Wall Street saw another rough day for AI-related stocks, with Nvidia and other chip companies losing ground for the second consecutive session. The S&P 500 dropped more than 3 percent, wiping out gains from the day before. The sell-off reflected growing investor concern about the concentration of AI megacaps in major indexes.
On the security front, Fortinet and Equinix launched FortiAIGate on the Equinix Distributed AI Hub, a solution designed to inspect AI traffic and enforce centralized security policies across large language model interactions. The architecture relies on NVIDIA GPU acceleration to handle inspection without slowing down AI workloads, and it lets organizations track per-token costs and set budgets while keeping data within their own infrastructure.
Sophos took a different approach to AI-era security with the release of Sophos Firewall v23, which includes over 30 improvements. The company also offers Sophos NDR, which uses AI to detect suspicious device activity without requiring agents, and a new Generative AI web category that helps organizations control which AI services employees can access.
Companies are spending more on cybersecurity, with 74.5 percent reporting larger budgets than a year ago, and 40.3 percent saying their next security dollar will target AI security and governance. However, 61.5 percent cannot confirm all their security tools are configured correctly, and employees are connecting to far more AI tools than IT teams realize.
Amazon entered the enterprise AI tools space with Web Answer on Bedrock AgentCore, which returns cited answers to natural-language questions in a single step. The tool combines search, retrieval, synthesis and citation, and supports enterprise controls like gateway authentication and per-account rate limits of 10 transactions per second in US East.
Certinia, a Salesforce-native platform, launched Veda AI agents for professional services firms, covering sales, staffing, project delivery and billing. The platform now includes 24 package agents, with plans to exceed 40 by year end. Its July acquisition of Moonnox added context graphs that connect emails, chats and calls to structured service data.
In China, Huawei, Tencent and Alibaba joined 20 other companies and institutions in supporting a voluntary AI agent safety pact launched at the World Internet Conference in Wuzhen. The agreement focuses on interoperability between AI agents from different providers and keeping users in control of key decisions.
Excelsoft launched four AI products developed from a pipeline of 38 ideas, 16 prototypes and nine customer pilots. The products include AI-levate App Suite, Saras SchoolAI, Saras Assessments in-a-Box and CareerSPARC. The AI stack runs on NVIDIA A6000 GPUs and uses models including Mistral and GPT-4o.
Orland Park is addressing problems with its AI-powered phone system, Humphrey, after residents reported long waits and failed connections. The village also approved a $43,200 yearly contract with Zion Cloud Solutions to use Google Gemini for data storage, which now answers 97 percent of staff questions.
Despite investor worries, Morningstar analysts say an AI spending slowdown is unlikely to trigger a recession. AI has driven nearly half of recent GDP growth from tech investment, and the median economic forecaster puts the chance of a recession in the next 12 months at 24 percent. Separately, Trefis estimates that a $5 million portfolio with 40 percent in AI megacaps could lose about $1.5 million in a 2000-style crash.
Key Takeaways
- The S&P 500 fell more than 3 percent during an AI stock sell-off, with Nvidia and other chip stocks declining for a second consecutive day.
- Fortinet and Equinix launched FortiAIGate on the Equinix Distributed AI Hub, using NVIDIA GPUs to inspect AI traffic with zero performance impact on workloads.
- 74.5 percent of companies report larger cybersecurity budgets than a year ago, and 40.3 percent say their next security dollar targets AI security and governance.
- Amazon introduced Web Answer on Bedrock AgentCore, offering cited answers to natural-language questions with per-account rate limits of 10 transactions per second in US East.
- Certinia launched Veda AI agents for professional services, with 24 package agents and plans to exceed 40 by year end, following its July acquisition of Moonnox.
- Huawei, Tencent and Alibaba joined 20 Chinese companies in a voluntary AI agent safety pact launched at the World Internet Conference in Wuzhen.
- Excelsoft launched four AI products including Saras SchoolAI and CareerSPARC, running on NVIDIA A6000 GPUs with Mistral and GPT-4o models.
- Orland Park approved a $43,200 yearly contract for Google Gemini, which now answers 97 percent of staff questions, while fixing issues with its Humphrey AI phone system.
- Morningstar analysts say a recession is unlikely from an AI spending slowdown, with the median forecaster putting the 12-month recession probability at 24 percent.
- A $5 million portfolio with 40 percent in AI megacaps could lose about $1.5 million in a 2000-style crash, according to Trefis.
Sophos combines performance and defense for AI-era networks
Sophos says AI-era networks need both high speed and strong security working together. The company released Sophos Firewall v23 with over 30 improvements. It also offers Sophos NDR, which uses AI to find suspicious activity on devices without agents. Sophos Workspace Protection adds secure access controls for web, SaaS, and AI tools. A new Generative AI web category helps organizations control which AI services users can access.
Fortinet and Equinix expand AI security partnership
Fortinet and Equinix expanded their partnership to secure distributed AI infrastructure. They created a validated reference architecture combining FortiAIGate with the Equinix Distributed AI Hub. The solution inspects AI interactions, enforces security policies, and tracks per-token costs for large language models. It supports data residency requirements and uses NVIDIA GPU acceleration to avoid slowing AI workloads. The partnership also covers unsanctioned AI usage through FortiGate network visibility.
Fortinet and Equinix launch distributed AI security hub
Fortinet and Equinix launched FortiAIGate on the Equinix Distributed AI Hub to secure AI at scale. The hub connects AI agents, models, data, and cloud providers through private interconnection. FortiAIGate inspects AI traffic and applies centralized security policies across sanctioned LLM and MCP traffic. Organizations can track per-token costs, set budgets, and keep data within customer infrastructure. The architecture uses NVIDIA GPUs for security inspection with zero performance impact on AI workloads.
AI stock crash hits Wall Street hard
A major decline in AI stocks sent Wall Street into a sell-off. The S&P 500 fell more than 3 percent during the trading session. The drop wiped out gains from the previous day. The article reports on the sharp market downturn driven by declining AI-related stock values.
Nvidia and AI chip stocks fall again on Wall Street
Nvidia and other AI chip stocks experienced further declines, erasing Wall Street's rally from the day before. The continued drop in technology and chip stocks followed the previous session's losses. The article covers the second day of falling prices for major AI-related companies on the stock market.
Companies increase AI security spending but lose control over shadow AI
Companies are raising cybersecurity budgets, with 74.5% reporting larger spending than a year ago. 40.3% said the next security dollar would go to AI security and governance. Many organizations lack proper oversight, as 61.5% cannot confirm all security tools are configured correctly. DNSFilter found that employees connect to far more AI tools than IT teams report.
Huawei Tencent Alibaba back AI agent safety pact in China
Huawei, Tencent and Alibaba joined 20 Chinese companies and institutions in supporting a voluntary AI agent safety pact. The agreement was launched at the World Internet Conference in Wuzhen. It aims to make AI agents from different providers work together while setting clear rules on authorization and responsibility. The pact also focuses on keeping users in control of key decisions.
Excelsoft launches four AI products from 38 ideas
Excelsoft launched four products from a pipeline of 38 AI-first ideas, 16 prototypes and nine customer pilots. The launched products are AI-levate App Suite, Saras SchoolAI, Saras Assessments in-a-Box and CareerSPARC. Saras SchoolAI offers teaching tools for lesson planning and quizzes at Excel Public School. Saras Assessments in-a-Box supports secure exam delivery for over 100 candidates per box. The AI stack runs on NVIDIA A6000 GPUs and uses models including Mistral and GPT-4o.
Orland Park fixes AI phone system and adopts Google Gemini
Orland Park is fixing its AI-powered phone system called Humphrey after residents struggled to reach village staff. Residents Karen Drake and Pam Stepansky reported long waits and failed connections. The village also approved a $43,200 yearly contract with Zion Cloud Solutions to use Google's Gemini AI for data storage. Gemini now answers 97% of staff questions and helps the village save money compared to its old system.
Amazon launches Web Answer tool on Bedrock AgentCore
Amazon introduced Web Answer on Amazon Bedrock AgentCore, a tool that returns cited answers to natural-language questions in one step. It combines search, retrieval, synthesis and citation into a single request. Web Answer runs current web searches for every question and uses responsible AI safeguards. It supports enterprise controls including gateway authentication, VPC endpoints and per-account rate limits of 10 transactions per second in US East.
Certinia launches Veda AI agents for professional services
Certinia introduced Veda AI agents to help professional services firms with sales, staffing, project delivery and billing. The Salesforce-native platform now includes 24 package agents, with plans to exceed 40 by year end. The July acquisition of Moonnox added context graphs and project memory that connect emails, chats and calls to structured service data. Certinia aims to reduce administrative work so firms can deliver projects faster and at lower cost.
Could an AI spending slowdown trigger a recession
Investors worry that a drop in artificial intelligence spending could hurt the economy, but Morningstar analysts say a recession is unlikely. AI has driven nearly half of recent GDP growth from tech investment. The Federal Reserve still has room to cut interest rates to offset weaker demand. The median economic forecaster puts the chance of a recession in the next 12 months at 24 percent.
AI-heavy portfolios face up to $1.5 million risk
A $5 million portfolio with 40 percent in AI megacaps could lose about $1.5 million in a 2000 style crash, according to Trefis. A 2022 style drop would cut roughly $1.48 million from the same portfolio. Index funds and stock compensation have pushed companies like Nvidia, Alphabet and Microsoft into large shares of major indexes. High capital gains taxes make trimming concentrated positions difficult for many investors.
Sources
- Securing the AI-Era Network: Why Performance and Defense Belong Together
- Fortinet and Equinix Expand Partnership to Secure Distributed AI Infrastructure
- Fortinet and Equinix Secure Distributed AI at Scale
- A wipeout for AI stocks sends Wall Street reeling, and the S&P 500 tumbles more than 3%
- More drops for Nvidia and other AI chip stocks erase Wall Street's rally from the day before
- Companies are spending more on AI governance and getting less control
- Huawei, Tencent back AI agent pact on safety, user control | MLex | Specialist news and analysis on legal risk and regulation
- Excelsoft has launched four products from 38 AI-first ideas
- Orland Park working to fix AI-powered call system, embraces new platform
- Getting started with Web Answer on Amazon Bedrock AgentCore
- Professional services reinvented: Certinia’s Veda AI agents
- Could an AI Spending Crunch Cause a Recession?
- The $1.5 Million Risk In AI-Heavy Portfolios
Comments
Please log in to post a comment.