Anthropic has actively blocked attempts by researchers linked to foreign governments to use its Claude model for creating biological weapons and advanced missiles. The company identified and stopped these dangerous projects, including efforts from accounts in northern Yemen controlled by Iran-backed Houthi rebels. While Anthropic shared five specific examples of these incidents, experts suggest these cases represent only a fraction of the broader threat.
Market reactions to AI developments remain mixed. Goldman Sachs boss Kim Posnett argues the sector is in an investment supercycle rather than facing a bubble, while Nvidia CEO Jensen Huang claims artificial general intelligence has arrived following OpenAI's release of the GPT-6 Astra model. Goldman analyst Rich Privorotsky notes the real shift lies in the model's increased intelligence, which could reshape market demand.
Corporate adoption continues to accelerate with tangible financial impacts. ServiceNow reported its AI products now exceed $1 billion in annual contract value, with customer usage of agentic AI tools growing ninefold. Meanwhile, Hugging Face CEO Clement Delangue dismissed AI extinction fears as misplaced priorities, comparing the concern to asking an air conditioning repairman about climate change.
Regulatory and operational shifts are also underway. California Governor Newsom signed a law mandating independent safety reviews for AI systems. In Australia, Nvidia partners with eight firms to more than double the nation's AI computing capacity from 1.6 gigawatts, utilizing DSX platforms and Quantum InfiniBand technology. Conversely, some sectors face new challenges, such as lawyers generating extra busywork despite email efficiency gains, and asset managers risking homogenized portfolios due to reliance on identical AI tools.
Key Takeaways
- Anthropic blocked foreign researchers from using Claude to create biological weapons and advanced missiles.
- Users in Iran-backed Houthi-controlled Yemen attempted to develop missiles using the Claude AI model.
- Goldman Sachs boss Kim Posnett describes the current market as an AI investment supercycle rather than a bubble.
- Nvidia CEO Jensen Huang states artificial general intelligence has arrived with OpenAI's GPT-6 Astra model.
- ServiceNow's AI products now exceed $1 billion in annual contract value with a ninefold growth in agentic AI customers.
- Hugging Face CEO Clement Delangue compared AI extinction fears to asking an AC repairman about climate change.
- Needham raised ServiceNow's price target to $155 due to strong AI-driven revenue growth.
- California Governor Newsom signed a law requiring independent safety reviews for AI systems.
- Nvidia partners with eight Australian firms to more than double the country's AI computing capacity.
- Asset managers risk producing similar portfolios as 55% have already integrated AI tools according to a Mercer survey.
Anthropic stops scientists from making bioweapons with Claude
Anthropic found that scientists tried to use its Claude AI model to create biological weapons this year. The company blocked these attempts and shared five examples of the incidents. These efforts were made by researchers connected to foreign governments who wanted to build deadly tools.
Anthropic blocked foreign researchers from using AI for bioweapons
Anthropic discovered multiple attempts by researchers linked to foreign governments to use Claude for creating biological weapons. The company successfully detected and stopped these dangerous projects before they could proceed. Experts believe these cases are only a small part of the larger problem.
Users in Yemen tried to build advanced missiles with AI
Anthropic reported that users in northern Yemen tried to use the Claude AI model to develop advanced missiles. These accounts were located in territory controlled by Iran-backed Houthi rebels. The company identified the attempts by analyzing the activity on the accounts but did not share more details.
Goldman boss says we are in an AI investment supercycle
Kim Posnett, the banking boss at Goldman Sachs, says people are asking the wrong questions about an AI bubble. He claims that the current situation is actually an AI investment supercycle. This view contrasts with concerns about a potential market crash in artificial intelligence stocks.
Hugging Face CEO compares AI extinction fears to AC guy
Hugging Face CEO Clement Delangue dismissed warnings from a former Anthropic researcher about AI extinction risks. He compared asking about these dangers to asking an air conditioning repairman about climate change. The CEO made this comment while discussing the role of AI safety experts.
Needham raises ServiceNow target to $155 due to AI growth
Needham analyst Mike Cikos increased the price target for ServiceNow to $155 because the company's AI products now exceed $1 billion in annual contract value. The number of customers using agentic AI tools has grown ninefold, and ServiceNow aims to reach $1.5 billion in AI ACV by the end of 2026. Investors see about 18% upside potential as the stock trades at $131.17, reflecting strong confidence in the company's future AI-driven revenue.
Nvidia CEO Jensen Huang says AGI has arrived with new model
Nvidia CEO Jensen Huang stated that artificial general intelligence has arrived after OpenAI released its GPT-6 Astra model. Although the term AGI has existed since 1997, Huang's comments suggest this new model might start the AGI revolution rather than being fully AGI itself. Goldman Sachs analyst Rich Privorotsky believes the real breakthrough is the increased intelligence of the model, which will reshape market demand.
AI helps lawyers write better emails but may create extra work
AI tools offer real efficiency gains for lawyers by helping them produce polished emails in English, French, and Spanish. Pierre Zickert, head of legal tech at Hengeler Mueller, notes that while these tools save time, they might also generate additional busywork for legal teams. The technology allows for better communication across different languages, but the net effect on workload remains a point of discussion.
AI tools may cause asset managers to think too similarly
Widespread use of AI could make asset managers produce very similar investment portfolios because they use the same tools and data. A 2026 survey by Mercer found that 55% of 131 asset managers have integrated AI, with 91% planning to increase usage. Experts warn that relying on identical AI infrastructure might reduce the unique insights that come from diverse human perspectives.
Five top artificial intelligence stocks to watch today
MarketBeat identifies five artificial intelligence stocks to consider: Hut 8, Tempus AI, Upstart, SentinelOne, and SoundHound AI. Hut 8 focuses on cryptocurrency mining, while Tempus AI analyzes medical data for healthcare professionals. Upstart uses AI for lending, SentinelOne provides cybersecurity, and SoundHound AI generates music based on patterns.
Astera Labs beats Marvell as top AI stock pick for 2026
Investors are comparing Astera Labs Inc and Marvell Technology Inc to find the best AI stock for 2026. Astera Labs is growing revenue at a triple-digit pace but relies on one customer for 70% of its sales. Marvell Technology has a large portfolio and a ten-year partnership with Google but trades at a higher price. The article concludes that Astera Labs is the better buy because its OPO and NPO product lines will drive future growth.
Newsom signs law for independent AI safety reviews
Governor Newsom signed a new law that requires independent reviews of artificial intelligence systems. This rule aims to create safeguards for AI technology in the state. The report by Darrielle Fair highlights this new step in regulating artificial intelligence.
Nvidia partners with eight firms to double Australian AI power
Nvidia is working with eight Australian companies to more than double the nation's AI computing capacity. The current capacity is 1.6 gigawatts, and the new projects will increase this amount significantly. The partners include Firmus, Sharon AI, IREN, Megaport, ResetData, CDC, NEXTDC, and AirTrunk. These companies will build AI factories using Nvidias DSX platform, accelerated computing, and networking tools. Sharon AI plans to deploy up to 68,000 Nvidia GPUs connected by Quantum InfiniBand and Spectrum-X technology. IREN is using the DSX architecture at its 800-megawatt campus in South Australia. CDC has a footprint of over 550 megawatts and is developing another 800 megawatts. Nvidia will also share access to its Nemotron open models with groups like Heidi and Atlassian. Nvidia recently stopped a credit backing program for cloud providers due to antitrust concerns but may restart it later.
Sources
- Claude users found ways around safeguards for bioweapons research
- Anthropic Says it Blocked Efforts to Use AI for Biological Weapons
- Users in Houthi-held Yemen tried to develop advanced weapons with AI, Anthropic says
- Our question about the AI bubble is wrong, Goldman's banking boss says
- Hugging Face CEO dismisses ex-Anthropic researcher's AI warnings
- Needham raises ServiceNow target to $155 as AI annual contract value tops $1B
- Nvidia's Jensen Huang Claims AGI Has Arrived. Here's How to Invest.
- AI Promises to Save Lawyers Time; But Is It Actually Creating More ‘Busywork’?
- AI May Be Making Asset Managers Think Too Much Alike
- Best Artificial Intelligence Stocks To Consider
- Astera Labs vs. Marvell Technology: Which AI-Driven Tech Stock Is a Better Buy in 2026?
- Newsom signs new AI safeguards requiring independent reviews of artificial intelligence
- Nvidia is partnering with eight Australian firms to more than double the country's AI computing capacity
Comments
Please log in to post a comment.