News & Updates
The latest news and updates from companies in the WLTH portfolio.
Japanet Expands Its VC Fund After Bets on Anthropic, xAI Pay Off
Pegasus manages startup investments on behalf of companies including Toyota Motor Corp.-affiliate Aisin Corp. and Taiwan's Asustek Computer Inc. Japanese home shopping company Japanet Holdings Co. is expanding its venture capital fund with San Jose-based Pegasus Tech Ventures Inc., following the success of early bets in SpaceX, OpenAI, Anthropic PBC and xAI. The Nagasaki-based retailer known for infomercials targeting seniors in aging Japan will allocate $200 million to the fund, up from an initial $50 million in 2021, following "significant growth" in investments so far, the companies said in a statement. The fund, of which Pegasus is general partner, will focus on areas such as generative AI, robotics and space technology. Its Japan portfolio includes startup Aillis Inc., which seeks to use artificial intelligence to analyze medical scans. Asian companies have struggled to win stakes in promising startups in Silicon Valley, hampered by a lack of personal connections and reputation for slow decision-making. Pegasus also manages startup investments on behalf of Toyota Motor Corp.-affiliate Aisin Corp., Japanese chemical maker Denka Co., Taiwan's Asustek Computer Inc. and Acer Inc. and Indonesia's pharma company Kalbe Farma Tbk. "Everybody wants a piece of the Silicon Valley AI action," Pegasus Chief Executive Officer Anis Uzzaman said on a video call. The venture capital firm manages some 40 funds with about $2 billion in total assets and investments in 290 startups including Airbnb Inc., DoorDash Inc. and Coinbase Inc., he said. The investments also often yield commercial relationships. "For the Asian corporations, it's an opportunity to invest as well as find new technologies and innovations in aligned fields," he said. "For the American AI companies, it's access to new global markets." A simmering feud with mega-influencer Alix Earle isn't the only source of drama surrounding Alex Cooper's growing podcast empire. Lately, Cooper -- the host of Call Her Daddy, the fourth-most popular podcast in the US -- has been grappling with multiple challenges inside her business, including employee turnover, a sputtering slate of shows and discord between staff members and her husband and co-chief executive officer Matt Kaplan, according to people familiar with the company's operations. Such issues are becoming more common across the creator economy as myriad hit performers from Jimmy Donaldson, aka MrBeast, to Ashley Flowers of Crime Junkie to the trick-shot artists at Dude Perfect, have begun to build diversified companies capitalizing on their internet fame, while also running demanding podcasts or YouTube channels. In 2023, Cooper and Kaplan -- a producer of over 50 films and TV shows, including Netflix Inc.'s To All the Boys franchise -- co-founded Trending, an umbrella company housing the power couple's growing business ventures. Initially, the portfolio included Unwell Network, Cooper's burgeoning collection of lifestyle podcasts, and ACE Entertainment, Kaplan's production company. The idea was to take Cooper's audience of Gen Z women and speak directly to them, building content around their interests. Cooper and Kaplan began hiring dozens of employees to staff their business as they expanded into live events, TV series, a creative agency and physical products. Many of the new recruits were young adults who knew Cooper's show and were eager to be part of her company's festive aura while working alongside a woman they admired. In interviews with news outlets and on podcasts, Cooper described herself and Kaplan as driven entrepreneurs who expected employees to "work hard, play hard." At one point, she told Marie Claire that as CEO, her door is "always open" to employees and that she and Kaplan foster a "very positive and safe" environment for their young staff. Cooper is often extremely busy keeping her show and other creative projects running, said the people who asked not to be identified discussing non-public information, leaving Kaplan to run most of the network's day-to-day operations. Along the way, he has earned a reputation for frequently yelling at staff members, the people said. The issues have grown so acute that crew members have threatened to walk off the job at film sets and live tours if he doesn't keep his distance. The company declined to comment for this story. Last month, Trending launched a program on YouTube called Unwell Winter Games in which influencers squared off against each other in events, ranging from tubing down a mountain to competing in the drinking game flip cup. According to TMZ, there were some logistical problems during the run-up to the series, resulting in an influencer-contestant dropping out unexpectedly and being replaced at the last minute by a surprised member of the show's production team. At another point during the filming, according to the people familiar with the events, Kaplan berated the staff on set and threatened to prevent them from ever working in Hollywood again if they messed up, resulting in a prominent, long-experienced crew member breaking down in tears. Crew members filed complaints through the formal, on-set process and eventually told leadership they would quit if Kaplan didn't behave better, said the people. Some of the internal tension has spilled out publicly. In Call Her Alex, a 2025 Hulu documentary about Cooper's first live tour, the production manager threatens to quit at one point. In a subsequent scene, Kaplan chalks up the situation to the demanding work schedule surrounding the tour. But according to a person close to the conflict, the manager's concerns had much more to do with Kaplan disrespecting staff by yelling and swearing at them. The documentary also captured a frenzied environment as the company rushed to put together a live show. In the past year, several high-level executives have left Unwell, including the company's head of brand marketing, head of the network -- its second in three years -- and chief growth officer. Cooper has said Unwell employs around 100 people. On LinkedIn, at least 20 have marked themselves as leaving in 2025 and 2026. Meanwhile, Unwell has struggled to increase hit programming beyond Call Her Daddy. In 2025, as part of a broader deal with SiriusXM Holdings Inc., the company launched three original programs designed for an Unwell-branded satellite radio station. Less than a year later, all three were canceled. In 2025, Cooper's company also parted ways with Earle, the host of Hot Mess, the network's buzziest creative expansion. At the time, Unwell offered little explanation for Earle's departure. But recently, Cooper drew attention to the rift publicly. On TikTok, she said she was "tired of waking up" and seeing Earle spreading "fake drama" -- referring, it seemed, to Earle's recent repost of a video criticizing Cooper's podcast interviews. Afterward, Earle responded, "Okay on it!!" The surprise salvo rattled the internet, drawing intense scrutiny by everyone from amateur online sleuths to the New York Times. Earle, through her publicity team, declined to comment. These days, rather than building new programs from scratch, Unwell is focusing on acquiring existing shows, Cooper said in a recent interview with a podcast on her network. The shift means Cooper's team has entered a highly competitive market against well-established players who have larger budgets and fewer ties to any one individual star. For now, Call Her Daddy continues to be Unwell's biggest program. Podscribe, a podcast data company, estimates most shows on the network receive under 100,000 downloads a month. Call Her Daddy receives an estimated 13.3 million. Trending continues to release projects at a rapid clip. In the past month, it has launched several films and TV shows, including Love Overboard for Hulu, a dating competition series, the Hannah Montana 20th Anniversary Special for Walt Disney Co. and a new season of XO, Kitty, a spinoff of Kaplan's original Netflix success. The series was No. 1 on the streamer when it premiered in April. Cooper's live tours have also been gaining traction with fans. Last year, she held a two-day festival in Las Vegas, complete with Chippendales dancers and a pool party with Paris Hilton spinning tunes as DJ. Three of her West Coast shows in 2024 grossed a total of $540,000 with 9,549 tickets sold, according to Pollstar. The company's ad agency has also picked up prominent clients, including Google, and Cooper has said sales in the department are now nearly on par with revenue from podcasting. Get the Tech Newsletter bundle. Get the Tech Newsletter bundle. Get the Tech Newsletter bundle. Bloomberg's subscriber-only tech newsletters, and full access to all the articles they feature. Bloomberg's subscriber-only tech newsletters, and full access to all the articles they feature. Bloomberg's subscriber-only tech newsletters, and full access to all the articles they feature. Plus Signed UpPlus Sign UpPlus Sign Up By continuing, I agree to the Privacy Policy and Terms of Service. Last week, Unwell announced it had hired new executive leadership, including Joanne Bradford, previously of Honey, an online shopping company, who will serve as president. Ellen Rocamora, who previously worked for more than a decade at The Ellen DeGeneres Show, has joined as an executive producer. The new team members will be tasked, in part, with helping Cooper as she continues to navigate new lines of business. It's a tricky career trajectory, but one that can make sense for financial reasons in the long term. "Creator work is not necessarily stable," said Jess Rauchberg, an assistant professor of communication technologies at Seton Hall University. "They are leveraging their visibility into something that can be consumed off the podcast." Whether an influencer can branch out successfully, she added, often depends on the level of the team around them. "It's partially skill, but also privilege," she said. Cooper, for her part, shows no sign of retreat, despite the difficulties. "Unwell is a really, really fabulous and incredible and, I think, inspiring place to work," Cooper said in a recent interview on The Burnouts with Phoebe & Sophia podcast. "But it is f------ hard."
Elon Musk's Strategic Investment: A Boost for SpaceX | Headlines
Elon Musk significantly raised his investment in SpaceX by acquiring $1.4 billion worth of shares from both former and current employees last year, according to a report by The Information. However, this report has not been independently verified by Reuters. Elon Musk has increased his ownership in SpaceX with a substantial purchase of $1.4 billion in stock from existing and former employees, a move that underscores his long-standing commitment to the aerospace company. The report, sourced from The Information, highlights Musk's strategic decision to strengthen his influence and control within the innovative company, renowned for its ambitious projects and technological breakthroughs. While Reuters was unable to immediately confirm the details presented in The Information's report, the increase in Musk's stake signifies potential positive impacts on SpaceX's future endeavors and market positioning.
Amazon To Invest Up To $25 Bn More In Anthropic - BW Businessworld
The commitment adds to the USD 8 billion the Seattle-based company has already invested in Anthropic Amazon said it will invest up to USD 25 billion in artificial intelligence startup Anthropic. The investment includes an immediate USD 5 billion infusion, with a further USD 20 billion to be deployed over time subject to commercial milestones, the company said. The commitment adds to the USD 8 billion the Seattle-based company has already invested in Anthropic. Under the expanded deal, Anthropic has committed to spending more than USD 100 billion over the next decade on Amazon's cloud services, the companies said, highlighting the scale of infrastructure required to train and run large AI models. Amazon has emerged as a key beneficiary of the AI boom by supplying cloud computing and custom chips, even as it has struggled to generate comparable excitement around its own AI models. The company said it expects capital expenditures of about USD 200 billion this year, largely tied to artificial intelligence. Anthropic said it plans to bring roughly 1 gigawatt of AI computing capacity online by the end of this year using Amazon's Trainium2 and Trainium3 chips, with a longer-term goal of securing up to 5 gigawatts of capacity. "Anthropic's use of Trainium chips reflects the progress we've made together on custom silicon," Amazon Chief Executive Andy Jassy said in a statement. Anthropic, which develops large language models (LLMs) with a focus on safety, has been racing to lock in computing resources as competition intensifies in the AI sector. The startup has been releasing models aimed at coding and design tasks, areas seen as critical battlegrounds for enterprise adoption. The latest deal also points to Amazon's strategy of backing multiple major AI players. Earlier this year, the company said it would invest up to USD 50 billion in OpenAI, the maker of ChatGPT.
Europe warned over security risks of Anthropic's Mythos AI | News.az
European financial regulators must act to prevent the misuse of advanced artificial intelligence, according to the head of Germany's central bank. Deutsche Bundesbank President Joachim Nagel warned on April 21 that the latest AI systems could introduce serious cybersecurity risks to the financial sector, calling for stronger oversight across Europe, News.Az reports, citing Reuters. His comments focused on Anthropic's newest model, Mythos, which he described as highly capable of identifying and exploiting vulnerabilities in financial institutions' software. Nagel said the model represents a "double-edged sword," as it could strengthen cybersecurity defenses but also be used for malicious purposes, including targeting weak points in critical banking systems. According to the Bundesbank chief, banking authorities must ensure that such powerful AI tools are not misused, while also making sure they remain accessible to all relevant institutions to avoid market imbalances and competitive distortions. The warning reflects growing concern among European policymakers about the rapid evolution of AI systems and their potential impact on financial stability and cybersecurity infrastructure.
Musk bought $1.4 billion SpaceX shares last year, The Information reports
April 21 (Reuters) - Elon Musk increased his stake in SpaceX last year by purchasing $1.4 billion worth of stock from current and former employees, The Information reported on Tuesday. The secondary stock purchase, made through Musk's trust, was disclosed in a draft of SpaceX's confidential IPO prospectus, the report said. SpaceX also approved a plan last month that would award the billionaire CEO 60 million additional shares if the company's market capitalization climbs from $1.1 trillion to as high as $6.6 trillion and completes an ambitious plan of building data centers in space to supply compute for AI developers, The Information said. Reuters could not immediately verify the report. SpaceX did not immediately respond to a request for comment. (Reporting by Chandni Shah in Bengaluru; Editing by Janane Venkatraman and Mrigank Dhaniwala)
Amazon To Invest Additional $5 Bn In Anthropic - BW CFO World
Deal includes long-term cloud commitments and access to AI chips to scale advanced models Amazon has agreed to invest an additional USD 5 billion in artificial intelligence startup Anthropic, with scope for up to USD 20 billion more over time, as part of a broader USD 100 billion-plus cloud computing arrangement over the next decade. The acquisition further integrates the financial and operational connection between the two firms, as capital intensity in the AI industry continues to increase and competition over access to compute resources expands. Anthropic, which develops the Claude chatbot and associated coding tools, has committed to spending more than USD 100 billion on Amazon's cloud technologies and AI-specific chips over the coming 10-year period. The deal encompasses a combination of general computing capacity and the home-built Trainium AI accelerators at Amazon, intended to scale up future versions of Claude. Amazon was already a major investor in Anthropic, having invested USD 8 billion in the startup. From a corporate finance perspective, the deal incorporates Anthropic's expensive model development and infrastructure investments into Amazon's cloud revenue stream, and bases further equity investments by Amazon on commercial milestones instead of open-ended commitments. Amazon will continue to be a minority shareholder and will not have a seat on the board of Anthropic, maintaining the governance structure of the startup despite increasing capital reliance on the Amazon cloud platform. For capital markets participants, the structure emphasises how large cloud providers are effectively underwriting the balance sheet stress of major AI companies through long-term consumption-based agreements and equity structures designed to de-risk incremental funding. The fact that Anthropic has to buy tens of billions of dollars of compute capacity, in turn, indicates that scaling advanced AI models is now a multi-decade, multi-hundred-billion-dollar capital investment for both vendors and providers.
FBI Warns Anthropic's 'Mythos' AI Can Hack Every Major Operating System in the World
The FBI is warning that Anthropic's 'Mythos' AI can hack every major operating system in the world, prompting federal officials to scramble amid urgent alerts to top government and financial leaders. The powerful new artificial intelligence model demonstrated the ability to identify and exploit vulnerabilities in virtually every major operating system and web browser. AI giant Anthropic's new system, known as "Mythos," is being kept under tight restrictions. However, insiders say the threat is already serious enough that the U.S. government is racing to understand it before it's too late. Zerohedge.com reports: Treasury Chief Information Officer Sam Corcos briefed the department's cybersecurity team on the technology last week and has directed efforts to gain access to the model "as soon as this week." The request comes days after Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell summoned top Wall Street CEOs to an urgent meeting at Treasury headquarters. Executives were warned that Mythos and similar frontier AI models could usher in a new era of heightened cyber risk. Anthropic itself has cautioned that the model may be capable of powering sophisticated cyberattacks unless companies proactively test it against their own systems and build defenses ahead of any wider release. At the meeting, bank leaders were strongly urged to take the model seriously and use it internally to detect vulnerabilities. What Is Mythos and Why the Restrictions? Anthropic introduced Mythos (also referred to as Claude Mythos Preview) as part of its new Project Glasswing initiative. In internal testing, the model demonstrated extraordinary offensive cybersecurity capabilities: it was able to identify and exploit vulnerabilities "in every major operating system and every major web browser when directed by a user to do so." In one documented case, it wrote a web browser exploit that successfully chained together four separate vulnerabilities. Project Glasswing brings together Amazon Web Services (AWS), Apple, Broadcom, Cisco, CrowdStrike, Google, JPMorganChase, the Linux Foundation, Microsoft, NVIDIA, and Palo Alto Networks to address growing concerns within the cybersecurity community that AI models are now capable of discovering and exploiting vulnerabilities at a faster pace than humans can keep up with. ... According to the post on Anthropic's website, the model's strong agentic coding and reasoning skills enable it to uncover and exploit security flaws when directed by the user that have existed for years, even decades without detection. Benchmarking results cited by the company suggest a notable performance gap between Mythos Preview and its previous models in cybersecurity-related tasks. -cxtoday.com What Mythos Has Discovered: Key Findings from Red Team Testing In controlled testing against real codebases in isolated containers, the model autonomously identified thousands of zero-day vulnerabilities across every major operating system and every major web browser. The testing used an agentic workflow: file prioritization based on a 5-tier vulnerability likelihood ranking, parallel Claude Code invocations, and secondary validation for severity and exploitability. Standout Zero-Day Discoveries Include: * 27-year-old remote crash vulnerability in OpenBSD (TCP SACK processing): An integer overflow in signed TCP sequence number comparison that enables a null-pointer dereference and remote denial-of-service against any responding host. The bug had survived decades of manual code review and extensive fuzzing campaigns. * 16-year-old bug in FFmpeg (H.264 parser): A slice number collision that triggers an out-of-bounds heap write when processing crafted frames with 65,536+ slices. The vulnerability originated in 2003, became exploitable after a 2010 refactor, and had evaded detection despite automated testing tools hitting the vulnerable path five million times. * 17-year-old FreeBSD NFS Remote Code Execution (CVE-2026-4747): A stack buffer overflow in RPCSEC_GSS authentication (96-byte buffer for 304-byte input) combined with NFSv4 information disclosure. Mythos autonomously constructed a 20-gadget ROP chain split across six sequential RPC requests -- a feat the prior model (Claude Opus 4.6) could achieve only with significant human guidance. Firefox JavaScript Engine Testing Results were especially dramatic: * Claude Opus 4.6: Developed only 2 working exploits out of several hundred attempts. * Mythos Preview: Developed 181 working exploits and achieved register control in 29 additional cases. OSS-Fuzz Results showed a similar leap: * Mythos generated 595 tier-1/2 crashes (plus several tier-3-5), including multiple tier-5 control-flow hijacks (full arbitrary code execution) on fully patched targets. These discoveries were achieved at remarkably low cost - many individual zero-day runs cost under $50, with full OpenBSD testing campaigns under $20,000 and Linux kernel N-day exploits under $2,000 each. Because of the dual-use risks, Anthropic has not released Mythos to the public. Instead, it is being provided on a tightly limited basis through Project Glasswing to a select group of vetted organizations - including major tech companies, cybersecurity firms, JPMorgan Chase, and the Linux Foundation - for defensive purposes only (scanning their own systems to find and patch flaws before attackers can exploit them). Anthropic has committed up to $100 million in usage credits to support these efforts. Several major financial institutions have already begun internal testing: * JPMorgan Chase was publicly named as part of Project Glasswing. * Goldman Sachs, Citigroup, Bank of America, and Morgan Stanley have also gained access or are in the process, according to people familiar with the matter. The company stated in its Project Glasswing announcement that it has been in "ongoing discussions" with government officials about the model and is "ready to work with local, state, and federal representatives." Pentagon Supply-Chain Risk Designation The Treasury's push for access is notable because the Pentagon formally designated Anthropic a US supply-chain risk earlier this year following a dispute over how the company's AI technology could be used by the military. The Defense Department gave Anthropic a six-month window to transition its services to another provider. Anthropic is actively fighting the designation in federal court. Despite this, Corcos - who previously encouraged the use of Anthropic's Claude AI tools inside Treasury before the Pentagon label - is now driving the department's effort to investigate Mythos.
Bitget launches IPO Prime with first offering preSPAX linked to SpaceX - Cryptopolitan
IPO Prime integrates early access with instant liquidity, enabling continuous price discovery and real-time position adjustments. Bitget, the world's largest Universal Exchange (UEX), opened subscription for IPO Prime, marking the next phase of its pre-IPO access platform as users begin participating in its first offering, preSPAX. The subscription window runs from April 18, 2026, 06:00 to April 21, 2026, 06:00 (UTC), followed by token distribution between April 21, 2026, 06:00 and 10:00 (UTC). Spot trading for preSPAX will commence at 12:00 (UTC) on April 21, enabling continuous market participation immediately after allocation. What makes IPO Prime different? IPO Prime is designed to provide structured access to pre-IPO opportunities through tokenized instruments that reflect the economic performance of underlying companies after listing. The first asset, preSPAX, is linked to SpaceX, offering users exposure to one of the most closely followed private companies globally. Participation is conducted through a subscription-based model, where users commit capital in stablecoins and receive proportional allocations based on total demand. This replaces fixed allocation models with a transparent mechanism that adjusts dynamically to participation levels. What distinguishes IPO Prime is the integration of early access with immediate liquidity. Once distribution is completed, preSPAX becomes tradable in a spot market, allowing users to enter, exit, or adjust positions ahead of a public listing event. This structure introduces continuous price discovery at a stage where traditional pre-IPO investments are typically locked. Bitget provides broader access "Early access has always been limited, but it's not just about getting in early," said Gracy Chen, CEO of Bitget. "It's also about what you can do after that. Giving users the ability to act and adjust in real time changes how these opportunities are approached." Assets listed on IPO Prime are issued through our compliant partner, Republic, and are designed to track economic outcomes rather than represent direct equity ownership. This framework enables broader participation while maintaining a structured approach to exposure. Within Bitget's Universal Exchange model, IPO Prime reflects a broader shift in how trading environments are being designed. As users increasingly look to manage capital across multiple asset classes and stages of the market cycle, the ability to access early-stage opportunities alongside liquid markets within a single system is becoming more relevant. By extending its infrastructure to support both participation and trading in one place, Bitget continues to evolve UEX as a more complete environment for navigating global markets.
SpaceX IPO filing shows Musk will keep voting control | News.az
SpaceX is planning to cement founder Elon Musk's control following its initial public offering (IPO), granting him and a small group of insiders super-voting shares that would outweigh other investors, according to excerpts of the company's IPO filing reviewed by Reuters, News.Az reports. The prospectus, which was filed confidentially this month, provides new details about the company's financial performance and corporate governance structure. According to the filing excerpts, Musk will continue serving as chief executive officer, chief technical officer, and chairman of SpaceX's nine-member board of directors after the offering is completed. Although Musk was paid $54,080 last year, the excerpts indicate he stands to gain billions in equity following the company's stock market debut. SpaceX is targeting a listing valuation of approximately $1.75 trillion and aims to raise $75 billion, which would make it the largest initial public offering in history. President and Chief Operating Officer Gwynne Shotwell received $85.8 million in total compensation last year, according to previous Reuters reporting, while Chief Financial Officer Bret Johnsen was paid $9.8 million. The IPO preparations are also involving senior executives, with Musk's team holding three days of meetings this week with Wall Street analysts. The schedule begins with a tour and briefings at SpaceX's Starbase launch facility in Boca Chica, Texas. The filing excerpts further show that SpaceX will adopt a dual-class equity structure. Under this system, Class B shareholders will hold 10 votes per share, concentrating voting power among Musk and a small group of insiders, while Class A shares sold to public investors will carry one vote each. The documents also outline provisions that could restrict shareholders' ability to influence board elections or pursue certain legal actions, directing disputes into arbitration and limiting where they can be filed. While such dual-class structures are common among founder-led technology companies, they generally reduce the ability of public shareholders to influence company strategy or challenge management decisions.
Musk bought $1.4 billion SpaceX shares last year, The Information reports By Reuters
April 21 (Reuters) - Elon Musk increased his stake in SpaceX last year by purchasing $1.4 billion worth of stock from current and former employees, The Information reported on Tuesday. The secondary stock purchase, made through Musk's trust, was disclosed in a draft of SpaceX's confidential IPO prospectus, the report said. SpaceX also approved a plan last month that would award the billionaire CEO 60 million additional shares if the company's market capitalization climbs from $1.1 trillion to as high as $6.6 trillion and completes an ambitious plan of building data centers in space to supply compute for AI developers, The Information said. Reuters could not immediately verify the report. SpaceX did not immediately respond to a request for comment.
German central bank chief calls for wide access to Anthropic's Mythos
FRANKFURT, April 21 (Reuters) - German central bank chief Joachim Nagel called on Tuesday for all institutions to have access to Anthropic's artificial intelligence model Mythos to keep the playing field even and to avoid it being misused. The Bundesbank head said banking authorities must act to prevent the misuse of Anthropic's most advanced AI model to date, as it opens the door to new and sophisticated cyber risks. "Mythos is an AI model that appears capable of quickly identifying and exploiting security vulnerabilities in financial institutions' software," Nagel said in a speech. Mythos has sparked fears across the banking industry that it could be misused to exploit legacy IT system vulnerabilities. "This AI model seems to be a double-edged sword, since it could be used not only to improve digital security systems, but also to leverage their vulnerabilities for malicious purposes," Nagel said at an event in Rome. The capabilities of Mythos to code at a high level have given it a potentially unprecedented ability to identify cybersecurity vulnerabilities, experts say, prompting greater scrutiny from regulators globally. Anthropic has rolled out a preview of Mythos to a select number of companies and some organisations that build or maintain critical software infrastructure, prompting calls for wider access to the technology. "All relevant institutions should have access to such technology to avoid competitive distortions," Nagel said. Its advanced coding and autonomous capabilities could dramatically accelerate sophisticated cyberattacks, particularly in sectors such as banking that rely on complex, interconnected and often decades-old technology systems, experts have said. While debuting Mythos, Anthropic said the model's ability to find software flaws at scale could, if misused, pose serious risks to economies, public safety and national security. In broader comments on AI, Nagel challenged the notion it could help lower inflation, the core focus of central banks. He said AI increases investment demand, could raise incomes and push up electricity prices, all of which may increase inflationary pressures. Moreover, the use of algorithms may facilitate the setting of prices above competitive levels, Nagel warned. "There is evidence that AI algorithms are able to consistently learn to charge excessive prices, without communicating with one another," he said. "From a central banking perspective, this uncertainty calls for particular vigilance," Nagel added. (Reporting by Balazs Koranyi, Editing by Louise Heavens and Alexander Smith)
Musk bought $1.4 billion SpaceX shares last year, The Information reports
April 21 (Reuters) - Elon Musk increased his stake in SpaceX last year by purchasing $1.4 billion worth of stock from current and former employees, The Information reported on Tuesday. The secondary stock purchase, made through Musk's trust, was disclosed in a draft of SpaceX's confidential IPO prospectus, the report said. SpaceX also approved a plan last month that would award the billionaire CEO 60 million additional shares if the company's market capitalization climbs from $1.1 trillion to as high as $6.6 trillion and completes an ambitious plan of building data centers in space to supply compute for AI developers, The Information said. Reuters could not immediately verify the report. SpaceX did not immediately respond to a request for comment. Reporting by Chandni Shah in Bengaluru; Editing by Janane Venkatraman and Mrigank Dhaniwala Our Standards: The Thomson Reuters Trust Principles., opens new tab
Musk's SpaceX tries to woo Wall Street with analyst meetings this week, sources say
SpaceX is moving ahead with plans for one of the most anticipated initial public offerings in history as it hosts analysts this week for three days of closed-door meetings at its launch facility in Texas and mega-size data center in Tennessee, according to three people familiar with the matter. Billionaire Elon Musk's company is holding the briefings for Wall Street's top aerospace and technology analysts as it looks to raise $75 billion, in what would be the world's biggest-ever IPO, with executives targeting a late June trading debut. The presentations kick off with an all-day meeting and analyst tour on Tuesday at the satellite and rocket maker's Starbase launch facilities in Boca Chica, Texas, the people said.
SpaceX bought around 20% of all Cybertrucks sold in Q4
It's hard to look at the Tesla Cybertruck as a succes, looking at sales figures over the past couple of years since its debut. But there is of course one way to prop up sales, and that's for Elon Musk to buy them himself. And that's, seemingly, kind of what he has done. A new report by Bloomberg suggests that nearly one in five Cybertrucks sold in late 2025 was bought by Elon Musk's own companies, with SpaceX doing most of the heavy lifting. Out of roughly 7000 trucks registered in the US during Q4, around 1279 went to SpaceX, with additional units snapped up by other Musk-linked firms like xAI, Neuralink and The Boring Company. That puts the total at just under 20% of all sales for the quarter. And it didn't stop there. The same pattern has continued into 2026, with more Cybertrucks being registered to those companies in the early months of the year. That makes for a murkier sales picture than before, now that we know Musk and his companies owns a significant chunk of the total amount of Cybertrucks in the US. Without those internal purchases, Cybertruck sales would look significantly weaker. In fact, some estimates suggest registrations could have dropped by more than half if you removed those company buys.
Amazon Expands Anthropic Partnership With US$5 Billion Investment
WASHINGTON, April 21 (Bernama-dpa) -- Amazon is expanding its partnership with OpenAI rival Anthropic with another multibillion-dollar investment, the company said on Monday. Amazon said it would initially invest US$5 billion in the developer of the Claude artificial intelligence (AI) chatbot, with further investments of up to US$20 billion possible in the future, German Press Agency (dpa) reported. At the same time, Anthropic will secure up to 5 gigawatts of capacity from current and future generations of Amazon's Trainium chips to train and power its advanced AI models, Amazon said.
Exclusive-Musk and insiders to retain voting control of SpaceX after IPO, filing shows
NEW YORK, April 20 (Reuters) - SpaceX plans to cement founder Elon Musk's control after its IPO, granting him and a small group of insiders super-voting shares that will outweigh other investors, according to excerpts of the company's IPO filing reviewed by Reuters. The prospectus, which was confidentially filed this month, provides fresh details of the company's financials and corporate governance. Upon completion of the offering, Musk will stay on as chief executive officer, chief technical officer, and will serve as chairman of SpaceX's nine-member board of directors. Though Musk was paid $54,080 last year, according to the excerpts, he stands to gain billions in equity after the company's stock market debut. SpaceX is targeting a listing valuation of roughly $1.75 trillion with a $75 billion raise, which would make it the largest initial public offering in history. President and Chief Operating Officer Gwynne Shotwell received $85.8 million in total compensation last year, Reuters previously reported, while Chief Financial Officer Bret Johnsen was paid $9.8 million. ANALYST DAY Some of the executives are driving Musk's IPO ambitions with three days of meetings planned this week for Wall Street analysts, starting with a tour and briefings at SpaceX's Starbase launch facility in Boca Chica, Texas. The filing excerpts show SpaceX will use a dual-class equity structure that gives Class B shareholders 10 votes each, concentrating power with Musk and a handful of other insiders, while Class A shares sold to public investors will carry one vote each. They also outline provisions that could limit shareholders' ability to influence board elections or pursue certain legal claims, forcing disputes into arbitration instead and restricting where they can be brought. While such structures are common among founder-led technology companies, they limit public shareholders' ability to influence strategy or challenge management. FIRST LOOK AT FINANCIALS The filing gives investors the first look at SpaceX's financial health, especially after Musk combined the rocket maker with his social media and AI company xAI this year. The combined company ended 2025 with about $24.8 billion in cash on hand, and had total assets of $92 billion against total liabilities of $50.8 billion. Its satellite internet business Starlink generated billions in profit last year, helping to offset heavy losses inherited when it bought founder Elon Musk's social media and artificial intelligence company xAI this year, the excerpts show.
Environmental Group To Protest At Starbase As Investors Visit SpaceX Ahead Of IPO
Texas-based environmental rights group The South Texas Environmental Justice Network said on Monday that it will hold protests outside SpaceX's HQ in Starbase as investors visit the facility ahead of the company's IPO. Pressuring Pension Funds To Avoid IPO The group's co-founder, Bekah Hinojosa, is urging investors to boycott the IPO amid concerns about the company's environmental impact, while also lobbying the New York City Comptroller's office to avoid the IPO for the city's pension fund, Reuters reported on Monday. Hinjosa, a resident of Brownsville, which is a town near Starbase, shared that she feels like the town is "being bombed by Elon Musk." The report said that a failed launch at the HQ in 2023 had sent a cloud of concrete over a nearby town, as well as started a fire spanning 3.5 acres on the Boca Chica State Park land. Cathie Wood Backs IPOEnvironmental Concerns With xAI The National Association for the Advancement of Colored People (NAACP) recently filed a federal lawsuit against xAI, accusing the artificial intelligence startup of polluting the air in the Memphis area. The group had earlier opposed the data center deploying gas-operated turbines, releasing harmful emissions, which were disproportionately impacting the majority-black communities in the area. Check out more of Benzinga's Future Of Mobility coverage by following this link. Photo courtesy: Shutterstock Market News and Data brought to you by Benzinga APIs To add Benzinga News as your preferred source on Google, click here.
Amazon puts up to $25 billion more into Anthropic
The deal is Amazon's second consecutive mega-round with a frontier AI lab, following a near-identical $50 billion investment in OpenAI two months ago. Anthropic's annualised revenue has reached $30 billion. The arrangement locks Anthropic into AWS's Trainium chips through Trainium4 and secures up to 5 gigawatts of compute capacity for Claude. Amazon has agreed to invest up to $25 billion in Anthropic, the AI safety company behind the Claude family of models, as part of an expanded agreement that simultaneously locks in more than $100 billion of cloud spending by Anthropic on Amazon Web Services over the next decade. Amazon is making an immediate investment of $5 billion, with up to $20 billion more tied to commercial milestones. This is on top of the $8 billion Amazon had already committed to Anthropic since 2023, bringing total potential Amazon investment to approximately $33 billion. The structure of the deal is worth examining. Amazon is not writing a cheque for $25 billion outright; the bulk of the new capital is contingent on performance. Simultaneously, Anthropic's $100 billion spending commitment to AWS is not a payment but a procurement pledge, a commitment to buy cloud compute, custom AI chips, and infrastructure services from Amazon over ten years. Amazon funds Anthropic, and Anthropic spends that capital on Amazon's infrastructure. The same structure was used two months ago when Amazon invested $50 billion in OpenAI as part of a comparable $100 billion cloud commitment. The compute at the centre of the agreement is Amazon's Trainium chip line, its in-house alternative to Nvidia GPUs for AI training and inference. The deal covers current Trainium2, the forthcoming Trainium3, and Trainium4, which has not yet been released. It also includes tens of millions of Graviton CPU cores. Anthropic secures up to 5 gigawatts of capacity to train and deploy its Claude models under the deal. Amazon said that significant Trainium2 capacity is coming online in the second quarter of 2026, with nearly one gigawatt of combined Trainium2 and Trainium3 expected by year-end. The compute commitment is a direct response to what Anthropic describes as rapidly growing demand: the company's annualised revenue has reached $30 billion, up from approximately $9 billion at the end of 2025. The partnership also deepens how Anthropic's models are distributed. AWS customers will now be able to access the full Claude Platform, Anthropic's native product interface, directly through their existing AWS accounts, without separate credentials or billing. This is a step beyond having Claude available through Amazon Bedrock, Amazon's managed AI marketplace. Claude remains the only frontier model available across all three major cloud platforms: AWS Bedrock, Google Cloud Vertex AI, and Microsoft Azure Foundry. The geopolitical backdrop matters to this deal. Anthropic is currently barred from Department of Defense contracts following a supply-chain risk designation that it is contesting in court. That dispute sits alongside a broader AI policy environment in which the Trump administration has been simultaneously pressuring AI companies to loosen safety constraints and courting them for national security applications. Dario Amodei, Anthropic's CEO, has been navigating that tension publicly. The commercial momentum of the AWS deal, and the implicit validation of Anthropic's infrastructure scale, is one answer to critics, including OpenAI, who suggested last week that Anthropic had made a strategic mistake by not securing enough compute. In November 2025, Microsoft invested up to $5 billion in Anthropic as part of a separate deal in which Anthropic committed to purchasing $30 billion of Azure compute capacity. The AWS deal dwarfs that arrangement and firmly establishes Amazon as Anthropic's primary infrastructure partner. The initial investment in this round is priced at Anthropic's latest valuation of $380 billion. Venture capital firms have reportedly been offering capital at valuations of $800 billion or more ahead of a potential IPO, though no such round has been announced.
Europe must prevent misuse of Anthropic's Mythos, Bundesbank chief warns
FRANKFURT, April 21 (Reuters) - Banking authorities must prevent the misuse of Anthropic's Mythos, its most advanced AI model to date, which opens the door to new and sophisticated cyber risks, Bundesbank President Joachim Nagel said on Tuesday. "Mythos is anAImodel that appears capable of quickly identifying and exploiting security vulnerabilities in financial institutions' software," Nagel said in a speech. "However, thisAImodel seems to be a double-edged sword, since it could be used not only to improve digital security systems, but also to leverage theirvulnerabilities for malicious purposes," he added. He added that all relevant institutions should have access to such technology to avoid competitive distortions. (Reporting by Balazs Koranyi, Editing by Louise Heavens)
Exclusive-Musk and insiders to retain voting control of SpaceX after IPO, filing shows By Reuters
NEW YORK, April 20 (Reuters) - SpaceX plans to cement founder Elon Musk's control after its IPO, granting him and a small group of insiders super-voting shares that will outweigh other investors, according to excerpts of the company's IPO filing reviewed by Reuters. The prospectus, which was confidentially filed this month, provides fresh details of the company's financials and corporate governance. Upon completion of the offering, Musk will stay on as chief executive officer, chief technical officer, and will serve as chairman of SpaceX's nine-member board of directors. Though Musk was paid $54,080 last year, according to the excerpts, he stands to gain billions in equity after the company's stock market debut. SpaceX is targeting a listing valuation of roughly $1.75 trillion with a $75 billion raise, which would make it the largest initial public offering in history. President and Chief Operating Officer Gwynne Shotwell received $85.8 million in total compensation last year, Reuters previously reported, while Chief Financial Officer Bret Johnsen was paid $9.8 million. ANALYST DAY Some of the executives are driving Musk's IPO ambitions with three days of meetings planned this week for Wall Street analysts, starting with a tour and briefings at SpaceX's Starbase launch facility in Boca Chica, Texas. The filing excerpts show SpaceX will use a dual-class equity structure that gives Class B shareholders 10 votes each, concentrating power with Musk and a handful of other insiders, while Class A shares sold to public investors will carry one vote each. They also outline provisions that could limit shareholders' ability to influence board elections or pursue certain legal claims, forcing disputes into arbitration instead and restricting where they can be brought. While such structures are common among founder-led technology companies, they limit public shareholders' ability to influence strategy or challenge management. FIRST LOOK AT FINANCIALS The filing gives investors the first look at SpaceX's financial health, especially after Musk combined the rocket maker with his social media and AI company xAI this year. The combined company ended 2025 with about $24.8 billion in cash on hand, and had total assets of $92 billion against total liabilities of $50.8 billion. Its satellite internet business Starlink generated billions in profit last year, helping to offset heavy losses inherited when it bought founder Elon Musk's social media and artificial intelligence company xAI this year, the excerpts show. SpaceX swung to a $4.94 billion consolidated loss in 2025 on revenue of $18.67 billion as it invested heavily in xAI's artificial intelligence infrastructure, from a $791 million profit and $14.02 billion in revenue the year before. It lost $4.63 billion on $10.4 billion in revenue in 2023. AI SPENDING Its losses stem from an almost fivefold increase in capital spending over two years to $20.74 billion last year, more than half of that on AI spending. The company's successful Starlink satellite internet service is subsidizing much of that spending, generating $4.42 billion in operating profit but accounting for less than a quarter of its total capital expenditures. Capital expenditure at the AI segment surged to $12.7 billion from $5.6 billion the prior year, pushing SpaceX's total capex above $20.7 billion, more than double the prior year. That remains a fraction of spending by the largest technology companies on AI infrastructure: Meta, with a comparable market capitalization of about $1.7 trillion, had capital expenditure of $72 billion in 2025. The Information previously reported some aspects of the financials. SpaceX did not immediately respond to a request for comment.
