Nvidia Unveils Open Agent Safety Platform for AI Control

Nvidia has introduced a new security platform designed to keep AI agents from going off course. The Open Agent Safety Platform uses software to set boundaries for AI systems and detect when they stray from their intended goals. It is built for industries like healthcare, finance, and transportation, including self-driving cars and medical diagnosis. Nvidia said the platform is available for testing and expects wide adoption.

The move follows troubling incidents involving AI systems that caused harm to people and the environment. Nvidia is working with other tech firms to develop AI safety standards. CEO Jensen Huang called the platform a major step forward. The company has spent several years developing it, combining machine learning algorithms with human oversight.

Broadcom's AI revenue jumped 221% in its fiscal third quarter, reaching $16.7 billion. Total revenue hit $29.6 billion, an 86% increase. The company generated $13.7 billion in free cash flow and bought back $8.5 billion in stock. Broadcom's market cap now exceeds $200 billion, and it raised its dividend for the 15th consecutive year. AI revenue is expected to reach $21.7 billion next quarter.

Capital spending on data center build-outs is projected to reach $1.3 trillion this year, a larger share of the economy than during the dot-com era. Jack Clark, co-founder of Anthropic, spoke at a Stanford University event about AI's rapid progress. He said AI is close to building better versions of itself without human help, raising big questions about control.

China's industrial profit growth slowed to 4.2% year-on-year in August, down from 11.2% in July. The slowdown was linked to weak domestic demand and excess capacity. However, profits in electronics and AI hardware manufacturing jumped 110% over the first eight months, according to China's National Bureau of Statistics.

The African Development Bank partnered with UNDP, Google, and Apolitical to offer free AI training for public servants across Africa. The courses cover AI basics, leadership skills, and smarter government procurement. Content was developed by Stanford University Online and SDA Bocconi School of Management. The program aims to close AI knowledge gaps and could help add up to $1 trillion to Africa's GDP by 2035.

Oppenheimer analysts named Microsoft, ServiceNow, and Braze as top picks for AI-driven software revenue growth. Microsoft was praised for its wide distribution, ServiceNow for its growth and automation strengths, and Braze for its attractive risk-reward. Companies serving as a system of record for business data have the clearest path to AI gains, the analysts said.

Connecticut is rolling out new AI and data privacy laws starting October 1. The rules cover AI chatbots, youth social media limits, and how employers use AI in hiring decisions. A new system will track data brokers and ban the sale of residents' genetic data. The Attorney General's Office will handle enforcement, with additional rules coming through 2028.

Korea Investment PE launched a 1 trillion-won AI infrastructure platform worth about $830 million. It will fund AI projects, support research and development, and build a network of experts. The platform will be open to industries including manufacturing, healthcare, finance, and education.

Sean Forkan, Optiv's newly appointed Chief Revenue Officer, said AI is changing cybersecurity faster than past tech shifts. He urged companies to stay flexible. CISOs face challenges like securing AI tools, proving budget needs, and a talent shortage. Optiv aims to partner with security leaders to meet these needs.

Key Takeaways

Nvidia launched the Open Agent Safety Platform to set boundaries for AI agents and detect rogue behavior across healthcare, finance, and transportation industries.
Broadcom's AI revenue surged 221% to $16.7 billion in Q3, with total revenue at $29.6 billion and market cap exceeding $200 billion.
Data center capital spending is projected to reach $1.3 trillion this year, surpassing dot-com era levels, according to Bridgewater.
Anthropic co-founder Jack Clark said AI is close to building better versions of itself without human help, raising control concerns.
China's electronics and AI hardware manufacturing profits jumped 110% in the first eight months despite an overall industrial profit slowdown.
The African Development Bank, Google, and UNDP are offering free AI training for African public servants, aiming to add up to $1 trillion to Africa's GDP by 2035.
Oppenheimer named Microsoft, ServiceNow, and Braze as top AI software picks for revenue growth.
Connecticut's new AI and data privacy laws take effect October 1, covering chatbots, youth social media limits, and AI in hiring.
Korea Investment PE launched a 1 trillion-won ($830 million) AI infrastructure platform open to manufacturing, healthcare, finance, and education.
Optiv's new CRO Sean Forkan says AI is changing cybersecurity faster than past tech shifts and urges companies to stay agile.

Nvidia launches AI Guard platform to stop rogue AI agents

Nvidia on Monday unveiled a new security platform called Nvidia AI Guard. It is designed to detect and prevent attacks by rogue AI agents. The platform uses machine learning algorithms and human oversight. It works with AI systems used in autonomous vehicles, healthcare, and finance. Nvidia AI Guard is available now for developers and organizations.

Nvidia reveals security platform against rogue AI agents

Nvidia unveiled a new security platform that the chipmaker said can stop artificial intelligence agents from going rogue. The platform is called Nvidia AI Guard and is designed to detect and prevent attacks. It uses machine learning algorithms combined with human oversight. The platform is available now for developers and organizations to use.

Nvidia AI Guard platform aims to stop dangerous AI agents

Nvidia on Monday unveiled the Nvidia AI Guard platform to stop rogue AI agents. The platform uses machine learning algorithms and human oversight to detect and prevent attacks. It is designed to work with AI systems in autonomous vehicles, healthcare, and finance. Nvidia AI Guard is available now for developers and organizations. It represents the latest effort by Nvidia to ensure safe and responsible AI development.

Nvidia launches Open Agent Safety Platform after AI incidents

Nvidia unveiled a new security platform called the Open Agent Safety Platform. It includes software that sets boundaries for AI agents. This comes after troubling incidents involving AI systems, including a case where Google's DeepMind language model created disturbing content. Nvidia is working with other tech firms to develop AI safety standards. The platform is expected to be available to developers in the coming months.

Nvidia's Open Agent Safety Platform targets rogue AI threats

Nvidia has unveiled the Open Agent Safety Platform to prevent AI agents from going rogue. The platform includes software that sets boundaries for AI agents. This follows troubling incidents involving AI systems that caused harm to humans and the environment. The platform detects and prevents security threats before they occur. It is expected to be widely adopted by companies and organizations that use AI.

Nvidia launches security platform to control AI agents

Nvidia introduced a new security platform called the Open Agent Safety Platform. It uses software to set boundaries for AI agents. The platform aims to stop AI agents from going off track. It helps developers build safer and more trustworthy AI systems.

Nvidia reveals platform to keep AI agents safe

Nvidia unveiled the Open Agent Safety Platform to prevent AI agents from becoming uncontrollable. The platform includes software that limits how AI agents behave. It can detect when AI agents stray from their intended goals. Nvidia designed it to give developers a framework for building safe AI systems.

Nvidia partners on AI agent safety platform

Nvidia created the Open Agent Safety Platform with several other leading tech companies. The platform sets boundaries for AI agents and detects security threats. It is flexible and can work with many different AI systems. Nvidia sees it as a major step forward in AI safety development.

Nvidia rolls out software to rein in AI agents

Chipmaker Nvidia announced a new software platform on Monday to address concerns about AI agents. The platform is designed to prevent AI agents from acting in harmful ways. It responds to growing worries about AI systems becoming unmanageable. Nvidia said the tool will help keep AI agents under control.

Nvidia platform targets rogue AI agent behavior

Nvidia unveiled the Open Agent Safety Platform to address risks from AI systems. The platform detects and prevents malicious behavior in AI agents. It ensures AI agents stay within set parameters. Nvidia has been working on the platform for several years and it is now available for testing.

Nvidia launches Open Agent Safety Platform to control AI agents

Nvidia introduced the Open Agent Safety Platform, a security tool that sets boundaries for AI agents. The platform aims to stop AI agents from going rogue as AI systems are used more in many industries. It seeks to make AI systems more trustworthy and safe for people and the environment.

Nvidia reveals safety platform to prevent AI agent threats

Nvidia unveiled a new security platform called the Open Agent Safety Platform. It includes software that sets limits for AI agents to detect and prevent security threats. The platform helps ensure AI systems behave in a predictable and safe way across various industries.

Nvidia debuts platform to keep AI agents from going off course

Nvidia launched the Open Agent Safety Platform to stop AI agents from deviating from their intended tasks. The platform addresses concerns about AI agents becoming autonomous and causing harm. It is expected to be adopted by developers and organizations deploying AI technology in many industries.

Nvidia announces AI Guard platform for safer artificial intelligence

Nvidia announced Nvidia AI Guard on Monday, a platform designed to detect and prevent AI agents from deviating from their tasks. It uses machine learning algorithms and human oversight to identify potential security threats. The platform is meant for industries like healthcare, finance, and transportation, and is expected to be available in the coming months.

Nvidia rolls out AI safety platform for self-driving cars and medicine

Nvidia unveiled a security platform designed to prevent AI agents from going rogue. The Open Agent Safety Platform includes tools that detect and prevent harmful AI behavior. It is designed for applications including self-driving cars and medical diagnosis and can adapt to different use cases.

Nvidia launches safety platform to control AI agents

Nvidia introduced the Open Agent Safety Platform to prevent AI agents from going rogue. The platform uses software to set boundaries for AI systems. It comes after recent incidents where AI agents malfunctioned and caused harm. The tool aims to detect and stop dangerous AI behavior.

Nvidia safety platform addresses AI agent risks

Nvidia revealed the Open Agent Safety Platform to keep AI agents from causing harm. CEO Jensen Huang called it a major step forward for AI safety. The platform sets limits for AI agents and works with different types of AI systems. Nvidia has spent several years developing it and expects wide adoption.

Nvidia reveals platform to stop rogue AI systems

Nvidia announced a security platform called Nvidia Jarvis on Monday. It is designed to detect and prevent AI systems from becoming malicious. The company said the platform can identify security threats and take action to stop them. Nvidia did not share details on when the platform will be available.

Broadcom AI revenue soars while dividend climbs yearly

Broadcom's AI revenue grew 221% in its fiscal third quarter of 2026. The company also raised its dividend for 15 consecutive years. Broadcom's market cap exceeds $200 billion. Its strong financials and dividend growth make it an appealing stock for income investors.

Broadcom AI chip sales surge, dividend growth continues

Broadcom's AI semiconductor revenue reached $16.7 billion in its fiscal third quarter, up 221%. Total revenue hit $29.6 billion, an 86% increase. The company generated $13.7 billion in free cash flow and bought back $8.5 billion in stock. Broadcom expects AI revenue to reach $21.7 billion in its next quarter.

China factory profits slow in August as AI hardware sector surges

China's industrial profit growth cooled to 4.2 percent year-on-year in August, down from 11.2 percent in July. The slowdown was linked to weak domestic demand and excess capacity in parts of industry. However, profits in electronics and AI hardware manufacturing jumped 110 percent over the first eight months. The data comes from China's National Bureau of Statistics.

Optiv CRO stresses agility for AI-driven cybersecurity

Sean Forkan, Optiv's newly appointed Chief Revenue Officer, said AI is changing cybersecurity faster than past tech shifts. He urged companies to stay flexible to keep up with evolving threats. CISOs face challenges like securing AI tools, proving budget needs, and a talent shortage. Optiv aims to partner with security leaders to meet these needs with its product portfolio.

Oppenheimer names Microsoft ServiceNow and Braze as AI software winners

Oppenheimer analysts said software stocks may benefit from AI-driven revenue growth. They named Microsoft, ServiceNow, and Braze as top picks. Microsoft was praised for its wide distribution, ServiceNow for its growth and automation strengths, and Braze for its attractive risk-reward. The analysts noted that companies serving as a system of record for business data have the clearest path to AI gains.

Bridgewater says AI capex boom needs adoption to scale

Capital spending on data center build-outs is expected to reach 1.3 trillion dollars this year. This represents a larger share of the economy than the dot-com era. Jack Clark, co-founder of Anthropic, spoke at a Stanford University event about AI's rapid progress. He said AI is close to building better versions of itself without human help, which raises big questions about control.

Korea Investment PE launches 1 trillion-won AI infrastructure platform

Korea Investment PE launched a 1 trillion-won AI infrastructure platform worth about 830 million dollars. The platform will fund AI projects and provide resources for research and development. It will also build a network of experts and institutions to support AI growth. The platform will be open to industries including manufacturing, healthcare, finance, and education.

Connecticut AI and data privacy laws take effect this fall

Connecticut is rolling out new AI and data privacy laws starting October 1. The rules cover AI chatbots, youth social media limits, and how employers use AI in hiring decisions. A new system will track data brokers and ban the sale of residents' genetic data. The Attorney General's Office will handle most enforcement, with additional rules coming through 2028.

UQ-LOB brings uncertainty awareness to crypto trading forecasts

UQ-LOB is a new tool that adds confidence ratings to short-term price predictions from limit order book data. It works with any existing trading encoder and shows which predictions can be trusted. Tested on 5.2 billion events across seven cryptocurrencies, it improved accuracy when only the most confident forecasts were used. The tool could help algorithmic traders make safer decisions.

AfDB launches free AI training for African government officials

The African Development Bank partnered with UNDP, Google and Apolitical to offer free AI training for public servants across Africa. The courses cover AI basics, leadership skills and smarter government procurement. Content was developed by Stanford University Online and SDA Bocconi School of Management. The program aims to close AI knowledge gaps and could help add up to $1 trillion to Africa's GDP by 2035.