News & Updates
The latest news and updates from companies in the WLTH portfolio.
Fluor Signs Contract With X-Energy for Advanced Nuclear Project in Texas
IRVING, Texas--(BUSINESS WIRE)--#Energy--Fluor Corporation (NYSE: FLR) announced today that it has entered into a contract with X-energy to support the company's proposed advanced nuclear project at Dow's UCC Seadrift Operations in south Texas. Under the agreement, Fluor will initially deliver Front-End Loading Stage 2 (FEL-2) services. FEL-2 focuses on project definition, strategic planning, feasibility assessment, cost control and risk mitigation. Fluor will recognize the undisclosed contract value for this initial portion of work in the first quarter of 2026. The X-energy project proposes to develop four, 80-megawatt small modular reactor (SMR) units to supply Dow's Seadrift site with safe, reliable, carbon-free electricity and industrial steam, replacing aging energy and steam infrastructure. The project is supported by the U.S. Department of Energy's (DOE) Advanced Reactor Demonstration Program (ARDP), which accelerates the commercialization of advanced nuclear technologies through cost-shared partnerships with industry. A construction permit application was submitted in March 2025 and is currently being reviewed by the U.S. Nuclear Regulatory Commission. "X‑energy's technology offers a powerful pathway for small modular reactors to deliver safe, reliable and fit-for-purpose baseload power in an industrial setting," said Pierre Bechelany, Fluor's Business Group President of Energy Solutions. "With eight decades of nuclear experience, Fluor brings the proven expertise and disciplined execution required to help advance this landmark project." X-energy was selected by the DOE in 2020 to develop, license and build its XE-100 advanced SMR and a first TRISO-X fuel fabrication facility. Since then, the company has completed engineering and preliminary reactor design, advanced development and licensing of its fuel facility in Oak Ridge, Tennessee. The Seadrift project is expected to become the first grid-scale advanced nuclear reactor deployed to serve an industrial facility in North America. Dow's UCC Seadrift Operations span 4,700 acres and produce more than 4 billion pounds of materials annually for applications including food packaging, footwear, wire and cable insulation, solar cell components, and medical and pharmaceutical packaging. About Fluor Corporation Fluor Corporation (NYSE: FLR) is building a better world by applying world-class expertise to solve its clients' greatest challenges. Fluor's nearly 23,000 employees provide professional and technical solutions that deliver safe, well-executed, capital-efficient projects to clients around the world. Fluor had revenue of $15.5 billion in 2025 and is ranked 257 among the Fortune 500 companies. With headquarters in Irving, Texas, Fluor has provided engineering, procurement, construction and maintenance services for more than a century. For more information, please visit www.fluor.com or follow Fluor on Facebook, Instagram, LinkedIn, X and YouTube. #EnergySolutions Brett Turner Media Relations 864.281.6976 Jason Landkamer Investor Relations 469.398.7222
Trump's 'craziest ever' press conference devolves into chaos as he promises to destro
Trump's 'craziest ever' press conference devolves into chaos as he promises to destroy Iran in one night https://www.dailymail.co.uk/news/article...-Iran.html President Donald Trump has sparked disbelief and mockery after an extraordinary White House briefing on Monday that critics claimed was more reminiscent of an SNL skit. Trump took the White House podium to detail the 'breathtaking' rescue of two US pilots shot down over Iran, showing the scope and resolve of the American fighting force. YouTube personality and comedian Jimmy Dore led the charge, calling the administration's narrative a 'clown show' after Trump claimed total military dominance despite the recent downing of a US fighter jet. 'Pretending that getting a fighter jet shot down by a military that Trump said was 'Completely Decimated' as a victory is pure clown show,' Dore wrote on X. 'Iran is humiliated? Everything they say is projection and a confession.' Other high-traffic financial accounts on X echoed the sentiment, with one viral post stating: 'That was one of the craziest press conferences Trump has ever given. Absolute comedy.' The comments come as Trump signals a frantic, optimistic push for a peace deal with what he calls a 'decapitated' Iranian regime - while simultaneously authorizing the most 'ferocious' wave of American strikes to date. And he wouldn't disclose whether he'll be winding down the bombing campaign. 'I don't know. I can't tell. It depends what they do,' Trump said, further sparking confusion. The President warned that the entire country of Iran could be 'taken out' in one night, adding that it could be on Tuesday
SpaceX Raptor Engine Test Seems to Have an Explosion | NextBigFuture.com
Brian Wang is a Futurist Thought Leader and a popular Science blogger with 1 million readers per month. His blog Nextbigfuture.com is ranked #1 Science News Blog. It covers many disruptive technology and trends including Space, Robotics, Artificial Intelligence, Medicine, Anti-aging Biotechnology, and Nanotechnology. Known for identifying cutting edge technologies, he is currently a Co-Founder of a startup and fundraiser for high potential early-stage companies. He is the Head of Research for Allocations for deep technology investments and an Angel Investor at Space Angels. A frequent speaker at corporations, he has been a TEDx speaker, a Singularity University speaker and guest at numerous interviews for radio and podcasts. He is open to public speaking and advising engagements.
Anthropic Ends Free Third-Party Access for Claude Code Subscribers
Anthropic has ended a pricing arrangement that allowed Claude Code subscribers to run the artificial intelligence assistant through external tools at no additional cost, requiring developers who rely on third-party frameworks such as OpenClaw to pay separately from their existing subscription. According to a customer email shared on Hacker News, Anthropic said that starting at noon Pacific on April 4, subscribers would "no longer be able to use your Claude subscription limits for third-party harnesses including OpenClaw." Instead, they must pay for extra usage through "a pay-as-you-go option billed separately from your subscription." The company said the policy applies to all third-party harnesses and will be extended to more tools shortly. Boris Cherny, head of Claude Code at Anthropic, said the company's "subscriptions weren't built for the usage patterns of these third-party tools" and that Anthropic is "prioritising our customers using our products and API." The technical justification centres on how agentic frameworks interact with Anthropic's infrastructure. A single OpenClaw instance running autonomously for a full day, browsing the web, managing calendars, responding to messages and executing code, can consume the equivalent of $1,000 to $5,000 in application programming interface (API) costs depending on the task load, making flat-rate subscription coverage unsustainable. To ease the transition, Anthropic is offering a one-time credit for extra usage equal to a user's monthly subscription price, redeemable until April 17, along with discounts of up to 30 percent on extra usage bundles. The announcement has drawn sharp criticism from parts of the developer community. OpenClaw creator Peter Steinberger, who recently joined rival OpenAI, said he and OpenClaw board member Dave Morin "tried to talk sense into Anthropic" but were only able to delay enforcement by a week. "Funny how timings match up, first they copy some popular features into their closed harness, then they lock out open source," Steinberger said. Anthropic's restrictions were announced and enforced within weeks of Steinberger's move to OpenAI in February 2026, a timeline that has not escaped notice. Sam Altman posted publicly that Steinberger would drive the next generation of personal agents at OpenAI, and that OpenClaw would be moved to an open-source foundation with OpenAI's continued support. Cherny pushed back on the characterisation, insisting that Claude Code team members are "big fans of open source" and that he personally contributed pull requests to improve prompt cache efficiency for OpenClaw specifically. "This is more about engineering constraints," he said. The move underscores a growing tension at the heart of the AI industry: power users want autonomous agents that run constantly, but AI companies are working to control costs, capacity and how their models are deployed. What industry observers have described as the "all-you-can-eat buffet" model of AI subscription pricing is giving way to more granular, usage-based billing as computational demands from autonomous agents continue to escalate.
Anthropic Expands Use of Google Cloud and TPUs
Latest expansion will provide Anthropic with multiple gigawatts of TPU capacity SUNNYVALE, Calif., April 6, 2026 /PRNewswire/ -- Anthropic today announced an expansion of its use of TPU chips and cloud services, as it scales its development of foundation models, agents, and enterprise applications. The expansion will provide Anthropic with multiple gigawatts of TPU capacity, expected to come online starting in 2027. This capacity expansion will be delivered through Google Cloud services, as well as access to Google-built TPUs supplied through Broadcom. The additional TPU capacity will support rapidly scaling needs for Anthropic's models. Anthropic also continues to grow its use of Google Cloud's broader cloud and AI solutions, including BigQuery, Cloud Run, AlloyDB, and others, which together help power Anthropic's data, AI development, and applications. Today, thousands of customers access Claude models through Google Cloud, including Coinbase, Cursor, Palo Alto Networks, Replit, and Shopify. About Google Cloud Google Cloud is the new way to the cloud, providing AI, infrastructure, developer, data, security, and collaboration tools built for today and tomorrow. Google Cloud offers a powerful, fully integrated and optimized AI stack with its own planet-scale infrastructure, custom-built chips, generative AI models and development platform, as well as AI-powered applications, to help organizations transform. Customers in more than 200 countries and territories turn to Google Cloud as their trusted technology partner. View original content to download multimedia:https://www.prnewswire.com/news-releases/anthropic-expands-use-of-google-cloud-and-tpus-302735047.html
Anthropic Expands Use of Google Cloud and TPUs
Latest expansion will provide Anthropic with multiple gigawatts of TPU capacity SUNNYVALE, Calif., April 6, 2026 /PRNewswire/ -- Anthropic today announced an expansion of its use of TPU chips and cloud services, as it scales its development of foundation models, agents, and enterprise applications. The expansion will provide Anthropic with multiple gigawatts of TPU capacity, expected to come online starting in 2027. This capacity expansion will be delivered through Google Cloud services, as well as access to Google-built TPUs supplied through Broadcom. The additional TPU capacity will support rapidly scaling needs for Anthropic's models. Anthropic also continues to grow its use of Google Cloud's broader cloud and AI solutions, including BigQuery, Cloud Run, AlloyDB, and others, which together help power Anthropic's data, AI development, and applications. Today, thousands of customers access Claude models through Google Cloud, including Coinbase, Cursor, Palo Alto Networks, Replit, and Shopify. About Google Cloud Google Cloud is the new way to the cloud, providing AI, infrastructure, developer, data, security, and collaboration tools built for today and tomorrow. Google Cloud offers a powerful, fully integrated and optimized AI stack with its own planet-scale infrastructure, custom-built chips, generative AI models and development platform, as well as AI-powered applications, to help organizations transform. Customers in more than 200 countries and territories turn to Google Cloud as their trusted technology partner. View original content to download multimedia:https://www.prnewswire.com/news-releases/anthropic-expands-use-of-google-cloud-and-tpus-302735047.html
SpaceX' IPO pitch centers on Elon Musk's ability to 'sell the dream'
SpaceX and investment bankers will host meetings to pressure-test the Elon Musk-led company's targeted $2 trillion valuation, a crucial step as it drives toward the largest-ever initial public offering, people familiar with the matter said. Two months after a merger with xAI, senior bankers are poised to begin talks around the rocket, satellite and AI company's initial public offering, to determine if its vaunted valuation will still lure in investors. Its market value is now expected to top $2 trillion, up from expectations of $1.75 trillion less than two weeks ago, people familiar with the matter said.
Anthropic expands partnership with Google and Broadcom for multiple gigawatts of next-generation compute
We have signed a new agreement with Google and Broadcom for multiple gigawatts of next-generation TPU capacity that we expect to come online starting in 2027. This significant expansion of our compute infrastructure will power our frontier Claude models and help us serve extraordinary demand from customers worldwide. "This groundbreaking partnership with Google and Broadcom is a continuation of our disciplined approach to scaling infrastructure: we are building the capacity necessary to serve the exponential growth we have seen in our customer base while also enabling Claude to define the frontier of AI development," said Krishna Rao, CFO of Anthropic. "We are making our most significant compute commitment to date to keep pace with our unprecedented growth." Demand from Claude customers has accelerated in 2026. Our run-rate revenue has now surpassed $30 billion -- up from approximately $9 billion at the end of 2025. When we announced our Series G fundraising in February, we shared that over 500 business customers were each spending over $1 million on an annualized basis. Today that number exceeds 1,000, doubling in less than two months. The vast majority of the new compute will be sited in the United States, making this partnership a major expansion of our November 2025 commitment to invest $50 billion in strengthening American computing infrastructure. The partnership deepens our existing work with Google Cloud -- building on the increased TPU capacity we announced last October -- as well as our relationship with Broadcom. We train and run Claude on a range of AI hardware -- AWS Trainium, Google TPUs, and NVIDIA GPUs -- which means we can match workloads to the chips best suited for them. This diversity of platforms translates to better performance and greater resilience for customers who depend on Claude for critical work. Amazon remains our primary cloud provider and training partner, and we continue to work closely with AWS on Project Rainier. Claude remains the only frontier AI model available to customers on all three of the world's largest cloud platforms: Amazon Web Services (Bedrock), Google Cloud (Vertex AI), and Microsoft Azure (Foundry).
Anthropic inks deal with Google, Broadcom for TPU capacity
Anthropic said it has inked an agreement with Google and Broadcom to for TPU-powered compute that'll come online in 2027. Anthropic has been expanding its compute infrastructure due to surging demand. The demand has been so strong that it has had to put caps on Claude Code tokens. That demand is also turning up in Anthropic's annual revenue run rate, which is now $30 billion, up from $9 billion at the end of 2025. Krishna Rao, CFO of Anthropic, said in a blog post that the extra capacity is needed "to serve the exponential growth we have seen in our customer base." In addition, Anthropic said it has doubled its base of business customers to top 1,000, double the tally from just two months ago. Anthropic added that most of the new compute will be in the US. Anthropic has a diverse base of suppliers. It works with Google Cloud and Broadcom for TPUs and runs and trains its Claude models on AWS Trainium, Google TPUs and Nvidia GPUs. The company said: "We train and run Claude on a range of AI hardware -- AWS Trainium, Google TPUs, and NVIDIA GPUs -- which means we can match workloads to the chips best suited for them. This diversity of platforms translates to better performance and greater resilience for customers who depend on Claude for critical work. Amazon remains our primary cloud provider and training partner, and we continue to work closely with AWS on Project Rainier. Claude remains the only frontier AI model available to customers on all three of the world's largest cloud platforms: Amazon Web Services (Bedrock), Google Cloud (Vertex AI), and Microsoft Azure (Foundry)." Google and Broadcom confirmed the Anthropic deal.
Federal workers can log back in to Anthropic's Claude following US injunction | MLex | Specialist news and analysis on legal risk and regulation
By Mike Swift ( April 6, 2026, 22:24 GMT | Insight) -- US federal workers are being told they can log back into Anthropic's Claude for Government service in the wake of a ruling by a California federal judge who granted an injunction blocking the Defense Department and other areas of the Trump administration from declaring Anthropic a supply chain risk.Federal workers who lost access to Anthropic's Claude for Government service are being notified they can once again use the generative artificial intelligence service following a US judge's order to block the Trump administration's designation of the company as a supply chain risk.... Prepare for tomorrow's regulatory change, today MLex identifies risk to business wherever it emerges, with specialist reporters across the globe providing exclusive news and deep-dive analysis on the proposals, probes, enforcement actions and rulings that matter to your organization and clients, now and in the longer term. Know what others in the room don't, with features including: * Daily newsletters for Antitrust, M&A, Trade, Data Privacy & Security, Technology, AI and more * Custom alerts on specific filters including geographies, industries, topics and companies to suit your practice needs * Predictive analysis from expert journalists across North America, the UK and Europe, Latin America and Asia-Pacific * Curated case files bringing together news, analysis and source documents in a single timeline Experience MLex today with a 14-day free trial.
Anthropic Expands Use of Google Cloud and TPUs | Weekly Voice
Latest expansion will provide Anthropic with multiple gigawatts of TPU capacity SUNNYVALE, Calif., April 6, 2026 /PRNewswire/ -- Anthropic today announced an expansion of its use of TPU chips and cloud services, as it scales its development of foundation models, agents, and enterprise applications. The expansion will provide Anthropic with multiple gigawatts of TPU capacity, expected to come online starting in 2027. This capacity expansion will be delivered through Google Cloud services, as well as access to Google-built TPUs supplied through Broadcom. The additional TPU capacity will support rapidly scaling needs for Anthropic's models. Anthropic also continues to grow its use of Google Cloud's broader cloud and AI solutions, including BigQuery, Cloud Run, AlloyDB, and others, which together help power Anthropic's data, AI development, and applications. Today, thousands of customers access Claude models through Google Cloud, including Coinbase, Cursor, Palo Alto Networks, Replit, and Shopify. About Google Cloud Google Cloud is the new way to the cloud, providing AI, infrastructure, developer, data, security, and collaboration tools built for today and tomorrow. Google Cloud offers a powerful, fully integrated and optimized AI stack with its own planet-scale infrastructure, custom-built chips, generative AI models and development platform, as well as AI-powered applications, to help organizations transform. Customers in more than 200 countries and territories turn to Google Cloud as their trusted technology partner. View original content to download multimedia:https://www.prnewswire.com/news-releases/anthropic-expands-use-of-google-cloud-and-tpus-302735047.html
Broadcom to Supply AI Chips to Google, Computing Capacity to Anthropic in Expanded Collaboration
Broadcom will develop and supply custom artificial-intelligence chips for Google and additional computing capacity to Anthropic in an expansion of the strategic collaboration between the three companies. The chip manufacturer said Monday it will supply Google with custom Tensor Processing Units, along with networking and other components for Google's next-generation AI data center racks through up to 2031 as part of a supply assurance agreement. Anthropic will access about 3.5 gigawatts of TPU-based computing capacity beginning in 2027 as part of its commitment for multiple gigawatts of compute capacity, Broadcom said. The expanded compute capacity is "dependent on Anthropic's continued commercial success," Broadcom said. "In connection with this deployment, the parties are in discussions with certain operational and financial partners," Broadcom added. Broadcom's agreement to supply chips to Google comes as the latter company is looking to expand the market for its TPUs, The Wall Street Journal has reported. Write to Elias Schisgall at [email protected]
Perplexity Launches AI Agent That Drafts Tax Returns | PYMNTS.com
By completing this form, you agree to receive marketing communications from PYMNTS and to the sharing of your information with our sponsor, if applicable, in accordance with our Privacy Policy and Terms and Conditions. Perplexity is trying to change that with Computer for Taxes, extending its artificial intelligence (AI) agent platform into U.S. federal tax preparation for the 2026 filing season. Computer does not answer tax questions in a chat window. Instead, it reviews uploaded financial documents, asks follow-up questions about a filer's situation and maps inputs to official IRS forms and drafts a federal return, according to the company's Thursday (April 2) announcement. Users can access it by selecting "Navigate my taxes" inside Perplexity Computer, which requires a Pro subscription at $17 per month. The distinction between Computer and general-purpose AI chatbots is structural. Tools like ChatGPT or Gemini respond to tax questions based on training data that has a fixed cutoff date and no direct connection to current IRS materials. A test of four major AI chatbots conducted by TaxSlayer and reported by The New York Times found the tools miscalculated the refund or amount owed by an average of more than $2,000 across eight fictional tax scenarios, even when provided with all necessary forms. Perplexity addresses this by packaging tax knowledge as loadable modules built on its Agent Skills protocol, according to the company. These modules are continuously updated and grounded in IRS materials and regulations, allowing the system to retrieve and apply current rules rather than rely on static model training. When tax laws change, the modules are updated independently, without retraining the underlying model. Perplexity also said Computer for Taxes can also review a professionally prepared return for accuracy and compliance, helping filers identify questions for their tax provider and confirm no money is left on the table. The company added that many tax professionals are still adjusting to recent changes in tax laws, and that Computer is designed to surface those gaps. The workflow begins with uploading documents. Computer then reviews the files, asks questions about the filer's financial situation and drafts a return on official IRS forms. The system can also build dashboards and tools for more complex parts of the tax code and support other filing workflows users want to design, Neowin reported. Perplexity positioned Computer for Taxes as a way to reduce the burden of self-filing. People pay for software, spend hours doing the work and often still need professional help when a return grows more complicated, the company said. Computer is designed to reduce that friction across the full workflow. The product arrives as consumer behavior around AI and taxes is shifting. As the April 15 tax deadline approaches, consumers who once turned to TurboTax or a CPA as their first resource are now opening ChatGPT, Gemini or Claude to ask whether they qualify for a deduction or what their refund might look like, PYMNTS reported. About a quarter of U.S. workers say they plan to use AI to help file their taxes this year, per the report, more than double the 11% who said the same last year. According to PYMNTS Intelligence, 62% of Generation Z consumers are open to using AI for financial planning guidance, a sign that generational appetite for on-demand financial tools now extends to filing season. What Perplexity is building is the next step in that progression. Computer for Taxes is built for the filers already asking AI for guidance and are ready to hand it the forms.
Broadcom announces long-term chip supply agreements with Google and Anthropic By Investing.com
Broadcom Inc. (NASDAQ:AVGO) announced Monday that it has entered into a long-term agreement with Google LLC to develop and supply custom Tensor Processing Units (TPUs) for future generations of Google's TPUs. The agreement also includes a supply assurance arrangement for Broadcom to provide networking and other components for use in Google's next-generation artificial intelligence racks through up to 2031. In a separate development, Broadcom, Google, and Anthropic PBC have expanded their current strategic collaboration. Beginning in 2027, Anthropic will access approximately 3.5 gigawatts of next-generation TPU-based AI compute capacity through Broadcom, as part of Anthropic's commitment to multiple gigawatts of AI capacity. The use of this expanded AI compute capacity by Anthropic is dependent on its continued commercial success, according to the filing. The parties are also in discussions with certain operational and financial partners in connection with this deployment. These announcements were disclosed in a press release statement based on Broadcom's filing with the Securities and Exchange Commission. In other recent news, Broadcom Inc. has announced a significant transition in its executive team. Amie Thuener will take over as Chief Financial Officer, succeeding Kirsten M. Spears, who is set to retire in June 2026 after 12 years with the company. Thuener, previously with Alphabet, will officially begin her role on June 12, 2026. In another development, Broadcom and Carahsoft Technology Corp. secured a five-year, $970 million contract with the Defense Information Systems Agency to provide cloud software solutions to various Department of War agencies. This agreement aims to consolidate software contracts and offer uniform pricing for private cloud infrastructure and security solutions. Meanwhile, Erste Group has downgraded Broadcom's stock rating from buy to hold, citing concerns about the company's software segment growth prospects. Despite this, Broadcom projects second-quarter sales to reach approximately $22.0 billion, marking a 47% increase from the previous year. The company also anticipates AI semiconductor sales to hit $10.7 billion for the quarter. Additionally, Seaport's Chief Equity Strategist, Jonathan Golub, suggested that Broadcom, alongside other semiconductor companies, presents upside opportunities as software stocks are expected to rebound. This article was generated with the support of AI and reviewed by an editor. For more information see our T&C.
Broadcom announces long-term chip supply agreements with Google and Anthropic By Investing.com
Broadcom Inc. (NASDAQ:AVGO) announced Monday that it has entered into a long-term agreement with Google LLC to develop and supply custom Tensor Processing Units (TPUs) for future generations of Google's TPUs. The agreement also includes a supply assurance arrangement for Broadcom to provide networking and other components for use in Google's next-generation artificial intelligence racks through up to 2031. In a separate development, Broadcom, Google, and Anthropic PBC have expanded their current strategic collaboration. Beginning in 2027, Anthropic will access approximately 3.5 gigawatts of next-generation TPU-based AI compute capacity through Broadcom, as part of Anthropic's commitment to multiple gigawatts of AI capacity. The use of this expanded AI compute capacity by Anthropic is dependent on its continued commercial success, according to the filing. The parties are also in discussions with certain operational and financial partners in connection with this deployment. These announcements were disclosed in a press release statement based on Broadcom's filing with the Securities and Exchange Commission. In other recent news, Broadcom Inc. has announced a significant transition in its executive team. Amie Thuener will take over as Chief Financial Officer, succeeding Kirsten M. Spears, who is set to retire in June 2026 after 12 years with the company. Thuener, previously with Alphabet, will officially begin her role on June 12, 2026. In another development, Broadcom and Carahsoft Technology Corp. secured a five-year, $970 million contract with the Defense Information Systems Agency to provide cloud software solutions to various Department of War agencies. This agreement aims to consolidate software contracts and offer uniform pricing for private cloud infrastructure and security solutions. Meanwhile, Erste Group has downgraded Broadcom's stock rating from buy to hold, citing concerns about the company's software segment growth prospects. Despite this, Broadcom projects second-quarter sales to reach approximately $22.0 billion, marking a 47% increase from the previous year. The company also anticipates AI semiconductor sales to hit $10.7 billion for the quarter. Additionally, Seaport's Chief Equity Strategist, Jonathan Golub, suggested that Broadcom, alongside other semiconductor companies, presents upside opportunities as software stocks are expected to rebound. This article was generated with the support of AI and reviewed by an editor. For more information see our T&C.
Inside Anthropic's Radical Experiment: Where Employees Argue With the CEO on Slack -- and He Argues Back
At most billion-dollar companies, telling the chief executive he's wrong is a career-limiting move. At Anthropic, it's practically a job requirement. The artificial intelligence company behind the Claude chatbot has built an internal culture so aggressively open that employees routinely challenge CEO Dario Amodei in public Slack channels -- debating everything from safety policy to product strategy to the company's public messaging. And Amodei doesn't just tolerate it. He participates, sometimes writing lengthy responses that read more like academic rebuttals than executive pronouncements. This isn't corporate theater. It's a deliberate organizational design choice that Anthropic's leadership believes is essential for a company building technology that could, by its own admission, pose existential risks. But as the company has grown from a scrappy research lab into a $60 billion enterprise with more than 1,000 employees, the question is whether this radical transparency can survive the gravitational pull of corporate normalcy -- or whether it even should. The Slack Debates That Define a Company Business Insider reported that Anthropic's internal Slack channels have become something like a permanent town hall, where rank-and-file researchers and engineers engage in substantive, sometimes heated exchanges with Amodei and other senior leaders. The debates aren't limited to technical questions. Employees have pushed back on the company's approach to government partnerships, its competitive positioning against OpenAI, and even the tone of Amodei's own public essays. One former employee described the dynamic as "the most intellectually honest workplace I've ever seen," while acknowledging it could be "exhausting." Another noted that the open culture attracted a specific kind of person -- someone comfortable with ambiguity and conflict, who viewed disagreement as a feature rather than a bug. The roots of this culture trace directly to Anthropic's origin story. Amodei and his sister Daniela, the company's president, co-founded Anthropic in 2021 after leaving OpenAI, where they had grown concerned about what they saw as insufficient attention to safety and a drift toward commercialization under Sam Altman. They brought roughly a dozen OpenAI researchers with them. The founding team's shared conviction was that building powerful AI safely required an organization where bad news traveled fast and dissent wasn't suppressed. So they built the company around that idea. Default-public Slack channels. A norm of writing long-form internal memos and inviting critique. Leadership that models vulnerability by admitting uncertainty. It's the organizational equivalent of "strong opinions, loosely held" -- except applied to a company worth tens of billions of dollars building systems that can generate code, write legal briefs, and soon, according to Amodei himself, perform at the level of a Nobel Prize-winning scientist. The culture has real consequences for how decisions get made. When Anthropic was weighing its approach to the Responsible Scaling Policy -- the framework it uses to determine when its AI models are too dangerous to deploy without additional safeguards -- the internal debate was extensive and, by several accounts, genuinely influenced the outcome. Engineers who believed the thresholds were too permissive said so. Researchers who thought they were too restrictive pushed the other direction. Amodei engaged with both camps. This stands in stark contrast to what's been reported at OpenAI, where multiple departures -- including co-founder Ilya Sutskever and former board members -- have been linked to disagreements about safety and governance that weren't resolved through open debate but through boardroom drama and legal maneuvering. At Google DeepMind, the culture is more traditionally hierarchical, with Demis Hassabis running a tighter ship. And at Meta's AI division, Mark Zuckerberg has made clear that he sets the strategic direction, with the open-source approach reflecting his personal conviction rather than a consensus-building process. Anthropic's model is different. Genuinely different. Whether it's better is a separate question. The Tensions That Growth Creates There's an inherent tension in running a company this way as it scales. When Anthropic had 100 employees, a culture of radical openness was manageable. Everyone knew everyone. Context was shared. Debates could be resolved because the participants understood the full picture. At 1,000-plus employees, it gets harder. Much harder. New hires don't always have the context to engage productively in debates about safety thresholds or deployment decisions. The signal-to-noise ratio in Slack channels degrades. And there's a real risk that the appearance of openness masks a reality where decisions are still made by a small group of senior leaders who have already reached a conclusion -- with the debate serving as a legitimizing exercise rather than a genuine input. Some current and former employees have raised exactly this concern. According to Business Insider's reporting, there's a perception among some staff that while Amodei is genuinely open to argument, the company's rapid commercialization -- driven by the need to justify its massive valuation and compete with OpenAI, Google, and others -- has made it harder for safety-focused dissent to carry the day. The pressure to ship products, sign enterprise deals, and keep up with competitors who are less cautious creates a structural headwind against the kind of slow, deliberative culture Anthropic aspires to maintain. The financial pressures are real. Anthropic has raised more than $15 billion in funding, with Amazon alone committing up to $8 billion. The company's annualized revenue has been growing rapidly, driven by Claude's adoption in enterprise settings. But it's still burning cash at an extraordinary rate -- the cost of training frontier AI models runs into the hundreds of millions of dollars per run, and the compute infrastructure required is staggering. Anthropic needs to grow revenue fast, which means shipping products fast, which means making decisions fast. Fast and open don't always coexist comfortably. There's also the question of what happens when the stakes get truly high. Anthropic's own framework acknowledges that its models could eventually pose catastrophic risks -- biological, cyber, or otherwise. When a company is making decisions about whether to deploy a system that could, in a worst case, help someone synthesize a dangerous pathogen, the "let's debate it on Slack" approach may not be sufficient. Or appropriate. Some decisions require speed, secrecy, and a clear chain of command. The challenge for Anthropic is building the institutional capacity for both modes -- open debate for strategic direction, rapid execution for crisis response -- without one undermining the other. Amodei himself seems aware of this tension. In his widely read essay "Machines of Loving Grace," published in late 2024, he laid out an optimistic vision of AI's potential while acknowledging the risks. The essay itself was reportedly debated extensively within the company before publication, with employees pushing back on specific claims and framing. That's the culture working as intended. But the essay also revealed something about Amodei's leadership style: he's a CEO who thinks in paragraphs, not bullet points, and who treats internal debate as an intellectual exercise as much as a management tool. That works when the CEO is the smartest person in the room -- or at least among the smartest. Amodei, a former VP of Research at OpenAI with a background in computational neuroscience, has the technical chops to engage substantively with his researchers. But not every future leader of the company will. And cultures built around a founder's personal qualities are notoriously fragile. What the Rest of the Industry Can Learn -- and What It Can't Copy Anthropic's experiment matters beyond Anthropic. The AI industry is grappling with a fundamental governance question: how should companies making decisions with potentially civilization-scale consequences be structured? The traditional corporate model -- board oversight, executive authority, shareholder accountability -- wasn't designed for this. Neither was the nonprofit model, as OpenAI's tortured governance history demonstrates. And government regulation, while advancing in the EU and under discussion in the US and UK, remains far behind the pace of technological development. Anthropic's answer -- build a culture where the people closest to the technology can challenge the people making deployment decisions -- is at least a coherent theory. It's grounded in the safety engineering principle that disasters are prevented not by infallible leaders but by systems that surface dissent before it's too late. The aviation industry learned this lesson decades ago. Nuclear power learned it the hard way. AI companies are still figuring it out. But culture isn't a moat. It's not even a strategy, really. It's an emergent property of the people you hire, the norms you set, and the incentives you create. And all three of those things are under pressure at Anthropic. The people are changing -- the company is hiring aggressively, bringing in employees from Google, Meta, and traditional enterprise software companies who may not share the founding team's almost religious commitment to open debate. The norms are being tested by the pace of competition -- when OpenAI ships a new model, the pressure to respond quickly can override the impulse to debate thoroughly. And the incentives are shifting -- as Anthropic's valuation has soared, the financial stakes for employees have grown, creating a natural reluctance to rock the boat that no amount of cultural programming can fully overcome. None of this means Anthropic's culture is failing. By most accounts, it remains remarkably open for a company of its size and ambition. Employees still argue with the CEO on Slack. The CEO still argues back. Memos still circulate. Dissent still gets aired. But the window during which this kind of culture is possible -- when a company is large enough to matter but small enough to maintain genuine openness -- is finite. Anthropic is somewhere in the middle of that window right now. What happens next will say a lot not just about one company, but about whether the AI industry can govern itself before governments do it for them. The debates on Anthropic's Slack channels, in other words, aren't just about Anthropic. They're a proxy for a much larger argument about power, accountability, and who gets to decide how the most consequential technology of the century is built and deployed. For now, at least, the arguing continues.
Claude Code's Memory Crisis: How a Simple Config File Is Breaking Anthropic's AI Developer Tool
A seemingly minor configuration feature in Anthropic's Claude Code -- the company's AI-powered command-line coding assistant -- has become a persistent source of frustration for developers who rely on it daily. The problem: the tool's memory system, built on a file called , doesn't work the way users expect it to. And in some cases, it doesn't work at all. The issue, cataloged as #42796 on GitHub, was opened on July 10, 2025, by a user named seanoliver. The complaint is straightforward. When developers tell Claude Code to "remember" something -- a coding preference, a project convention, a workflow instruction -- the tool is supposed to write that information into a file so it persists across sessions. Instead, users report that Claude Code frequently ignores these stored memories, overwrites them without warning, or fails to read them altogether when starting a new conversation. "Claude regularly ignores instructions in CLAUDE.md," seanoliver wrote. "It also frequently overwrites or removes existing memories when adding new ones." The response from other developers was immediate. Within hours, dozens of users chimed in with similar experiences, suggesting the problem isn't isolated but systemic. Some described spending significant time carefully curating their files only to find Claude Code disregarding the contents entirely during subsequent sessions. Others reported that the AI would acknowledge the file existed but then proceed to violate every instruction contained within it. This matters more than it might appear at first glance. Claude Code isn't a toy. It's Anthropic's bid to compete directly with tools like GitHub Copilot, Cursor, and Google's Gemini Code Assist in the rapidly growing market for AI-assisted software development. The tool operates in the terminal, reading and writing code, running commands, and managing entire development workflows with minimal human intervention. For it to function effectively in real projects, it needs persistent context -- the ability to remember how a particular codebase is structured, what conventions the team follows, which libraries are preferred. That's exactly what is supposed to provide. Think of it as a project-specific instruction manual that the AI reads before doing anything. Anthropic's own documentation describes it as the primary mechanism for customizing Claude Code's behavior on a per-project basis. Developers can place the file at the root of their repository, and Claude Code should treat its contents as standing orders. Should. The gap between design intent and actual behavior is what's driving the frustration. One commenter on the GitHub issue described a scenario where they had explicitly instructed Claude Code, via , to always use TypeScript strict mode in a particular project. Claude Code ignored the instruction repeatedly, generating JavaScript files instead. When confronted, the AI would apologize, acknowledge the instruction, and then immediately violate it again on the next task. The pattern -- acknowledge, apologize, repeat -- is a familiar one to anyone who has worked extensively with large language models, but it's especially problematic in a tool marketed for professional development workflows. Another user reported that Claude Code would sometimes read the file at the start of a session but then "forget" its contents partway through a long interaction, reverting to default behaviors. This suggests the issue may be related to context window management -- the way the tool handles the finite amount of text it can process at any given time. As conversations grow longer and more complex, earlier context, including the contents of , may get pushed out or deprioritized. The overwriting problem is arguably worse. Multiple users described situations where they asked Claude Code to add a new memory, and it replaced the entire contents of with a single new entry, destroying carefully organized instructions that had been built up over weeks. No warning. No confirmation. Just gone. "I had about 30 lines of project-specific instructions," one developer wrote. "Asked Claude to remember one new thing. It replaced everything with a single bullet point." Anthropic has not publicly responded to the specific GitHub issue as of this writing, though the company's engineers have been active in other Claude Code issue threads. The tool is still technically in beta, which Anthropic has used as a general disclaimer for known limitations. But beta or not, developers are building real workflows around it, and broken memory fundamentally undermines trust in the tool. The timing is notable. Anthropic has been aggressively pushing Claude Code as a differentiator in its competition with OpenAI and Google. The company released Claude Code publicly in February 2025 and has since rolled out features including GitHub integration, multi-file editing, and agentic task execution -- the ability for Claude Code to autonomously plan and carry out complex development tasks across multiple files and commands. Memory persistence is foundational to all of these capabilities. An agent that can't remember its instructions is an agent that can't be trusted with autonomy. The broader AI coding tool market is watching. GitHub Copilot, powered by OpenAI's models, has its own approach to persistent context through custom instructions and repository-level indexing. Cursor, the AI-first code editor that has attracted significant developer enthusiasm, uses files for similar purposes and has generally received better marks for adherence to stored instructions, though it has its own issues. Google's Gemini Code Assist takes yet another approach, integrating with Google's broader cloud infrastructure for context management. What distinguishes the Claude Code memory problem from typical AI inconsistency is the explicit contract it creates. When a tool provides a specific mechanism for persistent instructions -- a named file, a documented format, a clear promise that "Claude will read this" -- and then fails to honor that mechanism, it breaks a fundamentally different kind of trust than when an AI simply generates an imperfect response. Developers aren't complaining that Claude Code writes buggy code sometimes. They're complaining that it ignores its own instruction manual. Some users on the GitHub thread have proposed workarounds. One popular approach is to include redundant instructions both in and in the initial prompt of every session. Others have started using version control specifically for their files, committing them to git so that overwrites can be quickly reverted. A few have written wrapper scripts that automatically re-inject contents at regular intervals during long sessions. These are clever hacks. They're also exactly the kind of thing that shouldn't be necessary. The issue touches on a deeper tension in AI tool development: the gap between what these systems can do in demos and what they can do reliably in production. Claude Code is remarkably capable in many respects. It can understand complex codebases, generate sophisticated implementations, and reason about architectural decisions in ways that would have seemed impossible two years ago. But capability without reliability is a hard sell to professional developers who need their tools to behave predictably. Anthropic's challenge is clear. The company needs to fix the memory system -- not just patch it, but make it deterministic enough that developers can trust it implicitly. That likely means changes to how contents are weighted in the model's context, how the file is protected from accidental overwrites, and how memory persistence is maintained across long sessions. It may also mean rethinking the architecture entirely, moving from a simple file-based approach to something more structured and harder for the model to accidentally ignore or destroy. Until then, developers are left in an uncomfortable position. They have a powerful AI coding assistant that can't reliably remember what they told it yesterday. For a tool that aspires to autonomous agency -- one that Anthropic wants developers to trust with significant, unsupervised coding tasks -- that's not a minor bug. It's a credibility problem. And credibility, once lost with developers, is extraordinarily difficult to win back.
Broadcom announces long-term chip supply agreements with Google and Anthropic By Investing.com
Broadcom Inc. (NASDAQ:AVGO) announced Monday that it has entered into a long-term agreement with Google LLC to develop and supply custom Tensor Processing Units (TPUs) for future generations of Google's TPUs. The agreement also includes a supply assurance arrangement for Broadcom to provide networking and other components for use in Google's next-generation artificial intelligence racks through up to 2031. In a separate development, Broadcom, Google, and Anthropic PBC have expanded their current strategic collaboration. Beginning in 2027, Anthropic will access approximately 3.5 gigawatts of next-generation TPU-based AI compute capacity through Broadcom, as part of Anthropic's commitment to multiple gigawatts of AI capacity. The use of this expanded AI compute capacity by Anthropic is dependent on its continued commercial success, according to the filing. The parties are also in discussions with certain operational and financial partners in connection with this deployment. These announcements were disclosed in a press release statement based on Broadcom's filing with the Securities and Exchange Commission. In other recent news, Broadcom Inc. has announced a significant transition in its executive team. Amie Thuener will take over as Chief Financial Officer, succeeding Kirsten M. Spears, who is set to retire in June 2026 after 12 years with the company. Thuener, previously with Alphabet, will officially begin her role on June 12, 2026. In another development, Broadcom and Carahsoft Technology Corp. secured a five-year, $970 million contract with the Defense Information Systems Agency to provide cloud software solutions to various Department of War agencies. This agreement aims to consolidate software contracts and offer uniform pricing for private cloud infrastructure and security solutions. Meanwhile, Erste Group has downgraded Broadcom's stock rating from buy to hold, citing concerns about the company's software segment growth prospects. Despite this, Broadcom projects second-quarter sales to reach approximately $22.0 billion, marking a 47% increase from the previous year. The company also anticipates AI semiconductor sales to hit $10.7 billion for the quarter. Additionally, Seaport's Chief Equity Strategist, Jonathan Golub, suggested that Broadcom, alongside other semiconductor companies, presents upside opportunities as software stocks are expected to rebound. This article was generated with the support of AI and reviewed by an editor. For more information see our T&C.
Broadcom Confirms Deal to Ship Google TPU Chips to Anthropic
Broadcom Inc. confirmed plans to deliver chips to AI startup Anthropic PBC using Google's tensor processing units, or TPUs, offering an alternative to technology developed by Nvidia Corp. Broadcom and Alphabet Inc.'s Google have entered a long-term agreement to provide the chips and a supply assurance pact that runs through 2031, according to a filing Monday. The three companies also are expanding a strategic collaboration that will let Anthropic access about 3.5 gigawatts' worth of computing power. That will begin in 2027. "The consumption of such expanded AI compute capacity by Anthropic is dependent on Anthropic's continued commercial success. In connection with this deployment, the parties are in discussions with certain operational and financial partners," Broadcom said in the filing. Broadcom shares climbed 3% in late trading after the filing was announced. The company's chief executive officer, Hock Tan, previously discussed the collaboration during an earnings call last month. He also said Broadcom expects its AI chip sales to top $100 billion next year, making it a bigger competitor to Nvidia. Google's TPUs were originally designed to speed up its ubiquitous search engine, but have become useful at creating and running AI software. Broadcom takes Google's specifications and creates fully-formed designs that can then be sent for manufacturing. Satellite startup Impulse Space is working with Anduril Industries Inc to develop space-based interceptor technology for President Donald Trump's planned Golden Dome missile defense shield, according to people familiar with the matter. The companies have been selected by the Pentagon to develop prototypes of space-based interceptors intended to track and destroy missiles from orbit, said the people, who asked for anonymity to discuss confidential matters. Impulse Space would work as a subcontractor to Anduril on the technology, which doesn't yet exist. Space-based interceptors are a key but unproven component of the Golden Dome effort, which seeks to protect the US and possibly other countries with layers of defense systems from the ground to space. Contracts and their recipients are one of the few indicators of how the Pentagon is progressing with Golden Dome, which has been shrouded in secrecy since Trump signed an executive order establishing the project more than a year ago. In November, the US Space Force said it awarded multiple contracts of less than $9 million each to unidentified companies to develop prototypes for the technology. While the contracts are relatively small, recipients will be able to compete for larger production contracts after demonstrating their technology, one of the people said. Led by Tom Mueller, a founding member and employee No. 1 at Elon Musk's SpaceX, Impulse Space develops spacecraft that can haul satellites across different orbits, known as "space tugs" in the industry. The California-based company, founded in 2021, has signed contracts with other government customers such as the National Reconnaissance Office, Space Systems Command and NASA. The joint contract work between Anduril and Impulse Space signals a deepening partnership between the two companies, which have teamed up to develop highly agile spacecraft for national security missions. Anduril's involvement in developing space-based interceptors was first reported by Reuters. Impulse Space's involvement hasn't been previously reported. Impulse Space, Anduril and the Pentagon didn't immediately respond to requests for comment. President Trump's Golden Dome is expected to cost $185 billion dollars and demonstrate operational capability by 2028, according to remarks made in March by General Michael Guetlein, the military official leading the effort. Critics say the budget constraints and timeline make the project overly ambitious, particularly due to the development of a space-based interceptor network. Space-based interceptors are considered the most technologically complex and probably most expensive aspect of Golden Dome since it would require the US manufacturing, building and launching hundreds or thousands of armed spacecraft into orbit, on top of research and development costs.
Broadcom Stock Climbs On New Deals With Google, Anthropic - Broadcom (NASDAQ:AVGO)
AVGO stock is up. See the chart and price action here. TPUs for Google, Anthropic In an SEC filing, Broadcom revealed an expanded long-term partnership with Google to power the tech giant's future AI infrastructure. Under the new long-term agreement, Broadcom will design and supply custom tensor processing units (TPUs) for Google's upcoming generations of AI accelerators, as well as provide networking and related components for Google's next-generation AI data racks through 2031. In a separate update, Broadcom, Google and Anthropic have also deepened their strategic collaboration. Beginning in 2027, Anthropic will secure access -- via Broadcom -- to roughly 3.5 gigawatts of next-generation TPU-based AI compute capacity, forming part of Anthropic's broader multi-gigawatt expansion plan. The agreement, which hinges on Anthropic's commercial growth trajectory, reflects rising demand for Broadcom's AI chips and infrastructure solutions. The companies said they are in talks with additional operational and financial partners to support the scaled deployment. AVGO Price Action: Broadcom shares were up 2.57% at $322.50 in Monday's extended trading, after closing slightly down at 0..38% in the regular trading session, according to Benzinga Pro data. Photo: Piotr Swat / Shutterstock Market News and Data brought to you by Benzinga APIs To add Benzinga News as your preferred source on Google, click here.
