The Hacker News · Jul 8
CISA orders patching of actively exploited Langflow AI-platform flaws
CISA added actively exploited Langflow vulnerabilities to its Known Exploited Vulnerabilities catalog after a June campaign chained two flaws to steal LLM provider keys and cloud credentials from self-hosted AI agent builders. Federal agencies must remediate by July 10.
▸ The MSP Angle
Are self-hosted AI agent builders safe to run?
Attackers have figured out that self-hosted AI tooling is where the keys live: LLM credentials, cloud tokens, client data. If any client is experimenting with self-hosted agent builders, treat them as production attack surface today. This is the strongest recent argument for running client AI in a managed, isolated platform instead of a hobby deployment.
Read at The Hacker News ↗
Microsoft Partner Center · Jul 1
Microsoft 365 Copilot SKUs go GA for SMB with new CSP promotions
As of July 1, Microsoft 365 Business Standard with Copilot and Business Premium with Copilot are permanent SKUs for organizations up to 300 seats, and new FY27 CSP promotions discount Copilot bundles through December. The same cycle adds CSP growth margins for AI workloads and repackages most commercial Microsoft 365 plans.
▸ The MSP Angle
What do the new Microsoft 365 Copilot SKUs mean for MSPs?
Every Microsoft renewal you touch this quarter will surface these SKUs, so know them cold. But be clear-eyed about what reselling Copilot seats makes you: the billing middleman, while Microsoft becomes the AI provider of record for your clients. The durable margin is in governance and management, and in the bigger question every MSP should sit with this quarter: whose brand is on the AI your clients use every day?
Read at Microsoft Partner Center ↗
Thomson Reuters · Jun 22
74 percent of professionals use AI weekly, but only 6 percent of firms deliver what clients want
Thomson Reuters surveyed 1,816 legal, tax, accounting, and risk professionals across 62 countries: 74 percent now use AI weekly, 91 percent say their firms fall short of AI's potential, and a third admit to using unsanctioned AI tools. The report puts roughly 143 billion dollars of US legal and accounting revenue under active client reconsideration within 12 months.
▸ The MSP Angle
Do professional services firms actually want AI help from their MSP?
Law and accounting firms are classic MSP clients, and they are telling surveys they want AI help their firms cannot build internally. The one-in-three shadow AI number is your door-opener: every unsanctioned tool is a data-governance incident waiting for an invoice. Show up with an implementation plan before their clients walk.
Read at Thomson Reuters ↗
Anthropic · Jun 30
Anthropic ships Claude Sonnet 5 with introductory pricing until August 31
Anthropic launched Claude Sonnet 5 as its new default model, claiming performance near Opus 4.8 at mid-tier cost, with cyber safeguards on by default and introductory API pricing that steps up after August 31.
▸ The MSP Angle
How should MSPs price projects around introductory model rates?
A near-flagship model at mid-tier cost is the sweet spot for most client workloads: summarization, drafting, ticket triage. The date that matters is August 31, when the price steps up. Anything you quote on introductory pricing needs to survive the increase.
Read at Anthropic ↗
Fox Business · Jul 7
Gallup: workers who skip AI face triple the layoff risk
New Gallup analysis finds US tech workers who use AI at least monthly face a roughly 6 percent predicted layoff probability versus 18 percent for infrequent users, and 62 percent of all laid-off workers were AI non-users. Only 1 percent of laid-off workers named AI as the primary cause of their layoff.
▸ The MSP Angle
Does using AI at work really lower layoff risk?
AI fluency is now job insurance, and the data finally proves it. Train your own technicians first, then package the same enablement for your clients' workforces. AI training is a service line with built-in urgency: nobody wants to be in the 18 percent column.
Read at Fox Business ↗
Council of the EU · Jun 29
EU locks in revised AI Act timeline; first obligations land next month
The EU Council gave final approval to the AI Act simplification package, delaying high-risk system rules to late 2027 and 2028. Transparency obligations and enforcement powers over general-purpose AI providers still take effect August 2, 2026.
▸ The MSP Angle
When do EU AI Act obligations actually start?
The headlines say delay; the fine print says next month. Transparency obligations arrive August 2 regardless, and clients selling into the EU will ask their MSP what that means for them. Being the partner with a straight answer is cheap differentiation.
Read at Council of the EU ↗
Fortune · Jun 26
One in three employers already replacing entry-level roles with AI
GMAC's survey of more than 600 recruiters, over half hiring for Fortune 100 and 500 companies, found a third of employers already replacing entry-level positions with AI. Technology is the most exposed sector at 40 percent, led by automation of routine coding, data processing, and customer service.
▸ The MSP Angle
Is AI really replacing entry-level jobs?
The work being automated first is tier-1 helpdesk work: exactly the MSP's own cost center. That cuts both ways. Your service desk economics improve if you automate before your competitors, and your clients will expect the savings to show up in pricing either way. Reprice the service desk on your schedule, not theirs.
Read at Fortune ↗
SiliconANGLE · Jul 7
CyberProof launches agentic MXDR claiming two-thirds of investigations run autonomously
CyberProof's new managed detection and response service pairs AI agents with human analysts and claims autonomous handling of up to two-thirds of SOC investigations, with agents scored on effectiveness, speed, and cost.
▸ The MSP Angle
Can AI agents really run most SOC investigations?
Productized agentic security services are now shipping with hard automation claims attached. Your clients will read numbers like these and ask why their MDR costs what it costs. Whether or not the claims hold, agent-assisted service delivery is becoming the reference point MSP pricing gets measured against.
Read at SiliconANGLE ↗
TechCrunch · Jul 2
Microsoft commits 2.5 billion dollars to a hands-on AI deployment company
Microsoft stood up Frontier Company, a 2.5 billion dollar operating unit with 6,000 engineers dedicated to hands-on enterprise AI deployments, days after Amazon committed a billion dollars to forward-deployed engineering. Early partners include LSEG, Unilever, and Land O'Lakes.
▸ The MSP Angle
Why is Microsoft spending billions on AI deployment services?
When hyperscalers build 6,000-person implementation armies, they are telling you where the money is: deployment services, not licenses. They are also telling you who they will serve, and it is not the 80-seat manufacturer or the regional law firm. That last mile belongs to MSPs, and it is the most defensible ground in the whole AI economy.
Read at TechCrunch ↗
Reuters via Yahoo Finance · Jul 7
Nvidia dips on report DeepSeek is building its own AI chip
Nvidia shares slipped after Reuters reported DeepSeek is designing an in-house AI chip to reduce reliance on Nvidia and Huawei hardware, the same week Nvidia denied that its next-generation platform had slipped to 2028.
▸ The MSP Angle
What does more AI chip competition mean for infrastructure costs?
GPU supply and pricing flow straight into the cost of the cloud AI infrastructure you resell. More chip competition generally means cheaper inference over time, which is good news for anyone building AI services on top of someone else's silicon.
Read at Reuters via Yahoo Finance ↗
Microsoft Partner Center · Jul 8
Microsoft opens Cloud Accelerate Factory to more Azure and AI partners
Microsoft expanded its no-cost, Microsoft-led deployment assistance across more than 30 Azure services to partners holding AI and automation specializations, and the benefit now stacks with existing partner incentives.
▸ The MSP Angle
What is Microsoft's Cloud Accelerate Factory and who qualifies?
Free Microsoft engineering on standardized deployment work is effectively found capacity. If you hold a qualifying Azure specialization, route the commodity deployment tasks to the Factory and keep your own engineers on the work clients actually pay a premium for.
Read at Microsoft Partner Center ↗
Forbes · Jul 2