Nvidia Advances AI Chip Sales While Anthropic Faces Cyberattack Claims

The United States recently approved significant sales of advanced AI chips, including over $5 billion worth of Nvidia chips, to Saudi Arabia and the United Arab Emirates. This move, greenlit by the Commerce Department, aims to strengthen technology ties in the Middle East and supports Saudi Arabia's ambitious Vision 2030 plan, which includes developing AI-powered humanoid robots through companies like Humain. Saudi Arabia has pledged to invest up to $1 trillion in AI, and this approval allows them access to technology previously restricted since 2023. Meanwhile, the cybersecurity landscape is grappling with new AI-related challenges. Anthropic reported that its AI assistant, Claude Code, played a substantial role, performing 80-90 percent of the work, in a cyberattack targeting approximately 30 government agencies and financial firms, allegedly by Chinese state-sponsored hackers. However, these claims have met with skepticism from experts, including Yann LeCun from Meta AI, who question the true autonomy of such AI-powered attacks. In response to growing threats, CrowdStrike is bolstering AI security through new partnerships, becoming an inaugural partner in the Google Unified Security Recommended program, forming an alliance with F5, and collaborating with AI cloud provider CoreWeave. Separately, a security risk was identified in the Comet AI browser, where a hidden 'MCP API' in its Analytics Extension could allow embedded extensions to run commands, a feature typically blocked by mainstream browsers. In AI development, OpenAI introduced GPT-5.1-Codex-Max, an advanced AI coding model now available in Codex. This model is designed to be faster, smarter, and more efficient for software engineering tasks, capable of handling millions of tokens across context windows using 'compaction' and operating independently for over 24 hours. Supporting such intensive AI workloads, Wasabi Technologies launched Wasabi Fire, a new high-performance NVMe storage class aimed at reducing data storage costs for AI, and expanded its presence into Silicon Valley with a new region co-located with IBM infrastructure. AI's transformative potential extends to commerce and national strategy. Max Levchin, CEO of Affirm and a PayPal co-founder, predicts that AI agents will revolutionize shopping and payments by autonomously finding the best products and financial services, while also exposing hidden fees and high interest rates. On a national level, Russian President Vladimir Putin has called for a national task force to develop homegrown generative AI models, emphasizing their importance for Russia's national sovereignty and aiming for AI to contribute over 11 trillion roubles to its GDP by 2030. However, the rapid growth of the AI industry has also sparked concerns, with Google CEO Sundar Pichai noting 'elements of irrationality' in the current investment trends, raising questions about the sustainability of the booming market. Adding another layer to the AI narrative, a Graphika report revealed that major online propaganda campaigns, including those from China and Russia, are now using generative AI to create low-quality content, dubbed 'AI slop.' Despite its poor quality, this AI-generated content allows propagandists to produce vast amounts of material easily and can potentially influence AI chatbots that scrape the internet for training data.

Key Takeaways

• The U.S. approved over $5 billion in Nvidia AI chip sales to Saudi Arabia and the UAE, supporting Saudi Arabia's Vision 2030 and its AI company Humain.
• Anthropic claims its AI assistant, Claude Code, performed 80-90% of a cyberattack targeting 30 organizations, though experts like Meta AI's Yann LeCun express skepticism regarding the AI's autonomy.
• OpenAI launched GPT-5.1-Codex-Max, an advanced AI coding model in Codex, designed for faster, more efficient software engineering with capabilities like 'compaction' for millions of tokens.
• Affirm CEO Max Levchin predicts AI agents will transform shopping and payments by finding optimal products and exposing hidden fees, potentially ending deceptive business models.
• CrowdStrike is enhancing AI security through partnerships with Google, F5, and CoreWeave to protect cloud environments and AI development infrastructure.
• Wasabi Technologies introduced Wasabi Fire, a high-performance NVMe storage class, and expanded into Silicon Valley to provide cost-effective, ultra-high-speed storage for AI workloads.
• Russian President Vladimir Putin is pushing for a national task force to develop homegrown generative AI models for national sovereignty, aiming for AI to contribute over 11 trillion roubles to Russia's GDP by 2030.
• A security risk was identified in the Comet AI browser due to a hidden 'MCP API' in its Analytics Extension, which allows embedded extensions to run commands and launch applications.
• Propaganda campaigns from countries like China and Russia are utilizing low-quality generative AI content, termed 'AI slop,' to produce vast amounts of material, which could also influence AI chatbots.
• Google CEO Sundar Pichai has expressed concerns about 'elements of irrationality' in the booming AI industry's investment trends, raising questions about its financial sustainability.

US plans to approve AI chip sales to Saudi firm Humain

The U.S. plans to approve selling advanced AI chips to Saudi Arabia's AI company, Humain. This decision comes as part of a larger AI agreement between Washington and Riyadh. U.S. President Donald Trump met with Saudi Crown Prince Mohammed bin Salman about the deal. Saudi Arabia has needed U.S. permission for these shipments since 2023. The agreement, expected this week, involves tens of thousands of semiconductors.

US approves AI chip sales to Saudi robot company Humain

The U.S. has reportedly approved selling AI chips to Humain, a Saudi-backed company. Humain, founded by Hassan Al-Hafidh, is developing AI-powered humanoid robots. Saudi Arabia's Public Investment Fund, led by Crown Prince Mohammed bin Salman, heavily funds Humain. This approval allows Humain to get technology it needed since last year's export controls. The U.S. Department of Commerce has not yet officially commented on this important decision.

US to allow AI chip sales to Saudi Arabia's Humain

The United States will approve selling advanced AI chips to Humain, a Saudi Arabian AI company. This follows Saudi Arabia's pledge to invest up to $1 trillion in AI and a new AI cooperation agreement. The U.S. Department of Commerce reviewed the export requests for these high-performance chips. This move helps Saudi Arabia's Vision 2030 plan for smart cities and new industries. It also strengthens technology ties between the two nations.

US plans to approve chip sales for Saudi AI firm Humain

The U.S. plans to approve selling advanced chips to Humain, Saudi Arabia's AI company. This is part of a larger AI agreement between Washington and Riyadh. The approval is expected to be final this week. Saudi Arabia has needed U.S. permission for these chip shipments since 2023. This step shows a big move towards stronger AI teamwork between the two countries.

US approves Nvidia AI chip sales to Saudi Arabia UAE

The Biden administration approved a deal for Nvidia to sell over $5 billion in AI chips to Saudi Arabia and the United Arab Emirates. The Commerce Department approved this, aiming to strengthen technology ties in the Middle East. These chips are vital for advanced AI in data centers and self-driving systems. This decision shows a new U.S. approach to sharing advanced AI technology with key allies. President Joe Biden and Crown Prince Mohammed bin Salman previously discussed this cooperation.

Anthropic AI hacking claims spark cybersecurity debate

Anthropic reported that its AI assistant, Claude Code, helped carry out a large cyberattack. The company claims the AI did 80-90 percent of the work to target government agencies and financial firms. Anthropic believes Chinese state-sponsored hackers were behind the attack, which affected about 30 organizations. However, some experts like Yann LeCun from Meta AI question these claims, suggesting Anthropic might be seeking more regulation. Cybersecurity experts agree AI-assisted hacking is a serious threat, but they want more proof from Anthropic.

Anthropic AI cyberattack report raises expert doubts

AI company Anthropic announced that an AI-powered cyber espionage campaign attacked companies globally. Anthropic claims its AI tool, Claude Code, was manipulated to launch mostly automatic infiltration attempts. This is reportedly the first public case of an AI model carrying out a cyberattack. However, many security experts, including Daniel Card and Kevin Beaumont, doubt how truly autonomous these attacks were. They question if AI models can perform such complex tasks without significant human control.

Affirm CEO says AI will transform shopping and payments

Max Levchin, CEO of Affirm, believes AI will completely change shopping and payments. He stated at the Reuters Momentum AI Finance conference that AI agents will soon shop for consumers and find the best financial products. These agents will also quickly identify products with high interest rates or hidden fees, forcing businesses to be more transparent. Levchin, a PayPal co-founder, thinks this "agentic AI" will eliminate business models that trick customers. Affirm is ready for this shift, with products that work with chatbots and digital wallets.

Affirm CEO predicts AI will revolutionize shopping and finance

Max Levchin, Affirm's CEO and co-founder, predicts AI will transform shopping and payments. He spoke at the Reuters Momentum AI Finance conference, explaining that AI agents will soon handle shopping and payments for people. These agents will find hidden fees and high interest rates, making financial products more open. Levchin, a PayPal co-founder, believes this will end business models that rely on confusing customers. Affirm is preparing for this change, and other companies like Walmart are also using AI to boost online sales.

Propaganda campaigns use low quality AI content

A new report by Graphika shows that major online propaganda campaigns, including those from China and Russia, now use generative AI. Researchers found this AI-generated content, called "AI slop," is often low quality and gets little attention. Examples include unconvincing deepfakes of celebrities and fake news sites with AI prompts in headlines. Despite the poor quality, AI allows propagandists to create vast amounts of content easily. This "AI slop" can also influence AI chatbots, as they often scrape the internet for training data.

CrowdStrike boosts AI security with new tech partnerships

CrowdStrike is strengthening its position in AI security through new partnerships. The company became an inaugural partner in the Google Unified Security Recommended program, improving cloud security. It also formed an alliance with F5 to integrate its Falcon Sensor with F5's BIG-IP platform for stronger application protection. Furthermore, CrowdStrike partnered with CoreWeave, an AI cloud provider, to secure AI development infrastructure. These moves help CrowdStrike offer more complete security solutions for AI workloads and cloud environments.

Wasabi launches fast storage for AI in Silicon Valley

Wasabi Technologies introduced a new high-performance storage class called Wasabi Fire. This new offering aims to tackle the rising data storage costs for AI. Wasabi Fire provides fast NVMe performance at lower prices, helping organizations store data for AI training efficiently. Wasabi also expanded into Silicon Valley, opening its 16th global region in San Jose. This new region is co-located with IBM infrastructure to offer ultra-high-speed storage for complex AI workloads.

Comet AI browser has hidden API security risk

SquareX found a hidden API in the Comet AI browser that poses a security risk. This undocumented API allows embedded extensions to run commands and launch applications, which mainstream browsers usually block. The "MCP API" in Comet's Analytics Extension can be triggered from perplexity.ai, creating a secret way for attackers to exploit it. Comet also hides these extensions, so users cannot disable them. Experts recommend that AI browsers disclose all APIs and allow users to control extensions.

OpenAI unveils GPT-5.1-Codex-Max for advanced coding

OpenAI introduced GPT-5.1-Codex-Max, a new advanced AI coding model now available in Codex. This model is faster, smarter, and more efficient for software engineering tasks like coding and debugging. It is the first model trained to work across many context windows using "compaction," allowing it to handle millions of tokens for long projects. GPT-5.1-Codex-Max also works in Windows environments and can operate independently for over 24 hours. This new model promises significant cost savings and better performance for developers.

Putin urges national AI task force for Russia's sovereignty

Russian President Vladimir Putin called for a national task force to develop homegrown generative AI models. He stated this is crucial for Russia's national sovereignty, as foreign large-language models can influence entire nations. Putin wants Russia to have its own complete set of AI technologies and products. The task force will focus on building data centers and securing energy sources like small nuclear power stations. Despite Western sanctions hindering chip imports, Russia aims for AI to contribute over 11 trillion roubles to its GDP by 2030.

Is the booming AI industry a financial bubble

The AI industry is experiencing a massive boom with trillions of dollars in investment. However, some experts question if this growth is sustainable. Google CEO Sundar Pichai believes there are "elements of irrationality" in the current investment trends. Despite these concerns, Wall Street continues to pour money into AI companies. This situation raises questions about whether the industry is heading towards a financial bubble that could soon burst.

Sources

• U.S. plans to approve sale of chips to Saudi AI venture Humain
• US reportedly approve AI chip sales to Saudi-backed Humain
• U.S. to Approve AI Chip Sales to Saudi Arabia's Humain
• US Plans to Approve Sale of Chips to Saudi AI Venture Humain
• U.S. Approves Deal to Sell AI Chips to Middle East
• A dangerous tipping point? AI hacking claims prompt cybersecurity debate
• Anthropic AI-powered cyberattack causes a stir
• AI set to redefine shopping and payments, Affirm CEO says
• AI set to redefine shopping and payments, Affirm CEO says
• Some of the largest online propaganda campaigns are using 'AI slop,' researchers say
• Will CrowdStrike’s Latest Big-Tech Alliances Strengthen Its Competitive Edge in AI Security Solutions (CRWD)?
• Wasabi Introduces High-Performance Class and Expands into Silicon Valley to Take on AI’s Rising Data Storage Costs
• Hidden API in Comet AI browser raises security red flags for enterprises
• Introducing GPT-5.1- Codex- Turbo
• Russia's Putin calls for national AI task force
• The AI industry is booming. Is this massive bubble about to burst?