Anthropic's Claude model integrated into SAP's Business AI Platform

Key Takeaways

SAP partners with Anthropic to bring Claude to its AI platform

SAP and Anthropic announced a new partnership to integrate Claude, an advanced AI model, into SAP's Business AI Platform. This collaboration will make Claude a core part of SAP's Joule agents, helping businesses handle complex tasks like closing books and managing supply chains. The integration uses SAP's existing systems and data to ensure safety and trust while giving customers more flexibility in choosing AI models. SAP CEO Christian Klein stated that this partnership is key to building an autonomous enterprise where AI understands business context and operates within necessary controls.

SAP invests in n8n and values the AI platform at $5.2 billion

SAP has made a strategic investment in n8n, an AI orchestration platform, raising its valuation to $5.2 billion. This value is more than double the $2.5 billion n8n was worth less than a year ago. As part of the deal, n8n's workflow automation tools will be built directly into Joule Studio, SAP's environment for creating AI agents. The partnership allows enterprise teams to connect SAP systems with over 1,400 different business tools and AI models using a secure, compliant automation canvas.

SAP integrates n8n to help enterprises scale agentic AI systems

SAP is integrating n8n into its ecosystem to help large companies scale their use of agentic AI. The integration gives developers a no-code and low-code canvas to build automation workflows that connect SAP systems with more than 1,000 external tools. n8n is already used by over 1,400 enterprise customers and 1.7 million developers. SAP CEO Christian Klein noted that this move supports their new SAP Business AI Platform, which is designed to run AI safely within real business contexts.

Supply chain attack hits TanStack, Mistral AI, and UiPath with malware

A new cyberattack called Mini Shai-Hulud compromised over 400 versions of 170 software packages, including those from TanStack, Mistral AI, and UiPath. Attackers used security flaws in GitHub to inject malicious code that steals credentials and spreads to other systems. The malware hid itself by using stolen digital certificates to make the infected packages look legitimate. Security experts warn that this attack targets the tools developers use to build software, creating a risk that could spread to thousands of applications.

Malicious Mistral AI and TanStack packages may have stolen cloud credentials

Microsoft and security firms found that compromised versions of Mistral AI and TanStack packages contained hidden malware designed to steal credentials. The malicious code silently downloaded a second-stage payload on Linux machines whenever the packages were imported. Attackers used this access to steal GitHub tokens, cloud API keys, and CI/CD secrets. Experts advise developers to immediately rotate any exposed credentials and block connections to the malicious IP address involved in the attack.

Prediction markets show high confidence in a Trump visit to China

Prediction markets indicate a very high probability that Donald Trump will visit China by April 2026. One market shows a 98.9% chance of the visit occurring, with over $6 million wagered on the outcome. These markets also show strong interest in other events, such as Bitcoin reaching $150,000 by June 2026 and a potential peace deal between the US and Iran. The high volume of bets on these geopolitical and crypto events highlights significant speculation about future market movements.

AI security startup Grego AI wins record $250,000 bounty for finding a bug

Grego AI, a new cybersecurity startup, announced it won a record $250,000 bug bounty for finding a flaw in a major blockchain protocol. The company uses AI agents to scan codebases and find complex security issues that human auditors might miss. The discovered bug could have allowed attackers to steal $27.7 million if left unfixed. Grego AI plans to expand its AI-driven security tools to protect enterprise software in finance, healthcare, and government sectors.

Nvidia commits over $40 billion to AI investments led by OpenAI stake

Nvidia has committed more than $40 billion to AI-related equity investments in early 2026. The largest part of this investment is a $30 billion stake in OpenAI. Nvidia also announced seven other multi-billion dollar deals, including a $5 billion investment in Cerebras Systems. These moves are part of Nvidia's strategy to strengthen its AI ecosystem through partnerships with major tech companies like Microsoft and Alphabet's DeepMind.

Secure Code Warrior partners with AWS to train developers on AI security

Secure Code Warrior and Amazon Web Services (AWS) signed an agreement to create training modules for Amazon Bedrock. These new courses teach developers how to secure AI applications against risks like prompt injection and data leaks. The training includes interactive labs and challenges that let developers practice fixing security issues in a safe environment. This partnership aims to help companies build AI tools that are secure by default while accelerating their adoption.

Sales team uses AI to write follow-up emails in minutes

A sales team built a system to use AI for writing follow-up emails after calls. The process involves using voice-to-text software to record call notes, then feeding those notes into an AI tool to draft a personalized email. The team can now send these emails in three to five minutes instead of the previous 20 to 60 minutes. This change saves them about one full workday every week and helps them close more deals.

Homeland Security chief supports voluntary rules for AI safety

Alejandro Mayorkas, the former Homeland Security chief, supports voluntary policies for securing advanced AI models. He believes that working with market forces is faster and more effective than creating new government regulations. Mayorkas warned that AI tools could allow less skilled hackers to launch powerful cyberattacks. He urged the federal government to act quickly to ensure AI safety without creating a confusing patchwork of state laws.

Trump administration plans no new agency to police AI products

Kevin Hassett, a director in the Trump administration, said there are no plans to create a new government agency to regulate AI. Instead, the administration is revamping existing cybersecurity programs to include AI companies. Officials are working with the private sector to ensure AI models do not cause extreme harm before they are released. Hassett stated that the current approach involves cooperation between the government and private companies rather than a large new bureaucracy.

Sources

• SAP and Anthropic Plan to Bring Claude to SAP Business AI Platform
• n8n valuation doubles to $5.2bn as SAP makes strategic investment and plans to embed the AI platform into Joule Studio
• SAP Integrates n8n to Scale Agentic AI for Enterprises
• TanStack, Mistral AI, UiPath Hit in Fresh Supply Chain Attack
• Compromised Mistral AI and TanStack packages may have exposed GitHub, cloud and CI/CD credentials in 'mini Shai Hulud'  malware infection — supply-chain campaign spreads across npm and AI developer ecosystems like wildfire
• Trump China Visit: Prediction Market Odds
• AI security startup Grego AI debuts, claims record $250,000 bounty for AI-found exploit
• Nvidia's AI Investment Bets Top $40 Billion In 2026, Led By OpenAI Stake
• Secure Code Warrior Signs Strategic Collaboration Agreement with AWS to Support Security-Conscious Developers with Amazon Bedrock AI Learning Modules
• A Simple System to Increase Sales With AI Follow-Ups
• Mayorkas endorses ‘voluntary’ policies for tackling threats posed by advanced AI models
• No new bureaucracy to police AI products in the US, Hassett says