenya told the press that they had to view graphic material captured by the glasses, such as users having sex or going to the toilet. Last month, Wired magazine also uncovered code for a facial recognition system that was embedded in Meta’s smart glasses platform but unreleased. The company later removed the system.