AI agents break infra: GitLab cuts 14%, Coralogix raises $200M

AI agents are breaking the developer infrastructure they were built to enhance. GitLab cut 14% of its staff, roughly 350 employees, and exited 22 countries as part of a restructuring specifically designed to rebuild its platform for AI workloads. The move comes as CEO Bill Staples acknowledged that agentic workloads are stressing developer infrastructure beyond its original design. GitHub encountered the same problem: AI-powered submissions disrupted uptime on a platform used by tens of millions of developers worldwide. The stress is structural, not superficial. Agentic workflows generate event volumes, API call frequencies, and state complexity that traditional DevOps platforms were never architected to handle. At the same time, capital is flooding into the new layer of "agent observability." Coralogix raised $200 million in Series F funding at a $1.6 billion post-money valuation, led by Advent and Canada Pension Plan Investment Board, bringing total funding to $550 million. The round came just 11 months after the company's $115 million Series E, a pace of fundraising that reflects investor urgency. Coralogix is building monitoring and observability tools specifically for AI agents in production, a market that barely existed 18 months ago. Cisco joined the race on June 2 at Cisco Live US, unveiling Cisco Cloud Control, a platform built for humans and AI agents to manage, monitor, and defend critical IT infrastructure. The product forms the foundation of Cisco's AgenticOps operating model. Both incumbents and startups are racing to build the management layer for AI agents, and the winners will determine how the next generation of software is built, deployed, and secured.

GitLab rebuilds for agentic workloads

GitLab's restructuring is not a cost-cutting exercise. It is a fundamental rebuild of the company's architecture for an AI-first world. The 14% staff reduction, which eliminated roughly 350 positions, came alongside a decision to exit operations in 22 countries and flatten management layers. CEO Bill Staples framed the move as a response to a structural shift: agentic workloads are stressing developer infrastructure beyond its original design. GitLab is now partnering with an unnamed AI lab to rebuild its APIs specifically for AI agents, meaning the company is re-engineering how its core product interacts with the outside world. The restructuring is a direct admission that the existing developer toolchain, built for human developers writing code manually, cannot simply be augmented with AI features. It requires a complete rethinking of how the platform handles authentication, rate limiting, event streaming, and state management when the primary users are autonomous agents rather than humans. The 22-country exit suggests GitLab is consolidating its go-to-market operations to focus on core markets where it can deliver the new AI-native platform most effectively. The broader context is brutal: the tech industry has cut over 100,000 jobs in 2026, and GitLab's move signals that even companies with strong product-market fit must rebuild from the ground up to survive the agentic shift. The company is betting that a leaner, more focused organization can move faster than competitors still weighed down by legacy operations.

Coralogix raises $200M for agent observability

Coralogix raised $200 million in Series F funding at a $1.6 billion post-money valuation, led by Advent and CPPIB, bringing total funding to $550 million. The round comes just 11 months after the company raised a $115 million Series E, reflecting the urgency investors see in the agent observability market. Coralogix is building monitoring and observability tools specifically for AI agents in production, a category that barely existed two years ago. The company is working toward profitability and does not expect to raise more capital, so this round is intended to fund a land-grab rather than survival. The thesis is straightforward: as enterprises deploy AI agents to automate software development, customer support, and IT operations, they need a new class of tools to monitor agent behavior, detect failures, and debug complex multi-agent workflows. Traditional observability platforms like Datadog and New Relic were built for human-driven applications, not for autonomous agents that can spawn sub-agents, make decisions, and execute actions without human intervention. Coralogix's $550 million total funding gives it the resources to build the category-defining product, but it faces competition from both established observability vendors and new entrants like Cisco.

Cisco launches AgenticOps at Cisco Live

Cisco unveiled Cisco Cloud Control at Cisco Live US on June 2, 2026, marking the company's entry into the agent management market. The platform is built for both humans and AI agents to manage, monitor, and defend critical IT infrastructure, and it forms the foundation of Cisco's AgenticOps operating model. Cisco is positioning Cloud Control as the management layer for enterprise AI agents, directly competing with startups like Coralogix. The product is available in Controlled Availability in the US immediately, with Global Availability to follow. Cisco's advantage is its existing enterprise relationships and its deep integration with network infrastructure, security, and IT operations. The company already manages the networks that connect AI agents to their data sources and execution environments. Cloud Control extends that reach into the agent layer itself, giving Cisco the ability to monitor agent behavior, enforce security policies, and manage agent lifecycles. The move is a direct response to the same trend that drove GitLab's restructuring and Coralogix's fundraising: AI agents are becoming first-class citizens in enterprise IT, and they require dedicated infrastructure. Cisco's AgenticOps framework is a bet that enterprises will prefer a single vendor for agent management rather than stitching together multiple point solutions. The timing is aggressive: Cisco launched the product while the agent observability market is still being defined.

GitLab, GitHub, and the agentic developer toolchain

GitLab's restructuring has direct implications for its competitive position against GitHub. Both platforms are struggling with the same problem: AI-powered submissions are stressing developer infrastructure beyond its original design. GitHub has also experienced uptime issues related to AI-powered submissions, and both companies are racing to rebuild their platforms for agentic workloads. GitLab's decision to partner with an AI lab to rebuild its APIs for agents shows the company is taking a more radical approach than simply adding AI features to its existing product. The partnership implies GitLab is re-architecting its API layer to handle the unique demands of agentic workflows: high-frequency event streams, autonomous code reviews, and automated pipeline management. An AI agent running continuous integration jobs generates API call volumes orders of magnitude higher than a human developer committing code manually. Existing rate-limiting logic, webhook delivery systems, and pipeline scheduling engines were sized for human-paced interactions, not for agents that can trigger hundreds of pipeline runs per hour. This creates an opportunity for new entrants. If GitLab and GitHub cannot adapt quickly enough, enterprises will turn to alternative platforms built from the ground up for agentic development. The 22-country exit also reshapes the competitive landscape: GitLab is concentrating its sales and support resources in core markets, which could create openings for regional competitors in the markets it is leaving. For enterprises, the choice between GitLab and GitHub is no longer just about features and pricing. It is about which platform can handle the scale and complexity of agentic development without breaking.

Downstream effects on hyperscalers, fabs, and buyers

The shift to agentic infrastructure has second-order effects across the entire technology stack. Hyperscalers like AWS, Microsoft Azure, and Google Cloud will benefit from increased demand for compute, storage, and networking as AI agents generate more data and require more processing power. But they also face pressure to build agent-native services that compete with platforms like Cisco Cloud Control and Coralogix. The semiconductor supply chain will feel the impact as well: AI agents require more inference compute than traditional applications, driving demand for GPUs, TPUs, and specialized AI chips. Fabs like TSMC and Samsung will need to allocate more capacity to AI inference chips, potentially squeezing supply for other segments. Enterprise buyers face a complex decision: they must choose between building agent management capabilities in-house, buying from established vendors like Cisco, or adopting startups like Coralogix. The choice will have long-term implications for vendor lock-in, security, and operational efficiency. The regulatory environment is also shifting: as AI agents become more autonomous, regulators will demand greater visibility into agent behavior, which will drive demand for observability and management tools. Companies that can provide auditable agent logs and compliance reporting will have a competitive advantage in regulated industries such as financial services, healthcare, and defense. The downstream effects of the agentic shift will ripple through the entire technology industry for years. Security vendors will need to rethink threat models that assume humans are the primary actors in software systems. Data platform vendors will face new ingestion and retention requirements as agent telemetry volumes dwarf traditional application logs. The entire software supply chain, from development to deployment to operations, is being re-priced around the assumption that AI agents will be the primary consumers of developer infrastructure within three to five years.

The agentic infrastructure market is still in its earliest stages, but the direction is clear. GitLab's restructuring, Coralogix's fundraising, and Cisco's product launch are all signals that the industry is moving from experimentation to production. The next 12 to 18 months will determine which companies own the agent management layer, and the stakes are enormous. Enterprises that bet on the wrong platform will face costly migrations and operational disruptions. The winners will be those that can provide a unified, secure, and scalable platform for managing AI agents across the entire lifecycle from development to production to retirement. Cisco has the enterprise relationships and the existing infrastructure footprint. Coralogix has the startup agility and the venture capital war chest. GitLab and GitHub are fighting to defend their developer toolchain dominance. The outcome will reshape the software industry for the next decade. Enterprises that choose wisely now will compound that advantage across every subsequent generation of AI tooling.