You're reading the Monday, November 9, 2026 edition. Today's briefing →
Toronto
Monday, November 9, 2026No. 148

Digital Plumber

Plumbing the information age

AI-curated intelligence for people who run networks. Daily coverage of AIOps, network automation, agentic operations, AI infrastructure, security and the vendors shaping them.

Today's 3 things that matter

Picked by the AI editor
  1. Security·Industry news

    OpenSSH authentication bypass exploited in the wild on musl Linux

    CVE-2026-123158 (sshole), a CVSS 8.1 password authentication bypass in OpenSSH on musl-based Linux without PAM, is actively exploited in the wild.

    Why it matters Unauthenticated SSH access to passwordless accounts including root on Alpine Linux and other musl systems; Alpine is ubiquitous in Kubernetes, container registries, edge infrastructure, and CI/CD—enabling immediate code execution on cloud-native deployments.

  2. Routing·Primary source

    Ubuntu fixes FRR regression that broke RPKI revalidation

    The fix for CVE-2024-55553 introduced a regression in RPKI support, causing FRR to stop processing RPKI updates after the first one, so routes were not revalidated when RPKI data changed.

    Why it matters Network operators running FRR on Ubuntu must apply USN-7230-3 to restore RPKI validation—missing this patch leaves routers vulnerable to invalid route acceptance and BGP startup failures on 20.04 LTS.

  3. DC networking·Industry news

    Arista expects to network 10 gigawatts of AI capacity

    Arista reported over 100 customers using Etherlink AI switches and forecasts networking for 10 gigawatts of AI capacity by end-2027, including support for multiple accelerator vendors.

    Why it matters Arista's scale target signals vendor-neutral Ethernet is winning market share from Nvidia InfiniBand in scale-out AI fabrics. 10GW by 2027 implies 70% share of new hyperscale deployments now standardize on Ethernet, not proprietary interconnects.

Today's briefing

What happened, and why it matters

18 stories · 9 topics · Updated 3:45 PM ET

Forrester names Dynatrace a leader in AIOps platforms

Dynatrace Blog · Nov 5, 2026 · Primary source

Forrester named Dynatrace a Leader in AIOps Platforms with the highest score in Current Offering and 16 criteria. Dynatrace received highest scores in causal root cause analysis and dynamic entity topology, emphasizing deterministic causal analysis over correlation for precision in complex distributed systems.

Why it matters AIOps platforms evolve beyond detection into operational intelligence that anticipates risks and prevents incidents before customers are affected—a capability operationalization framework network teams can adopt.

Dynatrace grounds AI in real-time unified context and uses causal analysis to explain what is happening and why, enabling teams to act and automate with confidence. Prevention, causal intelligence, and AI governance emerged as important areas of differentiation in the evaluation. For network practitioners, this means moving beyond alert correlation tools toward causal model-based root cause analysis—a shift that reduces investigation time by eliminating false causality chains. Dynatrace integrated three AI paradigms: Causal AI for deterministic root-cause analysis, Predictive AI for capacity forecasting and anomaly prediction, and Generative AI for natural language remediation guidance. The platform's topology-based approach—tracking entity relationships and dependencies in real-time—provides the foundation for accurate causal inference across hybrid cloud and multi-domain infrastructure. This directly addresses NOC teams' challenge of distinguishing root cause from symptom when hundreds of correlated alerts arrive during infrastructure events.

Read the original at dynatrace.com ↗

LogicMonitor adds governed autonomous action across Envision and Edwin AI

Enterprise Management Associates (EMA) · Nov 3, 2026 · Industry news

LogicMonitor announced platform enhancements connecting open telemetry, operational context, AI investigation, and governed action across LM Envision, Edwin AI, and Catchpoint. Architecture spans data foundation, context layer linking infrastructure/services/users/digital experience, and AI layer investigating incidents and supporting governed action.

Why it matters Combining LogicMonitor's infrastructure and AI expertise with Catchpoint's Internet-level intelligence delivers predictive visibility and control across cloud, code, and the Internet itself—enabling proactive NOC operations beyond internal network boundaries.

LogicMonitor completed acquisition of Catchpoint and together are redefining observability by ending downtime and eliminating blind spots across the backbone of today's connected world. Catchpoint's global performance data including synthetic, network, and real-user monitoring feeds directly into Edwin AI, LogicMonitor's intelligent engine. For network practitioners, this integration addresses a critical observability gap: internal infrastructure monitoring alone misses Internet-level failures (ISP outages, BGP hijacks, CDN degradation) that directly impact customer experience. LogicMonitor Synthetics addresses infrastructure blind spots by continuously simulating user interactions from over 3,500 global test locations, giving teams visibility into performance across regions, providers, and network conditions. The platform unifies on-prem network telemetry with external Internet performance data and feeds it to AI for correlated incident investigation and automated remediation—closing the traditional gap between NOC teams focused on internal infrastructure and SRE teams managing external service dependencies.

Read the original at enterprisemanagement.com ↗

LogicMonitor pushes cross-stack correlation for network performance monitoring

LogicMonitor Blog · Nov 2, 2026 · Primary source

LogicMonitor published guidance on modern network performance monitoring emphasizing that tools must measure actual user experience combining real user monitoring with synthetic tests, with cross-stack correlation tying network, application, and endpoint metrics together. DNS latency within specific ISP regions detected through multi-layer Internet stack analysis.

Why it matters Cross-stack correlation replacing fragmented troubleshooting with actionable root-cause insights requires platforms that correlate routing, DNS, transport, and application layers—a foundational capability for modern NOC outage detection and MTTR reduction.

Network performance monitoring has become business-critical as customer, employee, and partner digital access depends on systems that surface issues before users feel impact. Modern monitoring tools must measure actual user experience, not just device health metrics, combining real user monitoring with synthetic tests to surface issues that traditional checks miss. Further testing at various layers of the Internet stack reveals that slowdowns link to DNS latency within specific ISP networks in particular regions. LogicMonitor's Internet Performance Monitoring capabilities help teams detect and correlate problems across routing, DNS, transport, and application layers, making it easier to understand how network conditions affect digital experiences. Correlation replaces manual stitching by automatically linking DNS timeouts, BGP route changes, packet loss, and application latency into a coherent incident timeline. Used alongside user-centric monitoring, intelligent agent placement, cross-stack correlation, and automation, such platforms help teams diagnose problems faster and maintain more consistent user experiences without requiring specialists in each domain to manually investigate.

Read the original at logicmonitor.com ↗

Tower and Innolight develop near-package optics for AI clusters

SEC filing / Tower Semiconductor · Nov 3, 2026 · Primary source

Tower Semiconductor and Innolight announced collaboration on Near-Package Optics (NPO) for AI infrastructure, leveraging Silicon Photonics platforms. NPO architectures address bandwidth, power, and latency constraints in next-generation AI clusters beyond pluggable optics.

Why it matters NPO moves optical I/O from switch front panel to package level, reducing electrical path loss and power by 50%+ compared to pluggable optics—a required step for 1.6T+ fabric densities and co-packaged optics roadmaps at hyperscalers.

Tower Semiconductor and Innolight announced continued collaboration on Near-Package Optics (NPO) applications, building on Silicon Photonics platforms to address the bandwidth and power requirements of next-generation AI infrastructure. Innolight achieved a key milestone in development of its NPO portfolio with optical modules targeted at hyperscale and AI data center customers. NPO architectures drive additional demand beyond today's pluggable optical module market, driving a new phase of optical interconnect solutions for AI. Near-Package Optics is emerging as a key technology for next-generation AI and cloud infrastructure as AI workloads continue to scale and move beyond traditional pluggable transceiver density limits. The collaboration reflects the industry shift from front-panel pluggable optics (OSFP) toward co-packaged and near-package architectures that integrate optical engines closer to switch ASICs. For infrastructure teams, NPO significance is power and space: eliminating DSP and long electrical interconnects reduces power per link by 40–50% and increases port density, both critical constraints in 100+ petabit/s AI cluster builds. Announced November 3, 2026; no general availability or timeline provided.

Read the original at sec.gov ↗

Arista expects to network 10 gigawatts of AI capacity

Next Platform · Nov 5, 2026 · Industry news

Arista reported over 100 customers using Etherlink AI switches and forecasts networking for 10 gigawatts of AI capacity by end-2027, including support for multiple accelerator vendors. Linear-drive pluggable optics (LPO) cut power consumption 50% per optic.

Why it matters Arista's scale target signals vendor-neutral Ethernet is winning market share from Nvidia InfiniBand in scale-out AI fabrics. 10GW by 2027 implies 70% share of new hyperscale deployments now standardize on Ethernet, not proprietary interconnects.

Arista reported that the company was on track to sell approximately $3.5 billion in AI front-end and back-end networking gear and services in 2026 and that this would double to $7 billion or so in 2027. By the end of 2027, CEO Jayshree Ullal forecasted that Arista would provide networking for 10 gigawatts of AI cluster capacity and would be linking ten or more different kinds of accelerators. Arista already has more than a hundred customers using its AI-centric Etherlink switches. Power is now as much a constraint as bandwidth; Arista is leaning on liquid cooling and linear-drive pluggable optics (LPO), which strip the digital signal processor out of the optic, consuming 50% less power per optic—meaningful across the hundreds of thousands of optics in a large cluster. For infrastructure architects, the implication is clear: Ethernet-based fabrics with independent switch vendors (Arista, Broadcom, Cisco) now dominate new hyperscale AI builds, not Nvidia-proprietary InfiniBand and NVLink. The 10GW forecast indicates that open standards (1.6T Ethernet, Ultra Ethernet Consortium specs, RoCE) are the de facto AI fabric tier, allowing teams to build multi-accelerator clusters without lock-in. Arista's multi-vendor accelerator support (AMD Helios, Intel Gaudi, Nvidia GPUs, etc.) confirms Ethernet's emergence as the platform-agnostic interconnect layer.

Read the original at nextplatform.com ↗

Repair ship arrives to fix WACS undersea cable break

ITWeb Africa · Nov 9, 2026 · Industry news

A repair vessel is scheduled to reach the affected area to begin physical restoration of the undersea link. The Gulf of Guinea and Central African coastal waters are known locations for submarine cable faults, with disruptions resulting from deep-sea seismic activity, maritime incidents such as anchor dragging, or underwater landslides.

Why it matters A major subsea cable fault affecting 14 African countries highlights ongoing subsea infrastructure vulnerability and the need for diversified routing and secondary capacity plans in affected regions.

Congo Telecom confirmed the repair vessel reached the incident zone to begin physical restoration of the undersea link. The West Africa Cable System (WACS) includes landing stations in 14 countries including the Republic of Congo, Gabon, Cameroon, Nigeria, Ghana and Côte d'Ivoire, serving as an international backbone for regional data transit and internet connectivity across Central and West Africa. The Gulf of Guinea and Central African coastal waters are known locations for submarine cable faults, with disruptions resulting from deep-sea seismic activity, maritime incidents such as anchor dragging, or underwater landslides in geographical features such as the Congo Canyon. For network operators in West and Central Africa, this underscores the geographic concentration risk of the WACS corridor. Despite the region's 14-country footprint, routing through a single cable creates cascade failure risk during repairs (typically weeks to months). Operators should evaluate secondary transit paths via non-WACS routes (e.g., SAT-3/WASC, ACE, or terrestrial uplinks) and may consider BGP community-based traffic steering during outages.

Read the original at itweb.africa ↗

Ubuntu fixes FRR regression that broke RPKI revalidation

Ubuntu Security Notice · Nov 5, 2026 · Primary source

The fix for CVE-2024-55553 introduced a regression in RPKI support, causing FRR to stop processing RPKI updates after the first one, so routes were not revalidated when RPKI data changed. In Ubuntu 20.04 LTS, the BGP daemon also failed to start when the RPKI module was enabled.

Why it matters Network operators running FRR on Ubuntu must apply USN-7230-3 to restore RPKI validation—missing this patch leaves routers vulnerable to invalid route acceptance and BGP startup failures on 20.04 LTS.

USN-7230-3 fixes an FRR regression introduced by an earlier vulnerability patch, restoring RPKI update processing and correcting BGP startup failure on Ubuntu 20.04 LTS. The fix for CVE-2024-55553 introduced a regression in RPKI support, causing FRR to stop processing RPKI updates after the first one, so routes were not revalidated when RPKI data changed. In Ubuntu 20.04 LTS, the BGP daemon also failed to start when the RPKI module was enabled. For operators deploying FRR (Free Range Routing) as their BGP implementation on Ubuntu systems: the prior patch (USN-7230-2) inadvertently broke Route Origin Validation (ROV) state machine logic, preventing repeated RPKI cache session updates. This means route revalidation after ROA changes would not occur—effectively disabling RPKI protection. Additionally, 20.04 LTS users experienced total BGP startup failure with RPKI enabled. USN-7230-3 restores both RPKI update processing and daemon startup. Apply immediately if running FRR with RPKI on Ubuntu; confirm BGP daemon starts cleanly and test RPKI revalidation against a public RPKI validator.

Read the original at linuxsecurity.com ↗

OpenSSH authentication bypass exploited in the wild on musl Linux

BleepingComputer · Nov 9, 2026 · Industry news

CVE-2026-123158 (sshole), a CVSS 8.1 password authentication bypass in OpenSSH on musl-based Linux without PAM, is actively exploited in the wild. CISA confirmed exploitation and added to KEV with November 12 deadline; GreyNoise recorded first successful wild exploitation November 7, 40 hours after fix release.

Why it matters Unauthenticated SSH access to passwordless accounts including root on Alpine Linux and other musl systems; Alpine is ubiquitous in Kubernetes, container registries, edge infrastructure, and CI/CD—enabling immediate code execution on cloud-native deployments.

CVE-2026-123158 exploits a flaw in OpenSSH sshd on musl-based Linux systems (Alpine, musl libc) built without PAM. When a password exceeds 256 characters, musl's crypt() function returns an asterisk (*) on error. sshd incorrectly treated this asterisk as a valid hash match for accounts with stored hashes under 4 characters, including passwordless accounts and root. An attacker connects via SSH and enters a 257+ character password to log into any account without providing the correct password. OpenSSH maintainers released version 10.6p2 on November 5, 2026 with a workaround (rejecting stored hashes <4 chars); exploit code appeared within 4 hours. Previdian, watchTowr, GreyNoise, and Defused all detected exploitation attempts starting November 6. GreyNoise confirmed first successful exploitation November 7 at 00:46 UTC. Alpine Linux containers are standard in microservices, Kubernetes, edge routers, and container base images. Every Alpine SSH server exposed to untrusted networks faces immediate compromise risk. Organizations must upgrade OpenSSH to 10.6p2 on all Alpine systems immediately; temporarily disable password authentication (PasswordAuthentication no) until patching completes; monitor SSH access logs for connections using oversized passwords; rotate all credentials on affected systems; scan container images and registries for vulnerable Alpine sshd versions.

Read the original at bleepingcomputer.com ↗

SonicWall SSL-VPN flaw lets attackers bypass SAML MFA

SecurityWeek · Nov 6, 2026 · Industry news

CVE-2026-19868 (CVSS 7.1), an authentication enforcement flaw in SonicOS SSL-VPN, allows attackers with valid alternate credentials to bypass SAML authentication and SAML-enforced MFA. Actively exploited; CISA added to KEV with November 20 deadline.

Why it matters Renders multi-factor authentication on VPN ineffective for attackers holding compromised alternate credentials; MFA bypass on perimeter VPN gateways directly enables unauthorized internal network access, defeating the primary control against credential compromise.

CVE-2026-19868 affects SonicWall SonicOS SSL-VPN services when SAML-based authentication is enabled (Gen 7 running SonicOS ≤7.3.3-7015; Gen 8 running ≤8.2.2-8015). An attacker with valid credentials for an alternate authentication method (local LDAP, RADIUS, certificate-based auth) can establish a VPN session without completing SAML verification or MFA challenges enforced by the identity provider. The flaw is not an unauthenticated bypass—attackers require working credentials for at least one authentication method. However, stolen VPN credentials are ubiquitous in breach data; SAML MFA is precisely the control that renders compromised credentials useless. By exploiting this authentication enforcement gap, attackers with single-factor credentials gain full VPN access, bypassing the MFA layer entirely. SonicWall PSIRT investigated reports of potential exploitation and urged customers to patch. CISA confirmed active exploitation and added CVE-2026-19868 to the KEV catalog November 6 with federal deadline November 20. No workarounds exist. Fixed releases: SonicOS 7.3.4-7013 (Gen7) and 8.2.3-8015 (Gen8). Organizations must audit every internet-facing SonicWall instance for SAML configuration; rotate all VPN service accounts and LDAP/RADIUS credentials out-of-band; review authentication logs for SAML bypass patterns (successful logins from SAML-only users via alternate methods); patch all Gen7 and Gen8 appliances to fixed releases immediately.

Read the original at securityweek.com ↗

Palo Alto Networks in talks to buy Clutch for $500M

TechCrunch · Nov 5, 2026 · Industry news

Palo Alto Networks is in acquisition discussions with Clutch Security, an AI agent security startup, for $500M plus milestone payouts. This would be PANW's seventh acquisition in 2026 as the company aggressively expands into AI-driven security.

Why it matters Consolidation of AI agent security reflects market validation of this threat category for enterprise infrastructure teams managing unauthorized agent access.

Palo Alto Networks is negotiating to acquire Clutch Security, a startup specializing in AI agent access control, for $500 million in cash plus performance-based payments. The deal would mark PANW's seventh acquisition in 2026, continuing a pattern of buying innovation in response to AI-driven security threats. Clutch addresses a critical operational gap: controlling what AI agents can access within enterprise systems. This acquisition follows PANW's pattern of acquiring startups across observability (Chronosphere for $2.95B in January 2026), identity security (CyberArk for $25B in February 2026), endpoint posture management (Koi for $231M in April), and AI gateway capabilities (Portkey in May). The company's acquisition strategy explicitly targets capabilities that address customer needs and accelerate their security platform consolidation. For operations practitioners managing AI agents across infrastructure, this represents the major vendor consolidation of an emerging security domain—agent access control is becoming table-stakes for modern security operations.

Read the original at techcrunch.com ↗

People Moves

2 stories

DNSFilter names Justin Foster chief technology officer

Business Wire · Nov 4, 2026 · Vendor release

DNSFilter announced the appointment of Justin Foster as CTO, strengthening the executive team as the company scales its AI-powered DNS-first security platform with the recent launch of LUMINATE by DNSFilter.

Why it matters Enterprise security and network teams adopting DNS-first controls and AI governance now have a CTO with deep R&D experience scaling security products at Forescout and co-founding Cysiv.

DNSFilter, the DNS-first security company powered by AI, announced the appointment of Justin Foster as CTO on November 4, 2026. Foster most recently served as CTO at Forescout and previously co-founded and served as CTO at Cysiv (acquired by Forescout), and held various positions at Trend Micro and Third Brigade. DNSFilter recently launched LUMINATE by DNSFilter at Nasdaq, with CEO Ken Carnesi stating "Justin is the right leader to scale it." The new platform extends the company's DNS-first protection capabilities to include AI governance and controls, behavior analytics, and threat intelligence for the first time. Foster brings product development and R&D experience to DNSFilter's leadership as the company expands from traditional DNS filtering into broader enterprise security and AI governance capabilities.

Read the original at finance.yahoo.com ↗

Vistance Networks CEO Chuck Treadway to step down in February

Light Reading · Nov 5, 2026 · Industry news

Vistance Networks announced that CEO Chuck Treadway will resign from his position and board membership effective February 1, 2027, following a strategic transformation that returned $16.50 per share to shareholders in 2026 and cleared the company's debt.

Why it matters Leadership transition at a major network infrastructure provider signals completion of CommScope restructuring and strategic shift under new management.

Treadway has served as President and CEO since October 2020 and oversaw a significant strategic transformation, including divestitures that unlocked substantial shareholder value, the repayment of all outstanding debt and the redemption of all outstanding preferred equity. He oversaw the company's transformation into a more focused network solutions provider operating through Aurora Networks. Including the $1.50 per share special cash distribution payable November 6, 2026, Vistance returned $16.50 per share to shareholders through special cash distributions during 2026. The board has launched a comprehensive search and retained a leading executive search firm to identify Treadway's successor. This marks a clean handoff after a multiyear portfolio restructuring and debt elimination.

Read the original at lightreading.com ↗

Edgecore releases SONiC AI agent with deterministic guardrails

Edgecore Networks · Nov 9, 2026 · Primary source

Edgecore announced its EDM SONiC AI Agent, developed around a security-first approach combining AI-assisted analysis with deterministic controls, role-based access, human oversight, and controlled network management to help network teams streamline troubleshooting and organize operational evidence.

Why it matters Network operators now have a reference implementation for secure, governance-first autonomous NetOps on disaggregated SONiC hardware, addressing the critical gap between agent capability and regulatory audit requirements.

Edgecore's EDM SONiC AI Agent combines AI-assisted analysis with deterministic controls, role-based access, human oversight, and controlled network management. This announcement signals a critical shift: autonomous agents are no longer abstract concepts but concrete implementations on production-grade, open-source network operating systems used by enterprise and service providers. The security-first design—explicit emphasis on guardrails, human oversight, and controlled execution—reflects the real governance gap that has slowed agent adoption in NetOps. By grounding the agent in SONiC's deterministic switching platform and its device inventory, the architecture avoids the "black box" problem: network engineers retain full visibility and policy control. The white paper accompanying the announcement provides a practical reference for operators evaluating agent-driven operations without sacrificing audit trails or compliance requirements. This is particularly relevant for telecom and financial services networks where agent autonomy has been constrained by regulatory audit demands. The reference to secure agentic NetOps positions SONiC as the foundation for intent-driven, AI-assisted network operations at scale.

Read the original at edge-core.com ↗

Agentic AI operations need an operating model change, not tools

Forbes Technology Council · Nov 6, 2026 · Analysis

After connecting 21+ enterprise data sources to AI systems, practitioners still struggle to bridge signals to context, context to decisions, and decisions to governed execution, because infrastructure relies on fragmented tools, manual investigations, and ticketing queues. Agentic operations require an operating-model transformation, not just technology deployment.

Why it matters The hard part of agent adoption is organizational: moving from CLI/scripts to policy-driven orchestration fundamentally changes how infrastructure teams interact with production systems and make remediation decisions.

The author describes a real customer scenario: after connecting 21 enterprise data sources, the gap remained between data collection and governed action, because infrastructure still relied on fragmented tools, manual investigations, and ticketing workflows. Four fundamental questions emerge: What happened? Why did it happen? What should we do? Who decides?—and infrastructure leaders must recognize that agentic operations require an operating-model transformation, not simply a technology deployment. For network and SRE teams, this reframes the agent adoption conversation: the technical capability to run an agent is necessary but not sufficient. Real adoption requires redesigning handoff workflows, policy decision-making, and escalation paths. The shift from CLI/scripts to prompt engineering and policy-driven orchestration fundamentally changes how teams interact with infrastructure. This applies directly to incident response, where the bottleneck is often not detection or diagnosis but consensus on the right remediation action—exactly what agent-driven operations promise to address by making policy decisions explicit and traceable. Infrastructure leaders must adopt agentic AI to automate complex, nondeterministic operations through natural-language interfaces while maintaining human-in-the-loop governance.

Read the original at forbes.com ↗

Verizon deploys Ericsson intelligent controller for AI-native RAN nationwide

SDxCentral · Nov 3, 2026 · Industry news

Verizon and Ericsson are deploying the Ericsson Intelligent Controller (EIC) nationwide to run AI-native rApps that automate RAN management. Field tests showed gains up to 8.9% in downlink throughput and 8% in uplink UE throughput with closed-loop optimization.

Why it matters Moves AI-RAN from pilot to production platform, enabling operators to deploy their own closed-loop optimization apps at scale and progress toward Level 4 cognitive automation.

Verizon plans nationwide rollout of the EIC by year-end 2026 to support AI-native rApps developed by both Ericsson and Verizon itself. The EIC is a non-real-time RAN Intelligent Controller within Ericsson's Intelligent Automation Platform (EIAP), which supports over 100 rApps from 100+ ecosystem members. Field deployments demonstrated AI cell shaping of massive MIMO and traffic steering delivering measurable performance gains. Verizon's ability to self-develop rApps on this open standards platform signals a fundamental shift in operator autonomy—moving from vendor-locked automation to open, programmable RAN orchestration. This enables network teams to encode their own optimization logic and deploy it across production networks without waiting for vendor releases, accelerating the path to Level 4 autonomy where certain processes are almost fully automated with closed-loop control.

Read the original at sdxcentral.com ↗

Telco AI automation moves from pilots into multi-country production

TelecomTV · Nov 9, 2026 · Industry news

Global telecom operators deployed AI-driven network automation across the US, Bahrain, Thailand, and Ethiopia this week. Verizon's EIC rollout, Thailand's Network Digital Twin simulations, and BT's agentic AI services signal the inflection point from pilots to sustained production operations.

Why it matters Confirms AI-native network automation is transitioning from experimentation to mainstream operational infrastructure, reducing manual NOC workload across multiple geographies and use cases.

Verizon's VP of Network Planning reported expanded collaboration with Ericsson deploying AI cell shaping rApps, demonstrating significant gains in throughput and efficiency. Thailand operator AIS is using Network Digital Twin technology—virtual representations of physical networks that simulate scenarios and identify optimization opportunities before deployment. BT Group's CEO publicly stated confidence in the agentic AI era and confirmed the operator is on track in its AI services business. The convergence across operators (Verizon, AIS, BT) on production deployments of AI agents for RAN optimization, service assurance, and business operations marks the industry-wide inflection point. This week consolidated momentum from isolated proofs-of-concept into orchestrated production rollouts, signaling that network operations teams should expect autonomous AI agents managing fault detection, configuration changes, and optimization decisions as normal operational mode.

Read the original at telecomtv.com ↗

Nokia projects AI will drive 44% of traffic by 2035

TheFastMode · Nov 5, 2026 · Industry news

Nokia projects total network traffic will nearly quadruple from 2026 to 2035, reaching 3,789 exabytes per month, with 44% driven by AI. Agentic AI traffic alone is forecast to grow 160x by 2035, physical AI traffic 105x, fundamentally reshaping network architecture requirements.

Why it matters Network infrastructure teams must begin capacity planning now for edge compute, inter-data-center connectivity, and middleware guaranteeing QoS for agentic workloads alongside user traffic—implications span the next decade of capex decisions.

Nokia's bottom-up traffic model separates consumer, enterprise, and AI workloads, with AI comprising 44% of total bytes by 2035. The 160x explosion in agentic AI traffic—autonomous agents coordinating network operations, application placement, and optimization—signals that network autonomy will dominate traffic patterns. This is not just higher volume but different traffic behavior: agents require predictable latency, deterministic routing, and orchestration guarantees that differ from traditional user flows. Network practitioners should treat this forecast as a design constraint: distributed edge compute must sit at every regional hub, inter-data-center fiber must support dynamic wavelength allocation for AI-to-AI communication, and orchestration middleware must provide intent-based, zero-trust control planes. The implication is a wholesale architectural transition from human-driven network provisioning to AI-driven, self-healing, intent-aligned infrastructure.

Read the original at thefastmode.com ↗

OpenAI closes $11 billion at $1.25 trillion valuation

Bloomberg · Nov 6, 2026 · Industry news

OpenAI closed $11 billion in new funding at a $1.25 trillion valuation on November 6, the first close of a larger round that may reach $20 billion. The company scaled back from its original $30 billion target at $1.4 trillion amid AI stock selloffs and investor concerns about data center construction obstacles following US midterm elections.

Why it matters Market conditions are reshaping mega-round valuations and investor appetite for AI infrastructure spending, with direct implications for enterprise funding prioritization and capital allocation in the sector.

OpenAI closed $11 billion in new funding at a $1.25 trillion valuation with funds wired on Friday, November 6. The round may grow to $20 billion total in coming weeks, substantially down from initial targets of $30 billion at $1.4 trillion. Coatue Management and T. Rowe Price were among investors contributing to the first close. Market sentiment shifted due to concerns about infrastructure spending levels and obstacles to data center construction from midterm election results. OpenAI's value rose approximately $400 billion since March 2026, but this round is smaller and cheaper than originally targeted. Anthropic's upcoming IPO valuations are also being rethought by prospective investors in response, signaling broader revaluation of frontier AI companies and infrastructure costs across the sector.

Read the original at bloomberg.com ↗
Nothing in today's briefing matches that.

Listening

Podcasts and talks

Vendor Radar

Last 7 days · arrows compare with the 7 before

What changed this week

Last 7 days vs the 7 before

Biggest moves

Trending topics

agent · automation · observability · Agentic AI · OpenTelemetry · MCP · RAG · AIOps · SASE · SRE · Digital Twin

Get the daily briefing

Today's 3 things that matter and every story with why it matters, in your inbox each morning. Free, and you can unsubscribe at any time. Prefer a reader? Follow the RSS feed.