AI coding tools are rapidly expanding, with spending projected to exceed $13 billion this year. These agents now possess human-like capabilities to modify files and execute commands, yet many organizations struggle to track their usage. Research indicates over 220,000 unmanaged software installations and 135,000 exposed systems appeared in a single week, with machines outnumbering human identities by a ratio of 109 to 14.
Palo Alto Networks addresses these gaps with Prisma AIRS, a platform designed to assign unique identities to AI agents and enforce strict data access rules. The solution helps companies manage costs, as 93% of organizations currently exceed their AI budgets. This tool allows security teams to monitor what agents actually do rather than just what they are permitted to access.
Security risks remain high, with a typical AI agent accessing 55 datastores compared to three or four for a regular employee. Alarmingly, 79% of these agents can reach stored secrets like API keys. To combat unreliable AI citations, experts suggest linking every claim directly to proof, including specific dates and names, to help search engines trust sources more effectively.
Investors are also watching companies facilitating workforce reskilling. ManpowerGroup reported $4.6 billion in revenue in France and $2.7 billion in the US, while Asure Software generated $155 million. Coursera earned $550 million from consumers and $336 million from enterprises as demand for new skills grows.
Hardware costs are another focus, with seven cheaper alternatives to the expensive DGX Spark now available. Options include the $2,600 Strix Halo box and the $3,000 OEM GB10 system. Meanwhile, Meta shares rose over 6% after launching Muse, a personal AI agent capable of booking travel and filling forms, signaling investor interest in tangible AI applications.
Key Takeaways
- AI coding spending is expected to exceed $13 billion this year.
- Over 220,000 unmanaged software installations were detected in one week.
- Machines now outnumber human identities by a ratio of 109 to 14.
- Palo Alto Networks launched Prisma AIRS to monitor and secure AI agents.
- 93% of companies are exceeding their AI budgets.
- A typical AI agent accesses 55 datastores versus three or four for employees.
- 79% of AI agents can reach stored secrets like API keys.
- ManpowerGroup generated $4.6 billion in revenue in France and $2.7 billion in the US.
- Asure Software reported $155 million in revenue from cloud-based HR tools.
- Meta stock rose more than 6% following the launch of its Muse AI agent.
Palo Alto Networks Secures AI Coding with New Platform
AI coding tools are growing fast, with spending expected to exceed $13 billion this year. These tools can now act like humans, changing files and running commands, which creates new security risks. Many companies lack visibility into these tools, and some employees use them without approval. Palo Alto Networks offers a platform to discover hidden AI tools, check their safety, and control what data they access. This solution helps companies manage costs and protect sensitive information while developers use AI.
Palo Alto Networks Helps Companies Secure AI Coding Tools
AI coding agents now have the same access as human engineers, which challenges old security systems. Research shows over 220,000 unmanaged software installations and 135,000 exposed systems in just one week. Machines now outnumber human identities by a ratio of 109 to 14, making it hard to track who is doing what. Companies also waste money, with 93% exceeding their AI budgets. Palo Alto Networks provides a platform called Prisma AIRS to monitor these agents, assign them unique identities, and enforce strict rules on data and spending.
Webinar Explains What AI Agents Access and How They Use It
AI agents are accessing enterprise data at a scale that traditional security controls cannot handle. Research from Bedrock Data shows a typical AI agent can access 55 datastores, while a regular employee accesses only three or four. Even worse, 79% of these agents can reach stored secrets like API keys. Security teams need to see not just what agents are allowed to access, but what they actually do. An upcoming webinar with Bedrock Data will teach leaders how to identify risks and regain control over agent behavior.
Three US Stocks Investors Watch for AI Reskilling Trends
AI is changing which jobs exist and which skills matter, creating demand for workforce training. Investors are watching three US companies that help businesses reskill their employees. ManpowerGroup provides global staffing and training services with revenue of $4.6 billion in France and $2.7 billion in the US. Asure Software offers cloud-based HR tools for smaller businesses, generating $155 million in revenue. Coursera runs an online learning platform for workers and companies, earning $550 million from consumers and $336 million from enterprises.
Seven Cheaper Alternatives to the Expensive DGX Spark
The DGX Spark is a powerful AI computer, but its high price makes it hard for many organizations to buy. There are now seven cheaper options available to run AI locally. The Strix Halo box costs about $2,600 and is a compact all-in-one system. The OEM GB10 system costs around $3,000 and allows users to choose their own parts. A 48GB card is the most affordable option at roughly $1,500, though it may not match the full performance of the DGX Spark.
Experts say fix AI search by linking claims to proof
An SEO expert explains that websites must change how they present information to work better with AI search engines. The article notes that 62% of AI citations are currently unreliable because systems struggle to find clear answers. To fix this, companies should connect every claim directly to its evidence, including specific dates, names, and context. Instead of vague statements, pages should clearly state who made a claim, what the results were, and when they happened. This approach helps AI systems understand the information and trust the source more effectively.
Spekit launches new AI deal rooms for sales teams
Spekit has released new AI-powered deal rooms that are available for use today. These tools help revenue teams manage buyer content, presentations, and sales pipelines in one place. The platform uses AI to suggest follow-up actions after calls and integrates directly with Slack to track conversations. It also provides real-time insights into the sales pipeline to help teams close more deals. Revenue teams can start using these features immediately to gain an edge in every customer interaction.
USC gets more funding to lead national AI research
The U.S. National Science Foundation awarded $20 million to renew the USC Artificial Intelligence Institute for Advances in Optimization through 2031. This project, led by Georgia Tech with partners USC and UC Berkeley, aims to use AI to improve power grids and manufacturing. USC will receive $5.6 million and expand its research team from two to five members. The researchers will also bring a summer camp to high schoolers in Los Angeles to teach data science skills. This funding helps move AI research from basic science to real-world applications that solve complex problems.
Cymphony raises $25 million for AI agent security
Cymphony has raised $25 million to build security tools for AI agents. The company, backed by Sequoia, treats AI agents like employees that need their own identity layer. This approach is part of a growing trend of startups focusing on securing the workforce of AI agents. The funding helps Cymphony continue developing its unique security stack for this new technology.
Meta stock rises after Muse AI agent launch
Meta shares rose more than 6% after the company introduced Muse, a new personal AI agent. Muse can perform tasks like booking travel and filling out forms, which investors see as a way to make money from AI investments. Other stocks moved differently, with ServiceTitan dropping 30% due to missed expectations and Navan facing criticism for rising expenses. Investors are now demanding clear proof that AI projects will lead to better future profits, not just current excitement. Cloudflare is also being viewed as a potential cybersecurity opportunity related to AI.
Sources
- Secure AI Coding: Governing every agent, artifact, and identity
- Secure AI Coding
- Webinar | What Your AI Agents Can Access, and What They Actually Do With It
- 3 US Reskilling Stocks Retail Investors Are Watching As AI Reshapes Hiring
- DGX Spark Alternatives in 2026: 7 Cheaper Ways to Run AI Locally
- How to fix your AI search problem
- Spekit Reimagines AI-Powered Deal Rooms to Give Revenue Teams an Edge in Every Deal
- USC’s role expanded in next phase of national AI research project - Beverly Press & Park Labrea News
- Cymphony Raises $25M as the Agent Identity Security Stack Keeps Forming
- Stock movers today: Meta leads AI optimism while ServiceTitan and Navan show the cost of missing expectations
Comments
Please log in to post a comment.