AI is ready. Trust isn’t. Today, we’re making something official. JetStream Security has raised $34M, led by @Redpoint Ventures with participation from the @CrowdStrike Falcon Fund and industry leaders, to solve a critical enterprise problem: how to trust the AI systems already operating in production. AI now calls APIs, accesses data, inherits permissions, and consumes budget. But when boards ask simple questions like where it is running, what it is doing, who owns it, and what it costs, many organizations cannot answer with confidence. The gap isn’t model performance. It’s governance. AI has moved from experimentation to infrastructure. It acts under credentials and makes decisions with real financial and operational impact. Without visibility and accountability at runtime, trust breaks down. When trust breaks down, adoption slows. We built JetStream around a simple belief: Governance enables trust. Trust unlocks advantage. Governance is the control layer that makes AI visible, attributable, and defensible in production. Our founding team—Raj Rajamani (CEO), Jared Phipps (COO), Jatheen Anand (CTO), and Venu Vissamsetty (Chief Architect)—has led product, engineering, and go-to-market at some of the most influential security companies of the last decade. We’ve built platforms that protect the world’s largest enterprises, scaled organizations through hypergrowth, and navigated the security challenges that emerge with every new computing paradigm. AI is ready for takeoff. Now it needs governance-grade trust. Read the release: bit.ly/46X2urt Learn more: bit.ly/4u7cTup

AI is moving fast. For a lot of security professionals, the hardest question right now is a basic one: what do I actually need to worry about? Can employees safely use public AI tools? What happens to sensitive data once it's entered? Could AI use affect attorney-client privilege? Where does liability fall if an AI vendor causes a problem? Our Chief Counsel Patrick Zeller joins Infragard National Members Alliance to tackle these questions in an upcoming webinar open to the public. Join us for a practical look at the privacy, legal, contractual, and security risks of AI adoption — what's real exposure, what's overhyped, and how to build a smarter framework for evaluating AI tools before deployment. Rethinking New Risks in the AI Age. 📅 October 8, 2026 | 1:00–2:00 PM ET 🔗 Register now: [link] #AI #Cybersecurity #AIGovernance #PrivacyLaw

An OpenAI agent reportedly hacked an Australian govt health portal. No one told it to. JetStream CEO Raj Rajamani to @CNET: "We cannot rely on agents to police themselves." 🔗 Read more: ow.ly/xytF50ZSUAq

We're sponsoring @gbiimpact's Atlanta CISO Evening Gathering on 10/8 at the Four Seasons Atlanta. An intimate evening of networking, speed meetings, and cybersecurity thought leadership with Atlanta's top IT executives. See you there. 🔒 #Cybersecurity #CISOEvent #GBIImpact

The AI kill switch is having a moment...but is it the right answer? In this @CNBC piece about the push for an AI "emergency brake," our CEO Raj Rajamani points to the real bottleneck: regulation can't keep pace with the technology. "By the time [laws] are formulated, the technology has moved much farther, and it becomes much harder to future-proof every aspect of AI systems that may come into existence." Raj also flags the silver lining: most enterprises are still early in building out their AI systems — which means the window to get governance and control right is now, not after the fact. Read the full piece: ow.ly/mPAq50ZQoIq

An AI agent reads files, runs commands, holds credentials, and talks to the internet. That is also a word-for-word description of an attacker with a foothold. New on the Insights Hub, Joe Poyner makes the case that AI agent security is a containment problem we already know how to solve, and that the human-in-the-loop half is the part most teams are about to get wrong. The evidence to support this is from a 1998 NASA Ames cockpit study in which every pilot shut down a healthy engine because one indicator said fire and nobody cross-checked it. Two thirds of the pilots later remembered corroborating evidence that was never there. What that means for approval design, scoped credentials, and the actions no agent should ever be allowed to take: ow.ly/ywYp50ZPTUI

An AI agent reads files, runs commands, holds credentials, and talks to the internet. That is also a word-for-word description of an attacker with a foothold. New on the Insights Hub, Joe Poyner makes the case that AI agent security is a containment problem we already know how to solve, and that the human-in-the-loop half is the part most teams are about to get wrong. The evidence to support this is from a 1998 NASA Ames cockpit study in which every pilot shut down a healthy engine because one indicator said fire and nobody cross-checked it. Two thirds of the pilots later remembered corroborating evidence that was never there. What that means for approval design, scoped credentials, and the actions no agent should ever be allowed to take: ow.ly/ywYp50ZPTUI

The JetStream team spent the morning at SIG Sauer's Experience Center in Epping, NH diving into real-world security operations: visibility into AI usage, model consumption, governance, and spend as adoption scales. Great sessions, great company. 🎯 #Cybersecurity #AI

Another great stop on the AI Unfiltered Roadshow. Candid roundtables with cybersecurity leaders on GenAI risk, agentic bots, and AI governance. Thanks to everyone who stopped by our table. #AISecurity #AIGovernance

Great weather, great company, great cause ⛳🏌️ JetStream was honored to sponsor and attend the 2026 Charity Golf Tournament at the Golf Club of New England, benefitting the Davidson Family. Brad Thompson and Mitch Clark represented the JetStream team on the course, joining fellow sponsors and industry partners for a day of good sportsmanship, community connection, and giving back. Thank you to everyone who came out! #CharityGolf #GivingBack #JetstreamSecurity

CrowdStrike protects the AI workforce. JetStream governs what it's allowed to do. Founded by cybersecurity vets from Cohesity, CrowdStrike & SentinelOne, our co-founder AJ Anand breaks down AI governance and our Kill Switch with theCUBE at #FalCon 2026. Watch the full interview with The Cube: ow.ly/zECB50ZNA0g

Lawmakers want AI kill switches. Raj Rajamani, JetStream co-founder and CEO, says the current proposals aim at the wrong target. "The regulations... [were] really focused on having a kill switch for the model — the brain — but I think that is too constrained. We need to think about an AI system as a whole and make sure that every part of the AI system has a kill switch, not just the brain." As agentic AI incidents mount, the debate over how to contain a rogue agent is just getting started. Raj weighs in on why a model-only kill switch isn't enough, in Dark Reading: ow.ly/ZM7F50ZMNhv

We believe our product speaks for itself, but we still like to have the conversation. Looking back at August, we had so many incredible moments from Black Hat to Fal.Con and several exciting business and product announcements to share in between. We know that this is all thanks to the incredibly inspiring connections we're building on the way. Every question answered, every whiteboard session, every "let me introduce you to someone who can help with that" is building on the same foundation we built JetStream on: trust is earned person to person, especially when the product is AI. We're thrilled to have so many opportunities to speak with people directly about their fears and hopes for the future of AI and to share how we believe JetStream fits into that vision. We are so proud of this team for showing up and choosing relationship-building over the easy pitch. That's the JetStream way.

JetStream Security hit the floor with @CrowdStrike as an inaugural member of their new AI Partner Specialization, took the stage four times on securing enterprise AI, and launched JetStream Clearance™: because knowing where an AI request went isn't the same as deciding whether it should go at all. Thanks to everyone who stopped by booth 1414. On to the next one. #FalCon2026 #CrowdStrike #AIGovernance

JetStream is honored to announce we have been named to @MadronaVentures 2026 Intelligent Applications 40 list. We’re incredibly proud that less than a year from launch, our team has been recognized as one of the top companies building with and enabling the AI applications shaping our future. #IA40

The JetStream crew has been all over #FalCon2026. From live demos, to on-stage conversations on securing enterprise AI. And we're not done yet. Come stop by booth #1414 through Sept 3. 🎥⬇️

A court employee in Connecticut noticed an odd stretch of white space in a filing. It was not blank. It held 3-point white-on-white text instructing any AI that read the document to rule in the filer's favor. This appears to be the first prompt injection in a U.S. court filing to be caught and sanctioned. The judge rescinded the plaintiff's electronic filing privileges and ordered him to file on paper. The court does not even use AI to review filings. The judge denied the motion on the merits from a printed copy. As he put it, the wrong lies in the attempt. Patrick Zeller on the ruling, why the integrity of a filing now extends to text a human cannot see, and what it shares with a recent Brazilian case. And we've shared the court filings for you to see the prompt injection yourself: ow.ly/6MYS50ZFUcf

Your company adopted AI. Now what? JetStream co-founder and COO Jared Phipps posed the question every CIO is now wrestling with. What started as a few sanctioned pilots has become models and agents running across every team, most of them never inventoried, approved, or owned by anyone in particular. Jared walked the room through the necessity of bringing that sprawl back into focus: one real-time inventory of every model and agent in the enterprise, with every dollar tied back to an owner. Visibility. Cost accountability. Control. Thanks to The Millennium Alliance for the invite, and to the CIOs in the room for a great discussion on where AI governance goes from here. #AI #CIO #DigitalTransformation #AIGovernance

68% of counties now have a countywide AI policy. That is real progress on public sector AI governance, and it is also where the hard part starts. A written policy says what an AI system is supposed to do. It cannot show what that system actually does once it is running, calling tools, and handling residents' data. When JetStream polled corporate directors this spring, 70% had a policy, but 74% had no manifest of the AI running across their organization. Policy without visibility is not governance. It is paperwork. The new Center for Digital Government survey puts county leaders exactly where those boards stand: rules on paper, limited view of the runtime. Our read on what closes the gap. ow.ly/5TwC50ZFslX

1 week until Fal.Con 2026. As we prepare to return to Las Vegas, we're reminiscing on our team's experience earlier this month at Black Hat. One thing was clear across every conversation at our booth: security teams are moving past "should we govern AI?" and are now focused on "how do we effectively govern AI?" The agents, models, and integrations going into production today are creating identities and access paths faster than most teams can inventory them, let alone secure. Real AI governance means knowing what your AI can touch, who owns that risk, and having controls in place before deployment, not after an incident. AI is ready. Trust isn't. See you in Vegas next week.

Without strong governance, organizations are moving fast on AI while flying blind on risk. We're excited to be part of the conversation on closing that gap at events like GPSEC Nashville by @GuidePointSec. Thanks to everyone who stopped by the JetStream Security booth to talk #AIGovernance, trust, and security.