Daily News Coverage: Samsung-TSMC HBM Deal, Anthropic Public Fund, Nvidia's $500B AI Infra Partnership
Daily News Coverage · October 19, 2026
The Big Headline: Anthropic Public Benefit Fund, Samsung-TSMC HBM Deal, Intel Fab Explosion, n0n Maritime Extortion Deadline
Good morning, folks! Anthropic commits 5% of its shares to a new public benefit fund as it circulates a draft AI safety bill that would let Washington block models only if they risk $1 billion in damage or 50 deaths. Samsung will use TSMC's 3nm process for Nvidia's custom HBM base dies in a historic reversal of its one-stop-shop model. An explosion at Intel's Kiryat Gat fab in Israel kills two workers and threatens CPU supply. The n0n extortion group's deadline to leak data from 38 maritime facilities — including an Indian port terminal — expires today. CISA sets a federal remediation deadline for BeyondTrust and Fortinet edge-device flaws. Proofpoint agrees to acquire Varonis for $6.3 billion. Workday moves to buy C1.ai. And AI stocks tumble as credit fears deepen. Here's today's briefing from The CODEW's The Newsroom.
Anthropic to Commit 5% of Shares to New Public Benefit Fund
The fund will support public-interest giving, including efforts to mitigate AI's impact on the job market, and the stake could increase as Anthropic's valuation grows.
What happened: Anthropic said on Sunday, October 18, that it will commit 5% of its shares to a new public benefit fund, subject to stockholder approval. The stake could potentially increase as the artificial intelligence company's valuation grows. The fund is designed to support public-interest giving, with uses including mitigating the impact of AI on the job market. The announcement comes as Anthropic continues to position itself as the safety-focused alternative to OpenAI and prepares for a possible public listing.
Key numbers/companies: Anthropic · 5% of shares · public benefit fund · stockholder approval · AI job displacement.
Why it matters: The commitment is an unusual governance move for a frontier AI lab, effectively tying a portion of its equity to public-interest outcomes. It creates a structural mechanism for Anthropic to fund safety and workforce transition efforts regardless of short-term commercial pressure. For investors, it adds a layer of complexity to valuation and control.
What's next: Whether stockholders approve the commitment and how the fund is governed.
Sources: Reuters; MarketScreener; TBS News; 财联社.
Samsung to Use TSMC's 3nm Process for Nvidia's Custom HBM Base Dies
The move breaks Samsung's years-long one-stop-shop model and marks its first collaboration with TSMC at the HBM manufacturing stage, driven by TSMC's yield exceeding 90%.
What happened: Samsung Electronics will use TSMC's 3-nanometre process to produce base dies for Nvidia's next-generation custom high-bandwidth memory (HBM) chips, the Seoul Economic Daily reported on Sunday. The move would mark the first time Samsung has worked with the Taiwanese contract chipmaker at the HBM manufacturing stage. Nvidia has asked Samsung, SK hynix, and Micron to use TSMC's process for NVHBM base dies and sixth-generation DRAM technology for core dies. NVHBM, unveiled in August, integrates Nvidia's custom memory controller into the base die and is aimed at standardizing implementation across multiple memory suppliers. Samsung accepted Nvidia's requirements after an internal review but plans to retain its in-house production model for standard HBM and custom products for other customers. It is also assessing whether its own foundry could produce NVHBM base dies if demand exceeds TSMC's capacity.
Key numbers/companies: Samsung · TSMC · Nvidia · NVHBM · 3nm · base die · SK hynix · Micron · 90% yield.
Why it matters: The reversal of Samsung's vertically integrated model is a significant strategic shift. TSMC's 3nm process and 90%+ yield are now critical to Nvidia's HBM roadmap, giving TSMC leverage in the memory supply chain it didn't have before. For Samsung, the deal ensures it remains an Nvidia supplier even if it cedes part of the manufacturing process to a competitor.
What's next: Whether Samsung's own foundry qualifies to produce NVHBM base dies if TSMC capacity becomes constrained.
Sources: Business Standard; Reuters; Seoul Economic Daily; UDN.
Explosion at Intel's Kiryat Gat Fab Kills Two Workers, Threatens CPU Supply
Fab 28 is a key production site for Intel CPUs, and the incident raises questions about supply for Taiwan's top four PC brands.
What happened: A gas explosion occurred at Intel's Fab 28 in Kiryat Gat, southern Israel, killing two workers and seriously injuring one. The facility is a major production base for Intel CPUs. The incident has raised concerns about potential supply disruptions for Taiwan's four major PC brands, which rely on Intel processors. Intel has not yet disclosed the cause of the explosion or the extent of damage to production lines.
Key numbers/companies: Intel · Fab 28 · Kiryat Gat, Israel · 2 dead · 1 injured · CPU supply.
Why it matters: Intel is already under pressure from AMD and Arm-based competitors, and a disruption at a key CPU fab could worsen its supply position. The incident also raises safety questions about Intel's global manufacturing operations.
What's next: Intel's assessment of damage and whether alternative fabs can absorb production.
Sources: SETN; Techmeme.
n0n Extortion Group's Deadline to Leak Data from 38 Maritime Facilities Expires Today
An Indian port terminal is among the targets, with ransom demands ranging from $2.4 million to $14 million in Bitcoin.
What happened: The n0n cyber extortion group has threatened to publish data allegedly stolen from 38 maritime facilities across 18 countries unless multi-million-dollar ransoms are paid in Bitcoin by October 19. The group, which emerged in September 2026, is financially motivated and monetizes stolen data through public leak postings rather than traditional ransomware deployment. SOCRadar reported that the facilities were added to n0n's dark web leak site in a single batch on October 12, marking the group's shift to maritime targets. The list includes an Indian facility identified as NCT Kandla, as well as Ocean Network Express, CULines, Santos Brasil, and NEOM Ports. Ransom demands range from $2.4 million to $14 million. SOCRadar noted that the relatively small, application-specific datasets claimed from several victims suggest a common application or shared vulnerability may have been exploited rather than 38 separate attacks. None of the listed organizations had confirmed breaches as of October 15.
Key numbers/companies: n0n · 38 maritime facilities · 18 countries · NCT Kandla · Ocean Network Express · CULines · Santos Brasil · NEOM Ports · $2.4M-$14M ransom · SOCRadar.
Why it matters: Maritime infrastructure is a high-value target because port operations depend on interconnected systems and the consequences of disruption are severe. The n0n campaign represents a shift from traditional ransomware to pure data extortion, which is harder for defenders to counter because there is no encryption key to recover.
What's next: Whether n0n follows through on its leak threat and how maritime operators respond.
Sources: The Week; SOCRadar; NewMoney.
CISA Sets October 19 Federal Remediation Deadline for BeyondTrust and Fortinet Edge Flaws
Both vulnerabilities are flagged for forensic triage, meaning agencies must check for compromise, not just patch.
What happened: CISA added two exploited edge-device flaws to its Known Exploited Vulnerabilities (KEV) catalog on October 16, both with a federal remediation due date of October 19 and both flagged for forensic triage. CVE-2026-102764 affects BeyondTrust Remote Support (BT26-05) and is an authentication bypass by spoofing that lets a remote attacker perform file operations on the appliance; Mandiant tied zero-day use to the ARMORY CASTLE threat cluster. CVE-2025-61624 affects FortiOS, FortiPAM, FortiProxy, and FortiSwitchManager and is a CLI path traversal vulnerability that lets a read-write admin write or delete arbitrary files; Fortinet marks it as Known Exploited, and Unit 42 linked it to the PEERSTITCH FortiGate implant. The forensic triage flag means agencies should check for signs of compromise, not just apply patches. Fixed versions are available for both products.
Key numbers/companies: CISA · KEV catalog · BeyondTrust · CVE-2026-102764 · Fortinet · CVE-2025-61624 · ARMORY CASTLE · PEERSTITCH · October 19 deadline.
Why it matters: Both vulnerabilities affect edge devices that sit at the perimeter of enterprise networks, making them high-value targets for attackers seeking initial access. The forensic triage requirement reflects the likelihood that some systems have already been compromised.
What's next: Whether agencies meet the deadline and whether additional victims are identified.
Sources: Dark Web Intelligence; CISA; Mandiant; Unit 42.
Anthropic Circulates Draft AI Safety Bill That Would Let Washington Block Models Only for Catastrophic Risks
The proposal would require a judge's approval and sets thresholds of $1 billion in damage or 50 deaths, while overriding some state laws including California's.
What happened: Anthropic confirmed it is circulating a draft bill among US lawmakers aimed at regulating artificial intelligence and addressing the absence of federal legislation on AI risks. According to the draft first reported by Politico, Washington could block a model's release or use only if there is an imminent risk of killing or seriously injuring more than 50 people, or causing more than $1 billion in damages. That intervention would require a judge's approval. The proposal would grant the federal government the ability to block or deter dangerous deployments of the most powerful models, require frontier developers to test their most powerful models, be transparent about findings, submit to independent evaluation, and maintain a robust security program. The draft would also override certain state laws, including laws in California. Anthropic co-founder Jack Clark said at the IMF-World Bank Annual Meetings in Bangkok that children's toys are regulated more than AI. The draft is not yet law, and it remains unclear whether the Trump administration will adopt it.
Key numbers/companies: Anthropic · draft bill · $1 billion damage threshold · 50 deaths · Politico · Jack Clark · California SB 53 · Commerce Department.
Why it matters: The draft represents the most detailed attempt by a frontier AI lab to shape federal AI legislation. By setting high thresholds for government intervention, Anthropic is positioning itself as a responsible actor while limiting regulatory risk. The proposal's override of state laws is controversial and could face opposition from California and other states that have moved ahead with their own AI rules.
What's next: Whether any lawmaker introduces the bill and how the Trump administration responds.
Sources: India Today; RTHK; Review of AI Law; Donga Biz; Hani.co.kr.
Proofpoint to Acquire Varonis for $6.3 Billion to Build Unified Data Security and AI Governance Platform
The combined company aims to serve more than 16,000 enterprises with data discovery, identity and AI access controls, and data loss prevention.
What happened: Proofpoint has agreed to acquire data security vendor Varonis Systems in an all-cash transaction valued at approximately $6.3 billion. The combined company aims to serve more than 16,000 enterprises with a unified platform spanning data discovery, identity and AI access controls, and data loss prevention. The deal reflects growing demand for integrated security platforms that can govern both human and AI-agent access to sensitive data.
Key numbers/companies: Proofpoint · Varonis · $6.3 billion · all-cash · 16,000+ enterprises · data discovery · AI access controls · DLP.
Why it matters: The acquisition is one of the largest cybersecurity deals of the year and signals consolidation around data-centric security. As AI agents proliferate, enterprises need tools that can govern what data AI systems can access — Varonis's data classification and identity controls fill a critical gap in Proofpoint's portfolio.
What's next: Regulatory approval and integration timelines.
Sources: TechPartner News; FintechGate.
Workday Agrees to Acquire C1.ai, Formerly ConductorOne, to Manage Identities for People and AI Agents
C1.ai has raised $111 million and develops enterprise software that manages identities for both human workers and autonomous AI agents.
What happened: Workday has agreed to acquire C1.ai, formerly known as ConductorOne, according to Portland Business Journal. C1.ai develops enterprise software that manages identities for both people and AI agents. The company has raised $111 million. The acquisition would strengthen Workday's ability to govern access for AI agents operating within enterprise HR, finance, and IT workflows.
Key numbers/companies: Workday · C1.ai · ConductorOne · $111 million raised · identity management · AI agents.
Why it matters: The deal reflects the growing recognition that AI agents require the same identity governance as human employees. Workday already provides HR and finance systems of record; adding C1.ai's identity controls would allow it to govern who — or what — can access sensitive enterprise data and actions.
What's next: Whether the deal closes and how C1.ai's technology is integrated into Workday's platform.
Sources: Tech and Business; Portland Business Journal.
Tether and HIVEN Back AIsa's Seed Extension as AI Agent Payment Startup Tops $9 Million
AIsa's platform lets AI agents access online services and make payments within preset limits, and it claims 250,000+ agents and 100 million+ transactions.
What happened: Tether and HIVEN have invested in San Francisco AI startup AIsa, completing a seed extension round that brings the company's cumulative funding to more than $9 million. AIsa develops technology that allows AI agents to access online services and complete payments within preset limits. The platform says it has integrated more than 5,000 APIs, software tools, and models, and claims to run over 250,000 agents and process more than 100 million transactions. Existing investors include Alibaba, Tribe Capital, Sumitomo Corporation, Draper Associates, and Saison Capital. Tether invested through its El Salvador-based Tether Capital as part of its AI payments and programmable money strategy, but did not disclose the specific amount, valuation, or ownership stake. AIsa said settlement can use fiat or stablecoins, and Tether did not require all transactions to be cleared in USDT.
Key numbers/companies: AIsa · Tether · HIVEN · $9 million+ · $2.5 million extension · 5,000+ APIs · 250,000+ agents · 100 million+ transactions · Alibaba · Tribe Capital · Tether Capital.
Why it matters: The investment reflects growing interest in infrastructure that allows AI agents to transact autonomously. If agents are to become economic actors, they need payment rails, identity, and spending controls — exactly the layer AIsa is building. Tether's involvement signals that stablecoin issuers see AI-agent payments as a growth market.
What's next: Whether AIsa's agent payment infrastructure gains enterprise adoption.
Sources: Mitrade; 金色财经; Gate; 377BTC; ChainCatcher.
Anthropic Offers Blackstone Over 10% Yield in $60 Billion "Project North Star" Chip Financing
The debt deal would help Anthropic procure about 1.5 gigawatts of Google-developed TPU chips and could become the largest chip financing transaction to date.
What happened: Anthropic is offering Blackstone a yield of more than 10% to participate in a $60 billion debt financing package codenamed "Project North Star," according to the Financial Times. The financing would help Anthropic procure approximately 1.5 gigawatts of custom semiconductors developed by Google, known as TPUs. The deal is expected to be the largest chip financing transaction to date. Anthropic's willingness to pay double-digit yields reflects the intense capital requirements of the AI compute race and the premium investors are demanding for AI-related credit risk.
Key numbers/companies: Anthropic · Blackstone · $60 billion · Project North Star · 1.5 GW · Google TPU · 10%+ yield.
Why it matters: The deal highlights the scale of capital required for frontier AI compute and the growing role of private credit in financing the AI buildout. A 10%+ yield on a $60 billion package is expensive capital, but Anthropic has little choice if it wants to secure dedicated TPU capacity outside of Google's own cloud.
What's next: Whether the financing closes and how Anthropic deploys the TPU capacity.
Sources: Financial Times; Investing.com; BitMart; East Money.
Nvidia Partners with Six Financial Firms to Mobilize $500 Billion for AI Infrastructure
CreditSights estimates the five largest hyperscalers will spend about $602 billion in 2026, while Bernstein warns only 35% of planned US data center projects are credible.
What happened: Nvidia announced partnerships with six major financial firms to create platforms that could mobilize more than $500 billion in third-party capital for AI infrastructure. The announcement comes as Amazon, Alphabet, Microsoft, and Meta are expected to spend more than $720 billion in capital expenditure this year, most of it on data centers and AI chips. CreditSights estimates the five largest hyperscalers will spend about $602 billion in 2026. Meanwhile, Bernstein analysts estimate that only about 35% of planned US data center projects are credible, as project delays, community opposition, and waning confidence weigh on the $10 trillion AI infrastructure boom.
Key numbers/companies: Nvidia · $500 billion · six financial firms · $720 billion capex · $602 billion hyperscaler spend · Bernstein · 35% credible projects.
Why it matters: Nvidia's move to mobilize third-party capital is a recognition that its own balance sheet cannot fund the entire AI infrastructure buildout. The gap between announced projects and credible projects is a warning sign: the AI infrastructure boom is increasingly dependent on debt and financial engineering, which raises systemic risk if returns disappoint.
What's next: Whether the financing platforms attract institutional capital and whether data center projects proceed on schedule.
Sources: Diario Bitcoin; Fortune Korea; Dünya; Hong Kong Economic Journal.
Boards Urged to Tighten AI, Identity, and Software Pipeline Oversight
SailPoint warns that 79% of organizations use AI agents but only 2% have identity tools to govern them.
What happened: Cybersecurity vendors and advisers are using Cybersecurity Awareness Month to urge boards and executives to tighten oversight of AI, software pipelines, and identity controls. Matthew Lowe, Regional Director at Anomali, said AI is moving from advising humans to acting on their behalf, raising the question of what organizations allow machines to decide. Eric Kong, Managing Director at SailPoint, said 79% of organizations globally use AI agents but only 2% have identity tools to govern them. He recommended discovering every agent, enforcing Zero Standing Privilege, granting just-in-time access, and revoking it after each task. JFrog's Itzik Swissa said AI-assisted development is producing code faster than security teams can govern it, and attackers are chaining low- and medium-severity flaws into serious exploits. JFrog's 2026 survey found 47% of Australian organizations use automated controls to block unapproved IDE extensions and MCP servers, but only 38% have adopted secrets detection.
Key numbers/companies: Anomali · SailPoint · JFrog · 79% use AI agents · 2% have identity tools · 47% block unapproved extensions · 38% secrets detection.
Why it matters: The data reveals a massive governance gap: organizations are deploying AI agents far faster than they are implementing controls to manage them. The fact that only 2% have identity tools for agents is a stark indicator that most enterprises are operating blind.
What's next: Whether boards take action and how quickly identity governance for AI agents is adopted.
Sources: IT Brief India.
Why It Matters
Anthropic's public benefit fund is a new model for AI governance. Committing 5% of shares to a fund that supports public-interest giving — including mitigating AI's impact on jobs — is not charity. It is a structural mechanism that ties a portion of the company's equity to outcomes beyond shareholder returns. If other labs follow, it could reshape how AI companies are governed.
Samsung's decision to use TSMC's 3nm process is a strategic reversal. Samsung has spent years promoting its one-stop-shop model, where it provides memory, foundry, and packaging services. Now it is ceding part of the HBM manufacturing process to its biggest rival because Nvidia demands TSMC's 90%+ yield. This is a humbling admission and a sign that Nvidia's standards are reshaping the supply chain.
Anthropic's draft AI bill is a calculated bet on regulatory influence. By proposing high thresholds for government intervention — $1 billion in damage or 50 deaths — Anthropic is positioning itself as a responsible actor while limiting its own regulatory risk. But the proposal to override state laws will draw opposition. The question is whether Congress will act before the midterms or whether the bill becomes a template for future legislation.
The AI credit selloff is intensifying. Nvidia is partnering with six financial firms to mobilize $500 billion for infrastructure, Anthropic is offering 10%+ yields to Blackstone, and Bernstein says only 35% of planned US data center projects are credible. The AI boom is increasingly debt-funded, and the market is starting to price in risk. The next few quarters will test whether the returns justify the leverage.
Enterprise identity governance for AI agents is the most underinvested area in cybersecurity. SailPoint's finding that 79% of organizations use AI agents but only 2% have identity tools to govern them is alarming. Proofpoint's $6.3 billion acquisition of Varonis and Workday's move for C1.ai show that vendors recognize the gap. But most enterprises are still operating without basic controls.
The CODEW Take
Anthropic's 5% public benefit commitment is the most interesting governance experiment in AI. The company is effectively saying: we will set aside a slice of our equity to fund public goods, including mitigating the job losses that our technology may cause. That is not a PR move — it is a structural commitment that survives leadership changes and commercial pressure. The question is whether 5% is enough and whether the fund's governance is independent.
Samsung's TSMC deal is the most important supply chain story of the week. Nvidia's demand for TSMC's 3nm process and 90%+ yield forced Samsung to break its one-stop-shop model. That is a stunning admission: Samsung's own foundry cannot meet Nvidia's standards for this product. The deal gives TSMC even more leverage over the AI supply chain and raises questions about Samsung's long-term competitiveness in advanced packaging.
Anthropic's draft bill is smart politics but risky policy. The company is proposing a federal standard that preempts state laws, including California's. That will not sit well with state regulators who have moved aggressively on AI. But the alternative — a patchwork of state rules — is worse for AI companies. Anthropic is betting that Congress will prefer a single federal framework with high thresholds over a fragmented landscape.
The Intel fab explosion is a reminder that semiconductor supply chains are fragile. A single gas explosion at a key CPU fab can disrupt supply for months. Intel is already struggling to compete with AMD and Arm; a production setback is the last thing it needs. The incident also raises safety concerns at Intel's global fab network.
The identity governance gap for AI agents is the cybersecurity story of the year. Only 2% of organizations have tools to govern AI agents, according to SailPoint. That means 98% are deploying autonomous systems with access to sensitive data and no way to audit or revoke that access. Proofpoint and Workday are making acquisitions to fill the gap, but most enterprises are not ready. The next major AI-agent breach will likely be traced to a permissions problem, not a model failure.
What to Watch
- Anthropic stockholder vote: Whether the 5% public benefit fund commitment is approved.
- Samsung-TSMC-Nvidia production: Whether Samsung begins producing NVHBM base dies on TSMC's 3nm process on schedule.
- Intel Fab 28 investigation: The cause of the explosion and the impact on CPU supply.
- n0n leak deadline: Whether the group follows through on its threat to publish data from 38 maritime facilities.
- CISA KEV remediation: Whether federal agencies meet the October 19 deadline and whether additional compromises are found.
- Anthropic draft bill: Whether any lawmaker introduces the proposal and how the Trump administration responds.
- Proofpoint-Varonis integration: How the combined platform governs AI-agent access to data.
- Workday-C1.ai: How identity governance for AI agents is integrated into Workday's suite.
- Anthropic Project North Star: Whether the $60 billion debt financing closes at the offered yield.
- Nvidia GTC Berlin (Oct 20-22): Jensen Huang's keynote and product announcements.
Editorial Note
The CODEW Daily News Coverage tracks the most important technology developments of the day, with a focus on AI, enterprise software, cloud computing, semiconductors, cybersecurity, startups, and digital infrastructure. This edition is a pure news aggregation and coverage page — what happened — while the Daily Tech Briefing provides the deeper strategic interpretation.
Coverage is based on company announcements, public disclosures, industry reporting, and other publicly available information. Reported figures and sourced-but-unconfirmed details are noted as such.
Reviewed by Erwin Castro
on
Monday, October 19, 2026
Rating:
No comments: