Cisco Investments backs startups pushing the boundaries of innovation in areas that matter to our customers. Each month, we spotlight how our portfolio companies are advancing the technologies shaping tomorrow’s enterprise.
New Investments
Elotl delivers AI-native workload scheduling and orchestration for Kubernetes environments, enabling organizations to dynamically deploy compute-intensive applications wherever GPU resources are available. Their technology empowers enterprises to optimize efficiency and cost while scaling AI workloads seamlessly across clouds and on-premises infrastructure. Learn more about our investment in Elotl in our blog, “Elotl: Harnessing Kubernetes Compute to Power AI Workloads.”
Artificial Intelligence
Cohere launched North, an enterprise-ready agentic AI platform that enables organizations with strict data security needs to deploy AI agents and automations at scale within their own infrastructure. It empowers teams to harness modern workplace intelligence, like connected search, chat, and workflow automations, while maintaining full control and privacy over their data.
Groq shared the inner workings of their Language Processing Unit (LPU) to reveal how it achieves ultra-fast inference by eliminating architectural bottlenecks and preserving model quality. It leverages innovations like TruePoint numerics, SRAM-based memory, and a software-scheduled network to deliver up to 40× performance boosts without trading off accuracy.
Collaboration
Theta Lake won Best Compliance Solution in the UC Today Awards 2025, and received a highly commended placement in the category of Best Use of AI.
Customer Experience
Fiddler delivered key insights from an AI strategist, highlighting why agentic AI adopters often stall in the “trough of disillusionment.” They emphasize that architectural trade‑offs, like choosing single‑agent versus multi‑agent designs, and robust observability are essential for turning agent prototypes into mission‑critical production systems.
Maven AGI introduced Maven Voice, a voice AI agent for live customer support calls that understands context and responds naturally in real time. It allows organizations to transform phone support into humanlike conversations by adapting tone, managing interruptions, and integrating with telephony and CRM systems to deliver trusted, empathetic experiences at scale.
Networking
Lightbits Labs raised the bar again in the MLPerf Storage v2.0 benchmark, showing performance gains across ResNet‑50, 3D U‑Net, and CosmoFlow models using just three commodity servers. They demonstrate that their software‑defined storage can outperform specialized systems by optimizing through software rather than scaling hardware.
Quantum
Qunnect unveiled Carina, the first turnkey quantum entanglement networking system, proven with metro-scale deployments over standard telecom fiber. It delivers secure, real-time quantum communication without specialized infrastructure, demonstrating fidelity, error correction, and stability in real-world environments.
Security
BigID outlined a practical guide for building AI governance programs, highlighting discovery, risk classification, and ongoing monitoring as foundational pillars. It helps organizations reduce risk, strengthen compliance, and build trust as they scale AI initiatives.
Expel published key insights from its Q2 2025 Quarterly Threat Report, showing that identity-based attacks accounted for 67.6% of incidents, non‑targeted malware fell significantly, and cloud infrastructure–related incidents rose to 2.5%. They highlight industry trends, noting that financial services had the highest volume of incidents at 12.6%, while manufacturing faced the most identity-focused threats, three out of every four incidents.
Orca Security expanded runtime protection to include hybrid, private, and on-prem environments, delivering unified runtime security with expanded detections and AI-assisted threat investigation. It lets security teams maintain consistent protection and accelerate response across complex infrastructures using a single, integrated platform.
Query introduced two new tools: Query Agents and Security Data Pipelines, designed to eliminate ETL and build structured, decision-ready security data workflows. They streamline how security teams ingest, normalize, and act on data directly in their cloud storage or SIEM, accelerating incident response with less engineering overhead.
SpecterOps rolled out BloodHound v8.0 and OpenGraph, expanding attack path management beyond Active Directory to ingest identity risk across platforms like GitHub, Snowflake, and SQL Server. It allows security teams to build custom threat models, visualize hybrid attack paths, and surface identity risks throughout the technology stack.
SGNL launched Continuous Identity Sessions, a new virtual series for identity and security professionals navigating the realities of modern IAM. There will be two episodes monthly, providing real insight, and highlighting what good Continuous Identity looks like in practice.
VISO TRUST upgraded its assessment reports with Smart Summaries, an AI-powered format that generates cleaner, clearer, and more shareable risk insights by including relationship context, artifact findings, and control evaluations. It lets teams edit summaries inline and export branded, audit-ready PDFs, making stakeholder communication faster, flexible, and more polished.