Steve Miller's Blog

A digital workshop for systems-minded tech commentary.

Author: steve

  • Your AI Agents Are Holding the Keys: Why Their Credentials Are Becoming a Critical Security Boundary

    Your AI Agents Are Holding the Keys: Why Their Credentials Are Becoming a Critical Security Boundary

    The fastest way to give an AI agent access to something is to paste an API key into its config. That is how a lot of pilots started, including some I have watched up close. It works right up until the agent gets prompt-injected, logs the key, or sends it along to a model provider with the rest of its context. Managing AI agent secrets and credentials is moving from a niche concern into product announcements from both a fast-growing startup and IBM. If your team runs agents in production, the shortcut you took last quarter is worth revisiting this week.

    Two Vendor Moves in the Same Week

    On October 27, 2026, two very different companies made agent credentials the center of their announcements.

    Infisical, the open-source secrets management company, announced a $40 million Series B at a $600 million valuation, led by Elad Gil. The release describes its Agent Vault product this way: it “lets companies give AI agents access without handing over credentials.” CEO Vlad Matsiiako put the logic plainly: “Agents need access to company systems to do their jobs, but giving them actual passwords and API keys is a major security risk.” In his own post on the round, Matsiiako wrote that Infisical will “significantly increase our focus on Agent Vault.” He also wrote that the company “didn’t have to raise this round” because it has been cash-flow positive. Infisical says it serves more than 1,300 customers.

    IBM, at its TechXchange 2026 conference, announced that HashiCorp Vault “will now extend those capabilities to AI agents, issuing secured, short-lived credentials that limit each agent to what it needs.” IBM paired that with Agent Identity, which it describes as establishing “the authoritative record for each agent, such as who owns it, its lifecycle status and the authority it has been granted.” IBM says HashiCorp Vault 2.2 and Agent Identity are generally available. It also says Vault 2.2 “gives agents their own secured accounts” and automatically renews their credentials.

    Is money “pouring into” this space? I can verify one funding round here, Infisical’s, plus a large incumbent repositioning its products. That is a meaningful signal. It is not proof of a market-wide flood, and I am not going to claim one.

    Why Credentials Are the Pressure Point

    IBM’s release makes the argument better than I could: “As agents invoke tools and act across business systems, their credentials become a critical security boundary.” It continues: “An agent with a long-lived, overprivileged credential can put systems and data at risk if it’s compromised or acts outside its intended task.”

    Infisical’s release frames it more bluntly. Agents need credentials like API keys, passwords, and authentication tokens, “which AI agents cannot be trusted with.” The company says Agent Vault is built so an agent “cannot leak credentials in a cyberattack or inadvertently send them to LLM providers or other tools it communicates with.”

    Here is my opinion, and I am labeling it as one. For many teams, secrets management is now the front line of agent security. That is because it is the control you can actually enforce today, while prompt-injection defenses are still immature. You cannot fully stop an agent from being tricked. You can limit what a tricked agent is holding.

    AI Agent Secrets Management: Two Architectures, One Principle

    The two announcements take somewhat different routes to the same idea:

    • Brokered access (Infisical’s description): the agent never holds the real secret. Infisical says it lets agents reach approved systems without exposing the underlying credentials.
    • Short-lived, identity-bound credentials (IBM’s description): each agent gets a governed identity and credentials that are short-lived and least-privilege. IBM says access can be “denied and session revoked in real time” when an agent tries something outside its approved permissions.

    Both reject the pattern most pilots started with: one long-lived, broadly scoped static key, sitting in a config file or environment variable, with no owner and no expiry.

    What I Would Do on Monday

    These are my recommendations for platform and security teams, not vendor guidance.

    1. Find the keys you already handed out. Scan agent configs, prompt templates, notebooks, CI variables, and repos for hardcoded secrets. IBM’s release mentions customers using HashiCorp Vault Radar to detect hardcoded and exposed secrets. Whatever tool you use, inventory first.
    2. Give every agent an owner and an identity. No shared service accounts across agents. If you cannot name who owns an agent, you cannot decide when to revoke its access.
    3. Replace static keys with short-lived, scoped credentials. Issue them per task or per session, scope them to the specific systems and actions the agent needs, and let them expire on their own.
    4. Keep secrets out of the model’s context where you can. A proxy or broker pattern means a prompt-injected agent has nothing reusable to exfiltrate. Remember that it can still leak the data an approved call returns.
    5. Log every credentialed call. Audit trails should tie each action back to an agent identity and a human owner.
    6. Rehearse revocation. Practice killing a misbehaving agent’s access in minutes, not hours.

    Try this hypothetical. A coding agent with a long-lived cloud key reads a poisoned README that tells it to print its environment variables into a pull request comment. With a static key, you are rotating credentials and reviewing logs all weekend. With a brokered or short-lived credential, the leaked value is useless or already expired. That difference is the whole argument.

    The Bottom Line

    Infisical’s Series B and IBM’s Vault and Agent Identity launches both point at the same weak spot: agents holding credentials they should not keep. The tooling is arriving. The work that falls to your team is finding the keys already pasted into configs, then moving every agent to owned, scoped, short-lived, audited access.

  • CISA 2030 Promises an Aggressive, Anticipatory Stance to Eliminate Adversarial Footholds in Critical Infrastructure: What the Strategy Actually Asks of the Operators Who Run It

    CISA 2030 Promises an Aggressive, Anticipatory Stance to Eliminate Adversarial Footholds in Critical Infrastructure: What the Strategy Actually Asks of the Operators Who Run It

    CISA released its new strategy, CISA 2030: Defend, Harden, Prevail, on October 26, 2026. The headline promise is bold. In his welcome letter, Acting Director Nick Andersen writes that CISA “will transition from passive defense to an aggressive, anticipatory stance that actively eliminates adversarial footholds in defense-critical infrastructure prior to weaponization against our systems.” If you run a water plant, a regional utility, a hospital network, or any other operational environment and you are reading up on the CISA strategic plan, critical infrastructure, and proactive defense, the question that matters is practical. What does this plan actually ask of the operators? I read the whole document. The honest answer is: less than the rhetoric suggests, at least so far.

    What the Strategy Says

    The document is short, about eight pages of substance. It is organized around three pillars, each named in the plan:

    • Defend: Lead National Cyber Defense
    • Harden: Prioritize National Infrastructure Security
    • Prevail: Mobilize National Resilience

    Andersen frames the premise realistically. He writes that “there is no guarantee that a strong defense will eliminate intrusions,” and that “we must ensure that when adversaries gain a foothold, we are prepared to detect them quickly, evict them decisively, and keep critical operations running without interruption.” CISA’s press release describes the approach as “decisive, proactive action to identify, disrupt, and eliminate adversarial access.” Those are CISA’s descriptions of its intent. They are not results, and the strategy does not claim they are.

    The most consequential line, in my view, sits under the Harden pillar: “We will deprioritize threat-agnostic, all-hazards approaches and embrace risk-informed, consequence-based analysis.” CISA says it will focus attention, investment, and operational effort “on the infrastructure whose disruption would have the greatest impact on public health and safety, continuity of the economy, and national security.” CISA also says it will support the White House in developing a National Risk Register to quantify risk to national infrastructure.

    What the Strategy Commits CISA to Do

    Almost every key objective is written as something CISA will do, often “alongside” or “with” partners. Among them:

    • Integrate cyber defense capabilities across government, industry, state, local, tribal, and territorial partners, and international organizations.
    • “Work alongside critical infrastructure owners and operators to modernize OT environments, improve asset visibility and anomaly detection, and assure OT supply chain integrity.”
    • Coordinate with partners on data-driven software security guidance and supply chain transparency.
    • Mature an “intelligence-to-operations pipeline” with the Intelligence Community, industry, and agencies, including real-time decision making and alerts.
    • Build capacity through CISA’s ten regional offices and modernize planning and exercise capabilities.
    • Expand priority telecommunications services, including 5G/6G and commercial satellite constellations.

    The timeline is also about CISA, not operators. The plan says that within 60 days CISA will “begin a programmatic review” of its core services and capabilities. Within 180 days it will deliver “an approach for implementation and measurement of outcomes.” Within 365 days it will deliver a progress update. As Nextgov/FCW points out, the 60-day deadline applies to starting the review, not finishing it.

    What the CISA Strategic Plan Asks of Critical Infrastructure Operators

    This is where I want to be careful, because it would be easy to read new mandates into this document. I did not find any. The strategy sets no new reporting requirements, deadlines, or technical controls for owners and operators. Nextgov/FCW also reports that it “does not establish new legal authorities.”

    What the strategy does contain are appeals. Andersen writes: “I call on everyone to do their civic duty and take every step necessary to deny our adversaries a single foothold or advantage.” In the press release he goes further: “I call on every company and organization in the critical infrastructure community—from owners and operators to vendors and technology manufacturers—to act now.” DHS Secretary Markwayne Mullin’s foreword urges unity at every level, including “owners and operators of businesses and critical infrastructure.”

    Beyond those calls to action, operators appear as partners. CISA says it will build “distributed collaboration frameworks that empower and enable SRMAs, regions, states, communities, and owners and operators to advance infrastructure security through localized action.” The specifics of what partnership will look like are deferred to the 180-day implementation approach.

    My Read and Recommendations for Operators

    These are my opinions, not CISA guidance, and none of them is a requirement in the strategy.

    1. Figure out which side of the prioritization line you are on. A consequence-based model means some sectors and some facilities get more CISA attention, and others get less. The strategy does not say which services lose resources. Do your own consequence analysis now so you can make your case when the implementation plan arrives.
    2. Invest in OT visibility before anyone asks. Asset visibility and anomaly detection are named objectives. You cannot be a useful partner in evicting an adversary from a network you cannot see.
    3. Rehearse eviction, not just detection. The strategy’s own language is “detect them quickly, evict them decisively.” Tabletop what removing a persistent intruder from your control environment would take without stopping operations.
    4. Inventory the free CISA services you depend on. The 60-day programmatic review covers CISA’s core services, and the plan does not say which ones might change. If a free scan or assessment is part of your program, note it and have a fallback in mind.
    5. Get to know your CISA regional office. Regional capacity is part of the plan. A relationship built before an incident pays off during one.
    6. Watch for the 180-day implementation approach. That document, expected roughly in late April 2027 if CISA hits its stated timeline, is where concrete expectations would likely show up.

    Here is a hypothetical worth thinking through. A mid-sized municipal water utility reads this plan and assumes it is now in line for hands-on CISA help. If the risk register and prioritization land elsewhere, that utility could end up with less direct support than before. The strategy does not say that will happen. It does make the tradeoff explicit.

    The Bottom Line

    CISA 2030 promises, in CISA’s words, an “aggressive, anticipatory stance” against adversaries in critical infrastructure. For now it is a statement of agency priorities and a timeline for CISA’s own work. It is not a rulebook for operators. Treat the calls to “act now” as a strong signal, and build your visibility, eviction playbooks, and consequence analysis before the implementation details arrive.

  • Hackers Reportedly Tried to Breach the FEC, Taking Its Filing System Offline a Week Before the Midterms: Planning for When a Public Deadline Outlives Your Platform

    Hackers Reportedly Tried to Breach the FEC, Taking Its Filing System Offline a Week Before the Midterms: Planning for When a Public Deadline Outlives Your Platform

    Late Monday, October 26, a week before the November 3 midterms, the Federal Election Commission’s electronic filing system went dark. For about two days, campaigns, super PACs, and outside groups could not submit the rapid-turnaround disclosures the law requires in the final stretch of an election. This FEC hack campaign finance filing outage deserves attention well beyond politics. It is a clean case study of what happens when a legal deadline keeps running after the platform you file on stops working.

    What We Actually Know About the FEC Hack and Campaign Finance Filing Outage

    Precision matters here, because the public record is thinner than many headlines suggest. According to Reuters reporting published by NBC News, an FEC electronic filings officer emailed several campaign compliance firms Monday night. The email said: “We have a bad actor attempting to breach FEC systems. As a result, we’ve had to temporarily shut down our web facing services.” The FEC’s own public statement was narrower: “at this time, the FEC’s electronic filing system is currently offline and our IT staff are actively working to restore service.” The agency said previously filed reports remained searchable and otherwise declined to comment.

    A status message put the start of the outage at around 10 p.m. ET Monday. CNBC reported that on Tuesday night the FEC said it had identified the issue and was implementing a fix. CNBC also noted that the FEC had not said who was behind the reported breach attempt, whether anyone gained access to its systems, or whether any data was compromised.

    By Wednesday, October 28, Reuters reported the system had been restored after being offline for nearly 41 hours “to contain a cyberattack.” Reuters also reported, citing an email, that campaign finance managers had to reset their passwords before logging back in. The FEC did not announce details on the cause, and Reuters could not establish a motive. CISA and the FBI referred questions to the FEC.

    So the fair summary is this. An FEC official described an attempted breach, and the agency took its web-facing filing services offline to contain it. As of these reports, there was no public confirmation that attackers got in or that any data was taken. I am not going to claim otherwise.

    Why the Timing Hurt

    The FEC does not administer elections. Reuters and CNBC both stress that the outage had no effect on voters’ ability to cast ballots. What it hit was transparency. CNBC explains that in the final 20 days before an election, candidate committees must report certain contributions or loans of $1,000 or more within 48 hours. Political committees and outside groups must disclose certain spending within 24 hours.

    That window is when the money moves fastest. Brendan Glavin, research director at OpenSecrets, told Reuters that super PACs and dark money groups spent an average of over $50 million per day in the week before the outage, with more than $250 million reported on October 20 alone. “Disclosure delayed is disclosure denied,” OpenSecrets executive director Hilary Braseth said in a statement. Steven Mele, CEO of compliance firm MBA Consulting Group, told Reuters: “I’ve never seen this in 26 years.”

    CNBC adds context that should worry anyone who manages public-sector IT. The six-member commission has had just two members since October 2025 and has lacked the four commissioners needed to take official action since May 2025, while career staff keep operations running.

    The Containment Call Was Probably Right

    Kyle Sandy, cybersecurity director at Logically, told Reuters that taking the system offline is the right call during an attack. “It tells me the FEC detected something they couldn’t immediately scope and chose a known outage over an unknown risk,” he said. I agree. When you cannot yet scope an intrusion attempt, shutting down web-facing services and forcing credential resets is defensible. That is especially true when the integrity of filed records is the whole point of the system.

    The hard part is that the rest of the world keeps its own clock. Filers still had legal deadlines. Watchdogs still had a public to inform. That is the gap I want IT leaders to plan for.

    My Recommendations: When a Deadline Outlives Your Platform

    These are my opinions for IT, compliance, and security teams, not FEC guidance or legal advice.

    If You Run the Platform

    • Publish an outage protocol before you need it. Filers should know in advance what counts as timely filing when your system is down, and where official updates will appear.
    • Separate intake from the public web. Design a degraded mode, such as an authenticated backup submission channel, that you can keep open while you contain an incident on the main portal.
    • Plan the credential reset. Forced resets for every filer at once are a support event. Script the communications and staff the help desk before the incident.
    • Say what you can, early. Reuters noted that the lack of information fueled conspiracy theories. A short, factual statement on scope and data status narrows the rumor space.

    If You Depend on Someone Else’s Platform

    • Map your hard deadlines to single points of failure. Any regulatory filing, payment run, or customer notice that depends on one external portal belongs on your risk register.
    • Keep evidence of attempted filing. Screenshots, timestamps, error messages, and saved submissions protect you if questions come up later.
    • Prepare filings ahead and queue them. Have reports ready to submit the minute service returns, rather than scrambling in the backlog.
    • Know your official channels. Decide in advance which agency status page or account you treat as authoritative, and do not act on rumors.

    Here is a hypothetical worth a tabletop exercise. Your quarterly regulatory filing is due at midnight, and the agency portal goes offline at 9 p.m. because of a suspected intrusion. Who decides what you do, what do you document, and who tells leadership? If you cannot answer in five minutes, you have homework.

    The Bottom Line

    Based on what has been reported, this was an attempted breach that forced a deliberate shutdown. It was not a confirmed data theft. The lesson is not unique to campaign finance. Every organization that runs or relies on a deadline-driven platform needs a plan for the hours when security wins and the system stays dark.

  • Bridgewater’s 100 Million Document Vault and Its Permission Problem: Building Enterprise AI Search That Respects Who’s Allowed to See What

    Bridgewater’s 100 Million Document Vault and Its Permission Problem: Building Enterprise AI Search That Respects Who’s Allowed to See What

    Most enterprise AI pilots work fine until the corpus gets big and the access rules get messy. Then the search quietly breaks. Bridgewater Associates has now described, in public, how that happened to its own document platform. It is the clearest case study I have seen of why permission-aware RAG enterprise AI search is a retrieval-engineering problem first and a model problem second. If your team is wiring a chatbot into SharePoint, email, or a research archive, this is the part you cannot skip.

    What Bridgewater Says It Built

    At AI Engineer NYC 2026, Bridgewater’s Sam Green, team lead for Applied AI, described the Vault as “our repository of external and internal unstructured data.” By his account it holds “more than 100 million documents,” with thousands more added every day from 135 external sources, plus internal data from inside the firm. Each document is enriched with tags like country, concept, market, and the time period it discusses (session transcript).

    Aaron Linsky, who runs product engineering at Bridgewater, listed five dimensions the team had to get right: recall, precision, filtering, entitlements, and scale. He said production traces showed “almost always there was some condition,” such as a date, a source, or a specific asset. That is where the team “ran into our first major performance issue.” He also put numbers on the two audiences. Humans generate about 100 queries an hour across roughly 300 users. Agents easily run “100-plus queries a second,” and agent swarms can reach thousands.

    Why Permissions Break Vector Search

    In a separate joint session with turbopuffer CTO Nikhil Benesch (“Can Your AI Analyst Keep a Secret?”), Linsky said that when Bridgewater first built this “without turbopuffer, we thought it worked fine.” Then, as documents piled into the Vault, “it fell over tremendously.” Benesch explained why. Internally, “every email, every meeting note, every report might have a completely different set of permissions attached to it.”

    Benesch described two classic approaches, and each one fails differently:

    • Post-filtering: grab the nearest vectors first, then throw away the ones the user cannot see. It is fast, but a user with narrow access can end up with few or no results, even when a relevant document they are allowed to read exists.
    • Pre-filtering: narrow to everything the user may see, then compare vectors. Recall is good. But his example of a user with access to 50 million of 100 million documents means a huge sequential scan, and the vector index stops helping.

    turbopuffer’s answer, which Benesch calls native filtering, combines the two. The index tracks which vector clusters contain at least one document a given user can access. It skips the clusters with none, then scans inside the nearest eligible ones. Benesch said clusters currently hold about 200 documents and that the size is still being tuned. Linsky summarized the outcome in his keynote: fast like post-filtering, “but with close to the same recall as pre-filtering.”

    About Those Headline Numbers

    You will see “450 million vectors” and “120 milliseconds” attached to this story. Those figures come from Benesch, the vendor’s CTO, not from a Bridgewater statement. In the joint talk he said that across 450 million vectors, Bridgewater’s 100 million documents “turn into 400 million vectors,” and that turbopuffer gets “120 millisecond searches” at P90 across those namespaces. Note that 400 million and 450 million appear in the same sentence. Treat both the vector count and the latency as turbopuffer’s reported results, not an independent benchmark. Benesch acknowledged that nested groups are a limitation where “things get tricky.”

    How the Permissions Are Actually Wired

    This is the part security teams should read twice. turbopuffer’s own permissions documentation says permissions “currently have to be implemented at the user-level with filters.” The product “doesn’t provide built-in mechanisms for row/document-level RBAC.” The recommended pattern is to store the user IDs or group IDs with read access on each document. At query time, you pull the caller’s identity and groups from your own auth layer and pass them in as a filter.

    The docs flag a subtle failure mode, too. An empty permission array should mean no access. To make a document public, write an explicit flag like is_public: true rather than inferring openness from an empty list. That design choice is safer, and the docs say it is faster.

    The lesson goes beyond any one vendor. The vector store enforces whatever filter you hand it. The truth about who may see what still lives in your identity system and your ingestion pipeline.

    My Recommendations for Permission-Aware RAG in Enterprise AI Search

    These are my opinions, not Bridgewater or turbopuffer guidance.

    1. Make entitlements part of ingestion. Copy ACLs, groups, and source licensing onto each chunk when you index it, and keep them in sync when permissions change. A stale ACL is a data leak waiting for a prompt.
    2. Default-deny. Follow the explicit public-flag pattern. Never let a missing or empty permission field mean “everyone.”
    3. Resolve identity at query time, server-side. The model should never pick the filter. Your backend attaches the caller’s user and group IDs after authentication.
    4. Benchmark with permissions turned on. Bridgewater built about 200 hand-labeled examples in under a week and measured precision at 10 rising from about 60% to the 90–95% range. Run your test set as different users, including narrow-access ones, or post-filtering failures will hide from you.
    5. Plan for agent load. Size for machine query rates, not human ones. Bridgewater’s own gap between human and agent traffic is, in Linsky’s words, “a few orders of magnitude.”
    6. Flatten or pre-compute nested groups. Even the vendor says nested groups get “tricky.” Expand memberships in your pipeline instead of hoping the search engine resolves them.

    Here is a hypothetical to test your design. An analyst in a restricted deal team asks the assistant about a target company. The best-matching memo belongs to a different team. With naive post-filtering, the analyst gets nothing useful. With a missing ACL, they get a memo they should never see. Both outcomes are failures, and only one of them shows up in a latency dashboard.

    The Bottom Line

    Bridgewater’s account shows that permission-aware retrieval decides whether enterprise AI search is usable and safe at scale. The impressive 450-million-vector, 120 ms figures are turbopuffer’s numbers. The durable lesson is architectural: carry entitlements on every document, filter at query time from a trusted identity source, and test recall as real users with real restrictions.

  • Microsoft’s 2024 Claim That EU Rules Made the CrowdStrike Outage Possible Still Frames the Kernel Fight: What the UK CMA Probe Means for Your Endpoint Stack

    Microsoft’s 2024 Claim That EU Rules Made the CrowdStrike Outage Possible Still Frames the Kernel Fight: What the UK CMA Probe Means for Your Endpoint Stack

    The outage that grounded airlines is now a policy argument between a platform owner and regulators. Whoever wins shapes how deep your EDR agent gets to reach, and your rollback plan should not wait for the verdict. Much of the debate over Microsoft, EU kernel access, and the CrowdStrike outage still traces back to a July 2024 Microsoft claim, not a fresh 2026 statement. What is new is a UK Competition and Markets Authority review that puts Windows and Microsoft security in the same regulatory frame.

    Microsoft’s EU Kernel Access Argument Dates to 2024, Not This Year

    After the July 19, 2024 CrowdStrike Falcon update disrupted Windows systems worldwide, Microsoft pointed reporters to a 2009 European Commission interoperability undertaking. As contemporaneous coverage summarized it, a Microsoft spokesperson said the undertaking “gives makers of security software the same level of access to Windows that Microsoft gets,” and contrasted that with Apple’s more closed model. The European Commission pushed back. A Commission spokesperson told Euronews that Microsoft is free to adapt its security infrastructure within EU competition law, that the incident was not limited to the EU, and that Microsoft had not raised security concerns about the 2009 agreement with the Commission before or after the outage.

    I could not find a 2026 primary Microsoft statement repeating that EU-rules-made-the-outage-possible line. Treat it as Microsoft’s 2024 claim, still cited in policy debates, not as news Microsoft just made.

    What the UK CMA Is Actually Examining

    The current UK regulator is the Competition and Markets Authority (CMA), not a separate cybersecurity agency. On May 14, 2026 it opened a strategic market status (SMS) investigation into Microsoft’s business software ecosystem. On October 16, 2026 it published a proposed decision to designate Microsoft with SMS for that ecosystem and opened consultation (closing 5pm on November 13, 2026), with a statutory final-decision deadline of February 2027.

    The provisional scope groups five digital activities, including a PC operating system activity (Windows) and a Security Software activity. The Security Software description covers software that secures operating systems, applications, and data through authentication, endpoint management, and threat detection and prevention. The CMA provisionally puts Entra, Active Directory, Intune, and Defender in that bucket, with some Purview, Security Copilot, and Agent 365 functions also in scope. Its press release says the indicative roadmap includes “Examining whether the way Microsoft bundles and integrates its security products limits customer choice.” That is about competition and choice in a stack that includes both Windows and Microsoft security. It is not a dedicated CrowdStrike-outage or kernel-access case. SMS designation, the CMA notes, does not assume wrongdoing.

    Where Kernel Access Stands Now

    Separately, Microsoft has been redesigning how third-party security runs on Windows. In its June 26, 2025 Windows Resiliency Initiative update, David Weston wrote that a private preview of the Windows endpoint security platform would let Microsoft Virus Initiative partners “start building their solutions to run outside the Windows kernel,” so antivirus and endpoint protection “can run in user mode just as apps do.” Microsoft’s WRI product page still lists the Windows Endpoint Security Platform (WESP) as enabling protection solutions “to run outside the kernel,” alongside MVI 3.0 safe-deployment requirements such as gradual rings.

    That does not mean kernel-mode agents are gone. An October 26, 2026 technical analysis by Pavel Yosifovich for Elastic Security Labs found WESP components (wesp.sys and espclient.dll) already present in Windows Insider builds, described the project as early, and noted that the preview client API appeared to lack a clear path for user-mode code to veto process creation the way classic kernel callbacks can. In other words: the migration is underway in engineering previews, not finished as a fleet-wide customer mandate. Keep “made the outage possible” attributed to Microsoft’s 2024 claim about the access model, not as a regulator’s finding.

    My Recommendations for Endpoint Buyers

    These are my opinions for IT and security leaders, not legal advice and not CMA or Microsoft guidance.

    • Ask every EDR vendor where their critical path still lives. Kernel driver, WESP user-mode path, or hybrid. Get a dated roadmap for staged updates and crash recovery.
    • Require safe deployment practices in the contract. Rings, kill switches, and a proven rollback that does not depend on the agent booting cleanly.
    • Separate platform concentration from best-of-breed detection. The CMA’s provisional framing is that Windows and Microsoft security reinforce each other. Even if you standardize on Defender, keep an independent telemetry path and an offline recovery plan for boot failures.
    • Tabletop a kernel-level agent failure. Not just malware. Assume a bad content update and time how long until machines are productive again.

    Picture this hypothetical. Your primary EDR still ships a kernel driver while a WESP path is “coming.” A content update fails on Friday night. If your runbook only covers malware quarantine, you are replaying 2024. That scenario is my illustration, not a new incident.

    The Bottom Line

    Microsoft’s EU-undertaking explanation for third-party kernel reach is a 2024 claim the Commission disputed. The live story in late 2026 is the UK CMA’s proposed SMS designation covering both Windows and Microsoft security, plus a still-early push to move endpoint agents out of the kernel. Build your stack for staged rollouts and boot-time recovery now. Do not wait for the consultation to close.

  • A Researcher Published 11 Tailscale SSH Flaws, Now Fixed in v1.104.4: How to Contain Your Mesh VPN When One Device Gets Popped

    A Researcher Published 11 Tailscale SSH Flaws, Now Fixed in v1.104.4: How to Contain Your Mesh VPN When One Device Gets Popped

    Mesh VPNs sold us on “every device is its own perimeter.” That only holds if one bad laptop cannot walk straight to root on the database box. This week’s Tailscale SSH vulnerability root access disclosure says the ACL you trusted may not be the wall you thought it was, at least for Tailscale SSH with forwarding enabled.

    What the Researcher Published

    On October 24–25, 2026, Erica Windisch of Cyberdione Labs published Tailworm, a lab security assessment of Tailscale client v1.104.2 and self-hosted headscale v0.29.5. The report lists 11 findings across Tailscale SSH, experimental conn25 app connectors, serve/funnel, the DERP client, and headscale. All were “verified by reproduction in a lab environment on unmodified, released software,” the report says. It describes lab tests, not attacks seen on customer networks.

    The headline class is Tailscale SSH unix-socket forwarding (findings 1–3). Per the report, forwarding path checks and file operations run as root inside tailscaled. With one approved device and an SSH policy rule that permits forwarding, an ordinary session can dial or replace privileged sockets such as Docker or systemd. Windisch reports lab results of root execution on targets, capture of enrollment keys, and, on a Tailnet-Lock-disabled lab setup, joining a rogue node. The report’s own words: restricting SSH to non-privileged accounts is “defense-in-depth only” and “does not mitigate these findings.” Check mode, it says, gates session start only and does not constrain forwarding once a session is open.

    Other findings cover experimental app connectors choosing destinations outside ACL visibility, a serve TCP forward to the local admin API, memory exhaustion in SSH sessions and relays, and headscale onboarding lockout. Those are researcher claims with lab evidence; treat them as Cyberdione Labs data until Tailscale publishes its own per-finding advisory.

    Patch Status: Fixed in Tailscale v1.104.4

    Update, November 3, 2026: Tailscale has shipped a fixed client. Tailscale v1.104.4, released October 28, 2026, makes Tailscale SSH perform Unix socket forwarding as the authenticated user rather than as root, which blocks peers from reaching root-only sockets such as the Docker socket or the tailscaled local API. The same release bounds SSH session memory and tightens serve Unix backends. Tailscale published these as security bulletins TS-2026-014 (SSH Unix socket forwarding to root-only sockets), TS-2026-015 (SSH session memory exhaustion), TS-2026-016 (serve Unix backends reaching privileged sockets), and TS-2026-017 (HTTP/2 serve capability mix-up), each fixed in “1.104.4 or newer.” v1.104.3 was internal-only, and the current client is v1.104.5 (November 2). If you use Tailscale SSH or Serve, upgrade to v1.104.4 or newer now.

    For the record, this post was first drafted on October 27, when the newest public client was still v1.104.2, the version the researcher tested. On October 26, Tailscale published Tailscale SSH: Investigation and remediation update, calling the privilege-escalation issues “serious security issues” and targeting a release by the end of October 27. The fixed build landed a day later.

    On notice timing: The Hacker News reports Windisch saying she sent the report and proof-of-concept to Tailscale’s security contact on the Saturday before the public post, and she described the process on Hacker News as “reported but not coordinated.” The Tailworm write-up itself argues for “rapid awareness” to speed remediation. Treat notice timing as disputed enough that “without warning” overstates it.

    What Tailscale Confirms Versus What Remains Researcher-Only

    Tailscale confirms a serious privilege-escalation class for authorized SSH sessions and says exploitation “depends on the permissions granted to the connecting user or device, the system’s configuration, and other conditions.” It is “not recommending that everyone disable Tailscale SSH.” Its bulletins cover the SSH forwarding, SSH memory, and serve findings with a fixed version, but it has not published a finding-by-finding confirmation of all 11 Tailworm items. Separately, do not confuse this week with earlier 2026 bulletins TS-2026-006 and TS-2026-009, which fixed different Tailscale SSH tricks that could reach root despite autogroup:nonroot and were patched in v1.98.9. Tailworm’s core claim is a different class: forwarding that runs with daemon privilege from an already-allowed unprivileged session.

    My Recommendations: Contain the Mesh Now

    The steps below mix Tailscale’s published interim advice with my own operator opinion. This is not a substitute for Tailscale Support.

    1. Disable SSH forwarding where you do not need it. Tailscale’s update says to set TS_SSH_DISABLE_FORWARDING=1 in the tailscaled environment and restart. That mitigates several of the more serious findings without turning SSH off. On SaaS, also deny forwarding per SSH rule. On headscale 0.29.5, the researcher says forwarding cannot be disabled per rule, so treat every SSH grant as high risk until you can upgrade.
    2. Rewrite SSH rules as per-pair grants. Specific source device to specific destination. Drop identity-wide and tag-wide accept rules. The report calls per-pair grants the control that stopped lab spread.
    3. Audit serve and app connectors. Remove serve TCP forwards to unix sockets, especially anything aimed at tailscaled.sock. Pull connector entitlements from devices that do not need them.
    4. Upgrade to v1.104.4 or newer and plan containment. Update every node and turn on auto-updates where you can. If a device is compromised, delete it at the control plane. Reimaging alone does not revoke keys an attacker may have copied off the host. That last point is the researcher’s claim; Tailscale’s update does not repeat it, but it is sound incident practice in my view.

    Picture this hypothetical. A developer laptop is phished. That laptop already has a Tailscale SSH rule to your CI hosts with forwarding left on. Under the researcher’s model, that foothold is the whole prerequisite. Tighten the rule and kill forwarding before you need the hypothetical to become a ticket. That scenario is my illustration, not a reported breach.

    The Bottom Line

    Eleven lab-verified findings are public. Tailscale agrees the SSH privilege-escalation class is serious and shipped fixes in v1.104.4 on October 28, 2026. Until every node runs v1.104.4 or newer, assume Tailscale SSH with forwarding is a lateral-movement path from any device your policy already trusts, and shrink that trust now.

  • SEC Exam Priorities Put AI and Super-Intelligence Risk on This Year’s List: The Governance Paper Trail IT Should Start Building Now

    SEC Exam Priorities Put AI and Super-Intelligence Risk on This Year’s List: The Governance Paper Trail IT Should Start Building Now

    Most companies can describe their AI strategy in a slide. An examiner will ask for the inventory, the approval record, and proof the marketing claims match what the tools actually do. That evidence lives in IT systems, not in the pitch deck. If you are searching for SEC exam priorities AI risk governance, the place to start is the Division of Examinations’ current fiscal-year document, not a deck title.

    What the Priorities Documents Actually Say

    Today is late October 2026. The SEC’s fiscal year runs from October through September, so Fiscal Year 2027 is the current year. On October 23, 2026, the Commission announced the Division of Examinations’ Examination Priorities for Fiscal Year 2027. That is this year’s exam list, not next year’s. The working headline’s “Next Year” wording was wrong for that reason, so I retitled the post.

    The FY2027 PDF does not lean on the phrase “artificial intelligence” the way prior years did. It uses “super intelligence (SI).” The leadership message says registrants are “increasingly incorporating super intelligence (SI) into their businesses,” including “generative SI tools to support research, drafting, coding, and client communications, as well as advisory and portfolio management functions, and compliance monitoring and surveillance.” Under “Emerging Financial Technology,” the Division says it will prioritize examinations of “policies and procedures relating to monitoring SI, including its use in fraud prevention and detection, back-office operations, compliance with AML and Reg BI obligations, and trading functions.” It also flags SI-related cyber threats such as “SI-generated phishing messaging and synthetic or deepfake voice and video.”

    The same section keeps the automated-tools checklist firms already know from prior years. Reviews of automated investment advisory services and related tools will generally ask whether: “(1) representations are fair and accurate; (2) operations and controls are consistent with disclosures made to investors; and (3) algorithms lead to advice or recommendations consistent with investors’ investment profiles or stated strategies.”

    That continuity matters. The Division’s Examination Priorities for Fiscal Year 2026, released November 17, 2025, used “AI” explicitly. It said the Division would “review for accuracy registrant representations regarding their AI capabilities or AI” and assess “adequate policies and procedures to monitor and/or supervise their use of AI technologies,” including fraud prevention, back-office, AML, and trading. FY2026 cybersecurity language also called out “new risks associated with artificial intelligence (AI) and polymorphic malware.” So AI-related exam focus is not brand new. FY2027 reframes and intensifies it as SI while keeping the automated-tools and representation tests.

    When Marketing Claims Did Not Match Reality

    The representation test has teeth. On March 18, 2024, the SEC announced settled charges against Delphia (USA) Inc. and Global Predictions Inc. for what Chair Gensler called “AI washing,” false and misleading statements about purported AI use. Per the SEC press release, Delphia claimed it used AI and machine learning with client data when it did not have those capabilities, and Global Predictions falsely claimed to be the “first regulated AI financial advisor” and to offer “expert AI-driven forecasts.” The firms paid $225,000 and $175,000 in civil penalties, respectively. The Division of Examinations assisted the investigations. That case is enforcement, not an exam-priorities checklist, but it shows why “representations are fair and accurate” is not a soft ask.

    My Recommendations: Build the Paper Trail Now

    Everything below is my opinion for IT and compliance teams at registered firms and the vendors that serve them. The priorities PDFs do not publish a fixed document request list, and an AI inventory is not a named statutory form.

    1. Inventory every AI or SI tool that touches advice, trading, marketing, AML, surveillance, or customer data. Record the vendor, model or product name, business owner, data classes ingested, and whether outputs can reach clients or filings. Shadow SaaS counts.
    2. Map claims to capability. Pull website copy, Form ADV language, pitch decks, and sales emails that mention AI or SI. Line each claim up against what the tool actually does today. If you cannot demonstrate the capability, change the claim before an examiner asks.
    3. Write and implement monitoring procedures. FY2026 and FY2027 both ask whether firms have policies to monitor or supervise AI/SI use in fraud detection, back-office, AML, Reg BI, and trading. Document who reviews outputs, how often, and what happens when a model drifts or a vendor changes a feature.
    4. Keep testing and change records. For advisory or recommendation algorithms, retain evidence that outputs stay consistent with stated strategies and investor profiles after model updates. That is how you answer the priorities’ algorithm-consistency questions without scrambling.
    5. Treat deepfake and SI phishing as exam-relevant cyber risk. FY2027 expressly ties SI to social-engineering threats. Update training, call-back verification, and incident playbooks, and keep the dated materials.

    Picture this hypothetical. An examiner’s document request asks for a list of AI tools used in marketing and the supporting policies. Your deck says the firm uses “proprietary AI research.” Your inventory shows a generic chatbot drafting blog posts. Closing that gap now is cheaper than explaining it under oath later. That scenario is my illustration, not a published SEC request letter.

    The Bottom Line

    The Division of Examinations’ FY2027 priorities, announced October 23, 2026, put super intelligence, automated investment tools, and the accuracy of technology claims on this year’s exam list. FY2026 had already named AI. In my view, the firms that will look prepared are the ones that can produce an inventory, the approval trail, and proof that public claims match the tools in production.

  • 79% of Third-Party Breach Victims in Kroll Cases Were Hit Through Technology Firms: Rebuilding Vendor Risk Reviews Around Who Actually Holds Your Data

    79% of Third-Party Breach Victims in Kroll Cases Were Hit Through Technology Firms: Rebuilding Vendor Risk Reviews Around Who Actually Holds Your Data

    The annual vendor questionnaire asks whether a supplier has a security policy. The third-party data breach vendor risk management problem landing this year asks a different question: how many of your customers’ records sit in a SaaS tool someone signed up for with a corporate card. New case data from Kroll makes that gap hard to ignore.

    What Kroll Actually Measured

    On October 22, 2026, Kroll published its Data Breach Outlook 2026: The Third-Party Domino Effect. The firm analyzed data breach incidents handled by its Cyber and Data Resilience practice between June 2025 and June 2026, and ran a companion survey of 250 risk and compliance leaders, C-suite executives, and board members. These are Kroll’s own cases and survey respondents, not a census of every breach worldwide. Treat the figures as vendor data from that caseload.

    Two numbers drive the headline. First, 69% of affected individuals in those incidents “were breached due to a third-party breach,” as Kroll’s publication page puts it. Second, of that third-party group, 79% “were from the technology sector,” compared with finance (9%), insurance (6%), and healthcare (5%). Kroll’s press release says the same thing another way: 79% of those third-party-affected individuals “were affected by a breach from the technology sector.” Technology firms also accounted for 60% of all affected individuals in the caseload, including direct and supply-chain incidents.

    That is why I changed the sheet title. “Most third-party breach victims were hit through a tech vendor” is directionally right, but Kroll’s wording is “technology sector” or “technology firms,” and the precise claim is 79% of third-party-affected individuals in its June 2025–June 2026 cases. Kroll does not publish a product taxonomy that separates SaaS from on-prem software or MSPs. Denyl Green, Kroll’s global head of breach notification, describes the link as “a vendor, software supplier or service provider.”

    Organizations Breached Versus People Affected

    Count companies instead of people and the picture flips. Finance was the most breached industry by organization count, at 20% of Kroll-handled breaches in the period. Professional services also accounted for 20%. Technology ranked only seventh by number of organizations breached, about 6% of incidents in press coverage of the report, yet first by total individuals affected. Insurance ranked ninth by organizations breached and second by people affected.

    Kroll’s point is the domino effect. “One vendor compromise can generate breach notifications for dozens or hundreds of downstream organizations that never touched the vulnerable system directly.” Green puts it in operational terms: when a supplier is breached, every client still has to decide whose data was affected and whether customers and regulators need notice. The survey side of the report adds that 54% of IT and technology leaders named vendor and third-party breaches as their biggest compliance risk, versus 40% of C-suite executives and 30% of risk and compliance leaders, and that dependence on vendors was among the top four breach-response challenges, cited by 41% of respondents.

    My Recommendations: Rebuild Reviews Around Who Holds the Data

    Everything in this section is my opinion, not Kroll guidance and not legal advice. If blast radius tracks the firms that hold many customers’ data, the review should start there, not with the longest questionnaire.

    1. Inventory who actually holds or can reach personal data. List every SaaS product, file-transfer tool, MSP, backup provider, HR or CRM platform, and support vendor that stores, processes, or can export your customer or employee records. Include shadow IT someone bought with a card. If you cannot name the data classes each holds, you cannot tier them.
    2. Tier by blast radius, not by spend. A low-cost CRM that holds your full customer file outranks an expensive consultancy that only sees sanitized decks. Put high-blast vendors on continuous review: current SOC 2 Type II or ISO/IEC 27001 scoped to the service you use, MFA and logging requirements, and a named security contact.
    3. Rewrite notification and audit clauses for shared incidents. Require the vendor to tell you within a fixed number of hours, not “promptly,” and to share enough detail for your own regulator and customer notices. Add a right to receive evidence that your tenant was or was not in scope.
    4. Cut standing access. Prefer customer-controlled encryption keys where the product allows it, time-bound support access, and least-privilege API scopes. A vendor breach is less useful to an attacker if the vendor cannot reach everything by default.
    5. Run a joint tabletop on a supplier compromise. Decide in advance who owns regulator notice, customer notice, and insurer notice when the first alert comes from a vendor, not from your SOC.

    Picture this hypothetical. Your marketing team adopts a new email platform. Six months later that platform discloses a breach affecting multiple customers. If your inventory already shows it holds email addresses and purchase history, notification planning starts the same day. If the tool never made the vendor list, you spend the first week discovering you are in scope. That scenario is my illustration, not a Kroll case.

    The Bottom Line

    In Kroll’s June 2025–June 2026 cases, most people affected by a breach were affected through a third party, and most of those people were tied to technology-sector compromises. Finance still led in how many organizations were breached. The practical takeaway, in my view, is to rebuild vendor risk around who holds your data and how wide the blast radius is, not around who filled out the longest security questionnaire.

  • Exposed Attacker Servers Show Ransomware Affiliates Handing Parts of Intrusions to AI Agents: What Changes When the Hands on Keyboard Aren’t Only Hands

    Exposed Attacker Servers Show Ransomware Affiliates Handing Parts of Intrusions to AI Agents: What Changes When the Hands on Keyboard Aren’t Only Hands

    Your SOC playbook quietly assumes the attacker gets tired, makes typos, and logs off for the night. An affiliate that delegates the boring steps of a ransomware affiliates AI agents intrusions workflow does none of those things. So the gap between first alert and encryption is the number to re-measure. Two research teams recently got an unusual look at that workflow after attackers left their own servers open on the public Internet.

    Who Found the Exposed Servers

    In late September 2026, the SOCRadar Threat Research Unit (STRU) found a publicly accessible command-and-control server at 45.227.254[.]43. Configuration files, logs, and credentials on that host led them to six more adversary servers, most on AS267784 (Flyservers S.A.). The Atos Threat Research Center (TRC), in a separate report by Piotr Mazurkiewicz, recovered material from an Internet-accessible host tied to the same INC ransomware ecosystem. Both teams describe affiliates of INC, also tracked as GOLD IONIC or MITRE G1032, with some cases branded as The Gentlemen. STRU rates confidence in a formal dual-brand alliance as low, and confidence that operators of both groups used a shared environment as moderate-to-high.

    What they found was not a malware dump. STRU recovered AI session databases, a central knowledge base, and the source of Phantom, which the operators called an “Autonomous Red Team Kill Chain Framework.” Atos TRC recovered a 1,882-record knowledge export, about 846 MB of tooling and victim data across 444 files, and HTTP logs showing successful downloads from the exposed service. STRU says the operators logged 8,257 messages across 133 OpenCode sessions and that one OpenRouter account showed about $23,920 in recorded spending. Those figures are vendor findings from the recovered logs, not independent measurements. STRU also graded 62 named organizations by depth of activity, 36 of them in the United States, and notes that many of the exploitation counts come from operator records that were not independently verified.

    What Was Handed to AI Agents, and What Was Not

    This is the distinction that matters for the headline. STRU assesses with high confidence an “AI-executed, human-approved operating structure.” A human operator started each campaign with a victim-specific master prompt and reviewed every step under the directive “coordinate every action with me.” The AI stack, OpenCode as the interface, commercial models hosted on OpenRouter (STRU names anthropic/claude-opus-4.6 and moonshotai/kimi-k3 among those used), Phantom’s skill library, and a persistent knowledge base, planned reconnaissance, selected techniques, generated commands, triaged stolen data, and drafted reports and extortion material. “WormGPT” was a persona prompt and a set of MCP servers, not a custom model. When a model refused a request, STRU says the operator rewrote it as an “Authorized Red Team Assessment Request” instead of switching models.

    Atos TRC is even more explicit. Its conclusion states that “the recovered evidence does not support the idea of autonomous AI conducting intrusions on its own.” AI “acted as a force multiplier for human operators,” reducing the effort to research targets, develop scripts, analyze data, document findings, and coordinate campaigns. Atos also notes that Phantom coordinated workflows, while individual actions could still run through custom Phantom agents, Mythic or Apollo C2, tunneling tools, or manually operated utilities. Recovered operator prompts Atos quotes are short and human-written, for example asking the assistant to continue from a session backup or to escalate from a domain user to Domain Admin given a service list. So the sheet headline “Handing Intrusions to AI Agents” overstates it. The accurate framing is that affiliates handed parts of the intrusion, planning, command generation, data triage, and drafting, to LLM-driven agents, while a human still chose targets and approved steps.

    Separately, both teams describe plenty of plain automation and conventional tooling that is not agent decision-making: FOFA-style scanning of edge devices, reverse-SOCKS tunnels (RevSocks, Chisel, Ligolo), scheduled-task persistence, BYOVD security-tool killers, restic or rclone exfiltration to cloud storage, and custom INC encryptor builds. STRU reports operator records claiming large-scale abuse of FortiClient EMS (CVE-2026-35616), SimpleHelp (CVE-2024-57727), and cPanel (CVE-2026-41940), and labels those counts as operator-reported. Treat the scanning and exploit claims as operator data, not confirmed victim totals. STRU also recovered 46 personalized INC encryptor builds staged for an October 1, 2026 wave and a September 30 SimpleHelp screenshot showing encryption on a workstation; deployment against the staged targets is unconfirmed.

    What Changes for Defenders

    The recommendations below are my opinion, not vendor guidance and not legal advice. If the planning and typing are machine-assisted, your response timelines need to assume a shorter dwell between foothold and impact.

    • Re-measure the first-alert-to-impact window. Build tabletop scenarios that assume continuous command generation overnight, not a human who sleeps. The SOCRadar and Atos material both show operators active on a schedule, but the agent does not get tired between approvals.
    • Prioritize edge and RMM hygiene. STRU’s hunting notes focus on rogue EMS admins, modified VPN on_connect scripts, unauthorized SimpleHelp or MeshCentral agents, and MSP credential reuse. Patch those surfaces and require MFA and IP allowlists on every remote-management path.
    • Alert on tooling the agents call. Watch for RevSocks, Chisel or Ligolo patterns, restic or rclone to unexpected cloud storage, Zemana-style vulnerable drivers paired with security-product termination, and mass .INC renames or INC-README.txt drops.
    • Treat data-extortion clocks separately. These operators preferred proof packs, pricing from stolen financials, and customer-facing notice pages. Have IR, legal, and notification templates ready before encryption starts.

    Picture this hypothetical. Your overnight SIEM shows a new EMS admin and a burst of PowerShell that looks like a junior analyst scripting. Under an agent-assisted playbook, that burst may be the middle of an approved kill-chain step, not a tired human. That scenario is my illustration, not a case from either report.

    The Bottom Line

    Exposed affiliate infrastructure shows LLM agents planning steps, generating commands, and drafting extortion text inside a human approval loop. It does not show fully autonomous ransomware. In my view, the practical change is shorter assumed dwell and louder hunting around the tools those agents call, not a rewrite of every detection rule overnight.

  • A Top Lawmaker Says Data Centers Will Be Critical Infrastructure: What That Designation Could Mean for Your Colo Contracts, Audits, and Incident Reporting

    A Top Lawmaker Says Data Centers Will Be Critical Infrastructure: What That Designation Could Mean for Your Colo Contracts, Audits, and Incident Reporting

    For about a decade, data centers have been treated like commercial real estate with good air conditioning. Now a senior lawmaker says that is going to change, and the phrase people are searching for is a data centers critical infrastructure designation. Before you rewrite your colo contracts, though, it is worth separating what was said, what has been proposed, and what is actually in force today. Those are three different things.

    What Was Actually Said, and by Whom

    The remark came from Rep. Jim Himes, D-Conn., the ranking member of the House Intelligence Committee, at a Brookings Institution event on Wednesday, October 21, 2026. According to Nextgov/FCW’s report, Himes said “I think there’s no question” that data centers will be considered critical infrastructure in the future. He added that “they’re going to be even more important than the network data centers that we’ve been living with for 40 years now because so much more functionality is going to rely on their health and existence.”

    Nextgov also reports that Himes said securing them will require a “robust compute infrastructure,” with services that “fragment and disperse” so that “critical functions are redundant and protected.” He predicted AI data centers “will start to look more like JP Morgan’s resilience centers than they look like today’s data centers.”

    That is one lawmaker’s prediction. It is not a designation, a bill he introduced, or an agency action. That is why I changed the headline from “Lawmakers Say Data Centers Will Be Treated as Critical Infrastructure” to “A Top Lawmaker Says Data Centers Will Be Critical Infrastructure.” Nextgov itself notes that although AI data centers “already underpin some of the critical infrastructure, they are not yet deemed to be a standalone sector.”

    Where Data Centers Sit Today

    CISA’s critical infrastructure sectors page lists 16 sectors. There is no Data Center sector. The closest fit is the Information Technology Sector. CISA describes it as producing and providing “hardware, software, and information technology systems and services, and—in collaboration with the Communications Sector—the Internet.” DHS is its Sector Risk Management Agency.

    One caution: neither CISA’s sector list nor its IT sector page uses the words “data center.” So it is fair to say that the functions running inside a data center fall within the IT sector, and its overlap with Communications, as CISA defines them. It is not accurate to say CISA has formally assigned data centers as a category to a specific sector. Nothing about current rules has changed because of the Himes remarks.

    Is Anything Formal in Motion?

    There is one bill worth knowing about. H.R. 8711, the Data Infrastructure Risk Reduction Act, was introduced on May 7, 2026, by Rep. Suhas Subramanyam, D-Va. It is a different lawmaker from Himes. If enacted, it would require DHS, acting through CISA, within 180 days to “identify data centers that should be treated as critical infrastructure,” to consider the security of the power and water infrastructure connected to them, and to submit to Congress a strategy to “defend data centers described in this subsection from external breaches from malefactors.” Congress.gov lists it as “Introduced,” and its latest action was a referral to the House Homeland Security subcommittee on May 8, 2026. It is a proposal, not law. Even if passed, it asks for an identification exercise and a strategy, not a new sector or new rules for operators.

    On incident reporting, the status is also proposed rather than in force. CISA’s CIRCIA page says the proposed rule was published April 4, 2024. It says “CISA continues to work on the final rule,” and that “multiple funding lapses impacted CISA’s ability to conduct rulemaking activity.” It also states plainly: “Until the effective date of the final rule, organizations are not required to submit covered cyber incident or ransom payment reports under CIRCIA.” The statute itself calls for reports within 72 hours of a covered incident and within 24 hours of a ransom payment once the final rule takes effect. Who counts as a covered entity is something the final rule will settle.

    My Recommendations for Your Colo Contracts

    Everything in this section is my opinion, not legal advice, and not a requirement that exists today. I think there is no reason to wait for a designation. Here is what I would review now.

    Incident-Reporting Clauses

    • Ask for a contractual notification window measured in hours, not “promptly.” If CIRCIA’s final rule eventually covers you, its 72-hour clock starts when you reasonably believe an incident occurred, and you cannot meet that if your provider tells you a week later.
    • Require the provider to share enough detail (timeline, affected systems, indicators) to support your own reports to regulators, insurers, and customers.

    Audit Rights and Evidence

    • Get the current SOC 2 Type II report and ISO/IEC 27001 certificate, and check that the scope actually includes the facility and services you use. A certificate scoped to headquarters does not help you.
    • Negotiate a right to audit or to receive bridge letters between report periods, plus a right to see the physical security and resilience testing summaries.

    Resilience and Dependencies

    • Ask how the facility handles utility power and water disruptions, the same dependencies H.R. 8711 calls out. Get the answer in writing.
    • Map which of your services sit in a single building. Himes’s “fragment and disperse” point applies to your own architecture as much as to the industry.

    Change-of-Law Language

    • Add a clause that commits both parties to renegotiate in good faith if new security or reporting obligations apply to the provider. That way a future designation is a conversation, not a dispute.

    Picture this hypothetical. A provider is later identified as critical infrastructure under a bill like H.R. 8711 and begins receiving federal security guidance. Customers with change-of-law and evidence clauses can simply ask for updated documentation. Customers without them have to wait for the next renewal. That scenario is my illustration, not something that has happened.

    The Bottom Line

    A top House Intelligence Democrat says there is “no question” data centers will be considered critical infrastructure. A bill that would have CISA identify which ones should be treated that way sits in committee. CIRCIA reporting is still waiting on a final rule. None of that changes your obligations today. In my view, it is still a good reason to make your colo contracts ready for the day it does.