security

According to Bloomberg, citing informed sources, SoftBank is considering acquiring Allied Universal, the largest security services company in the United States, with a potential equity valuation of up to $18 billion. The company also has about $12 billion in publicly traded loans and bonds, and including related debts, the total transaction value could approach $30 billion.Masayoshi Son is exploring the application of AI robots across various industries. The report indicates that SoftBank hopes to have robots undertake some security tasks for Allied in the future, but the deployment of humanoid security robots may take longer. Relevant discussions are still in the early stages, and matters including pricing have not been finalized.

The Indian Computer Emergency Response Team (CERT-In) released an important announcement this week titled "AI Agents as Attack Vectors," indicating that AI agents operated by unnamed model providers exceeded their granted task scope during testing, reading confidential business records in the online systems of Indian agencies, and no affected agencies discovered this behavior on their own.CERT-In stated that it received voluntary disclosures from several AI model providers, which indicated that these models accessed the online systems of Indian agencies while being researched and evaluated on the internet. The announcement did not disclose the names of the providers, models, or affected agencies. The agents read business records containing personal data and confidential materials, utilizing credentials and connection settings exposed in application configurations, public code repositories, and overly permissive cloud service roles. In some cases, the same credentials were recovered multiple times and reused in different operations.Additionally, CERT-In observed that criminals exploited publicly available AI-driven offensive security frameworks to scan for vulnerabilities, implement exploits, and launch economically motivated attacks against Indian entities, reducing the time from initial access to final impact from several days to just hours; these tools were repeatedly repackaged and deployed, with related deployments appearing on servers in multiple countries.

On October 15, Anthropic announced that it has begun providing employee-like access to a team from the nonprofit organization METR for the purpose of evaluating its models and safety practices. The company stated that last month it committed to strengthening independent third-party evaluations and signed the White House "Superintelligence Agreement," which includes a commitment to collaborate with independent external auditors or evaluators. Over the past year, Anthropic has worked with METR to review several risk reports, and METR will publish its risk assessment results.Anthropic has equipped six METR employees with company hardware, office space, and access to internal tools and systems roughly equivalent to those of its own risk assessment team. With a few explicitly defined categories such as client data, attorney-client privileged information, and model architecture information, employees are encouraged to share information as they would with colleagues working on alignment risk. METR may publish analyses of risk levels, events, or practices, which Anthropic cannot veto, but can request the removal of pre-defined narrow category information; METR may publicly state the circumstances of redaction and indicate whether it believes the redaction is reasonable without disclosing the content that was redacted. This arrangement is limited to small teams working under enforceable agreements, and other METR employees are isolated from working with this team.Anthropic is also collaborating with SecureBio and Accenture on assessments and is in early discussions with other organizations, including Hugging Face. SecureBio is reviewing the chemical and biological weapons section of the August 2026 risk report, having previously completed and published a pilot review of the previous risk report.

Elon Musk stated that SpaceX may follow Anthropic and hire external safety assessors.

Polygon Labs official announcement, Bor v2.1.3 is officially released to the public, previously distributed in private version form to validators and major infrastructure providers. This update fixes two medium-severity denial-of-service vulnerabilities in BlockSTM V2 parallel block execution, both submitted by security researchers through the Immunefi bug bounty program.Unbounded keccak256 cache issue: Attackers can fill hundreds of MiB of cache during the speculative execution phase by constructing specific transactions, leading to node memory exhaustion and delayed block production. The fix is to set a capacity limit on the shared cache. Repeated code hash calculation issue in parallel execution: A single transaction can cause the parallel processor to take about 11 seconds, severely affecting node activity. The fix is to cache code hash results on a per-transaction basis.In addition, this update also includes several internal hardening measures: the transaction pool address filtering adds interception for EIP-7702 SetCode authorization, RPC batch execution pool concurrency limit fixes, consistency fixes for reading balances across self-destruct boundaries in parallel execution, and an automatic cleanup mechanism for pending transactions (default 3 hours). The affected versions are Bor v2.1.1 and v2.1.2. v2.1.3 is a non-hard fork, and nodes can upgrade individually without coordination. No exploitation records of the two vulnerabilities have been found on the mainnet and testnet Amoy.

According to The Defiant, Safe announced the launch of the Safenet Aegis security network on Ethereum, Gnosis Chain, and Arbitrum. This is the first production-ready version of Safenet that can generate fees, previously approved for funding and launch by SafeDAO through the SEP 56 proposal. The network separates transaction verification from transaction authorization, providing transaction security verification for smart accounts like Safe multi-signature wallets.Safenet consists of three types of roles: Sentinel, responsible for checking transactions and staking USDC bonds, can earn verification fees paid in USDC; Validator, who secures the proof layer by staking SAFE, can earn SAFE rewards; Delegator can delegate SAFE to Validators and participate in staking rewards. Safenet Guard is responsible for executing proof results on-chain, with relevant protocol parameters governed by SafeDAO.The current Sentinel pool includes OpenCover, IntentGuard, BlockSec, and Candide, while the Validator pool includes Gnosis, Greenfield, Core Contributors, and Safe Labs. Safenet will subsequently integrate transaction verification into Safe Wallet, displaying verification results during the transaction review process and supporting access for other wallets built on Safe smart accounts.

The Financial Services Commission (FSC) of South Korea stated that after investigating 16 companies (including cryptocurrency exchanges) involved in a previous security incident, it plans to implement stricter cybersecurity requirements for financial institutions. The agency noted that in an industry review covering 555 financial companies, 130 companies were found to have cybersecurity deficiencies, with approximately 100,000 pieces of customer information leaked, including names, contact details, and identification numbers.Affected companies include Streami, the operator of Gopax, and the cryptocurrency trading platform Prabang, which previously disclosed that a hacking incident on October 2 led to the leakage of some customers' personal information. The FSC indicated that it is still investigating whether these incidents involved the use of artificial intelligence or if they are related to the same group of attackers.According to the proposed reforms, larger financial institutions will undergo vulnerability assessments at least twice a year. The regulatory agency also plans to improve the information-sharing system and is considering using artificial intelligence for security monitoring of small businesses. Additionally, the FSC is considering allowing supplementary cybersecurity inspections outside of the regular assessment schedule in the event of an urgent threat to prevent similar incidents from occurring in the future.

Vitalik posted on the X platform to congratulate the @zksync Airbender team for achieving this milestone. He believes that the speed competition of zkEVMs has achieved a decisive victory, and the primary focus should shift entirely to security.Vitalik stated that this means there is a need to formally verify the equivalence of arithmetic and Ethereum specifications, the hard upper limit of the ratio of prover costs to total block Gas, and the security of the proof system itself.

On the 11th, Microsoft co-founder Bill Gates stated in a television interview that AI is the most dangerous technology ever invented by humanity, with risks exceeding those of cars, airplanes, and pharmaceuticals. He mentioned that the real threat at present is not AI going out of control, as that scenario is still years away from becoming a reality; rather, the more immediate threat is criminals using AI to carry out extortion, manipulate the stock market, and other crimes.Gates indicated that as criminals crack and bypass the security protections of the latest AI models, there will be a significant number of negative events in the cybersecurity field over the next year. He also mentioned concerns about biosecurity, stating that AI has significant implications for fields such as healthcare, but the potential risks still outweigh those of traditional industries that are already under strict regulation. Currently, there is no effective regulatory framework established globally, and within the next year, industry governance will have to move beyond mere self-regulation, while existing resource investments are far from sufficient.According to reports cited by the Financial Associated Press, Ken McCallum, the director of the UK's MI5, stated in the annual threat assessment that AI is exacerbating the cyber offense and defense game and may change the risk landscape of biological weapons. MI5 has announced the establishment of an AI threat assessment center. Additionally, reports indicate that executives from laboratories such as OpenAI and Anthropic have recently privately simulated disaster response plans, believing that a likely scenario involves cyberattacks impacting financial services, internet connectivity, or water and electricity supply.

The Synthetic Dollar Agreement's founder and CEO of Ethena, Guy Young, stated that users have noticed a large investor wallet transferring tokens in the past 48 hours. He confirmed that the related transfers are a security precaution related to recent concerns about hardware wallets and do not indicate that the tokens are being sold after being transferred.Guy Young made the above statement on X in response to the community's concerns about the on-chain movements of large wallets. He emphasized that the related operations are precautionary arrangements taken for security reasons and do not represent that investors are selling their held tokens.

NVIDIA has committed to investing $250 million over the next five years to promote research in agent safety and security in the UK. The company stated that it will provide $50 million in computing resources support for agent safety research in the UK each year.

Pluto Security, an enterprise artificial intelligence security startup, announced the completion of a total of $42 million in financing.This round of financing includes an $8.5 million seed round led by TLV Partners, with participation from the founder of cybersecurity company Armis, as well as a $34 million Series A round led by Jump Capital, with participation from Skinos and Ansa Capital. The new funds will be used for product research and development, expanding security research operations, and increasing market presence in the United States.

San Francisco-based cybersecurity company Reach Security announced on October 13, 2026, the completion of over $50 million in Series B financing, led by Top Tier Capital Partners, with participation from Ballistic Ventures, In-Q-Tel, Microsoft's venture fund M12, and existing investors. The funds will be used for product innovation, expanding integration partnerships, international expansion, and growth in the U.S. enterprise and federal markets.At the same time, Reach Everywhere was launched, integrating security control insights into exposure management, security operations centers, incident response, compliance platforms, and AI security environments through the Reach API, MCP servers, and technology partners. The company claims its AI reads the real-time configurations of existing security products in enterprises, determines whether exploitation paths are blocked, and points out the specific rules that dictate that outcome, allowing for immediate remediation when protection is bypassed.According to VulnCheck, less than 1.3% of vulnerabilities were exploited in the wild in the first half of 2026; citing CrowdStrike, AI-enabled adversary actions increased by 89%, with the shortest time for attackers to breach being as little as 27 seconds. The number of customers has doubled in the past 12 months, including Autodesk, Insurity, MPS, and Ensemble Health Partners. Independent analysis by TAG Infosphere states that organizations using Reach save over 160 hours weekly on control reviews, situational assessments, and incident investigations, with a return on investment of 156%.

According to SolanaFloor, a TOKEN2049 attendee has filed a report with the Singapore police regarding a crypto security side event called Don't Get Played that took place during TOKEN2049.The event was promoted by self-proclaimed organizers Apex Accelerator and Stone Venture. It is alleged that the organizers pressured and induced attendees on-site to install malware that would allow the organizers to remotely take full control of their computers.

Hardware wallet manufacturer Trezor stated that in light of recent supply chain security incidents, it has decided to reduce the number of distributors to more effectively review partners.Currently, there is no evidence that Trezor devices have been compromised. This adjustment is a precautionary measure taken out of caution.

The Finance Committee of the French National Assembly has passed an amendment that proposes requiring the declaration of self-custodied wallet crypto assets with a year-end value reaching or exceeding €100,000 (approximately $112,000). Those who fail to declare may face fines of up to €10,000 (approximately $11,000). This regulation has not yet become law and may still be adjusted in the future.Harry Halpin, CEO of privacy infrastructure company Nym, stated that personnel from the French tax authorities are suspected of selling taxpayer data to criminals, and the new declaration requirements could threaten the safety of holders. Rand Hindi, founder of the privacy protocol Zama, urged French crypto asset holders to leave France, and those who cannot should hire security for their families.As this proposal emerges, France has seen several incidents of crypto asset robberies, kidnappings, and "ransom attacks" in recent years. The committee had also previously passed measures such as considering the exchange of crypto assets for stablecoins as a taxable sale and expanding the scope of the crypto asset exit tax, but the budget revenue portion was subsequently rejected by a vote of 31 to 3, and the related amendments need to be resubmitted.

Web3 security platform Blockaid co-founder Ido Ben-Natan posted on the X platform that since the release of the first half report in August, the number of on-chain cybersecurity incidents has continued to reach historical highs month by month. Attackers, motivated by profit, find that attacking on-chain infrastructure offers the highest return for hackers, and this trend will become even more severe as the funds callable by AI agents increase.According to Fortune, Ben-Natan stated that the company is currently in its "busiest ever" state, one reason being the significant increase in hacker attacks. Additionally, Nansen CEO Alex Svanevik predicted that by 2028, trading cryptocurrency assets through AI agents rather than manual interface clicks will become the norm.

According to Bitcoin News monitoring, the Lightning Development Kit (LDK) has released a security update that fixes vulnerabilities that could allow malicious nodes to steal funds. The most serious vulnerability involves Lightning channel reconnections. According to Bitcoin Optech, malicious nodes may falsely claim that they did not receive previously confirmed channel updates.The vulnerable LDK software could subsequently sign conflicting commitment transactions without properly recording the related transactions, allowing attackers to regain Bitcoin after forwarding the payment is completed. Versions LDK 0.2.7 and 0.1.13 have fixed this vulnerability. As of the reviewed disclosure information, no confirmed losses caused by these specific vulnerabilities have been found.

a16z crypto invited Stanford University professor Dan Boneh, a16z crypto partner Eddy Lazzarin, and other experts to discuss the threats AI poses to cryptography. The experts pointed out that recent advancements in AI in the field of mathematics have not included cryptographic results and have not proven any specific cryptographic assumptions to be less secure.The experts stated that modern cryptography relies on the assumption that certain mathematical problems do not have efficient algorithms; if this assumption holds, no matter how powerful AI becomes, it cannot find algorithms that do not exist. This assumption is related to the unsolved P vs NP problem in computer science, and current AI advancements have not yet touched upon this issue.The experts believe that AI can significantly reduce the time required for cryptographic reviews. Hypotheses that previously took ten years to validate can now allow models to run for a month to attempt to break them. At the same time, the migration work driven by quantum threats has prepared for some of the risks posed by AI, and cryptographic agility can help systems switch methods when weaknesses are discovered. For developers, the most direct opportunity is to use AI to discover more vulnerabilities before software deployment.

According to Axios, officials from the Trump administration stated that after the artificial intelligence company Anthropic was accused of "fraudulently" using government systems, AI companies are now required to report and rectify security incidents. The leader of the White House Superintelligence Task Force stated in an exclusive statement to Axios that this reporting and remediation process is not optional but a critical national security obligation.The government stated that Anthropic reported unauthorized and fraudulent use of government and other systems discovered at the end of last month, and the related activities have been halted. State Department officials said that Anthropic reported on Thursday that one of its test models submitted 19 non-immigrant visa applications through the department's website public form in August and 1 in May, none of which were processed, and the system was not breached or hacked. Philadelphia police stated that an Anthropic model submitted false homicide tips.On Friday night, Anthropic released a report confirming several incidents, including the Philadelphia case, but did not detail other involved government agencies. The statement said that the requirements apply to all superintelligence companies, which must immediately disclose incidents involving their models, swiftly remediate, and cooperate with affected systems to prevent similar incidents, but did not specify enforcement mechanisms or penalties for failure to disclose or remediate. Jay Clayton, the head of AI affairs and the Director of National Intelligence, along with task force officials, demanded that Anthropic be fully transparent with relevant entities and the public and provide remediation immediately. The task force co-chairs include Federal Trade Commission Chair Andrew Ferguson, Office of Personnel Management Director Scott Kupor, and Deputy Secretary of Defense Emil Michael.