Operations | Monitoring | ITSM | DevOps | Cloud

AI gross margin is what remains of SaaS profitability after inference, model routing, and AI infrastructure land in cost of revenue. The numbers have moved. AI products averaged 45% gross margin in 2025 and are projected near 53% in 2026, against the 70% to 85% that SaaS built its valuations on. The compression is real, measurable, and manageable for companies that can see their cost to serve.

The recent mystery around Ox Alpha last week and the week before was a fun slice of what makes social media fun. The hunt for the model provider and how people did that discovery should be studied. But, this post is more about stealth models in general. You might be wondering what the phrase “stealth model” even means and if you already know, you might still be curious about why companies release stealth models.

Canonical is now part of the Open Secure AI Alliance, announced by NVIDIA with partners across cloud computing, cybersecurity, enterprise software, open source foundations, and AI research. The alliance aims to develop and share open technologies, techniques, and tools to safeguard software and agents in the age of AI.

Codex is OpenAI's terminal-based autonomous coding agent, bundled into ChatGPT plans and billed through token-based credits. Claude Code is Anthropic's terminal-based coding agent, sold through Claude subscriptions or the API. Codex is cheaper to enter; Claude Code is stronger on complex work. Both run $100 to $250 per developer monthly at active use.

Tuning guardrails can take time. To accelerate this, we combined the power of Claude Code, Coralogix Telemetry and the CX CLI, to build a loop that iterates on guardrails. To test our loop, we built a fictional chat interface for an iGaming operator, that took bets, gave odds on markets and more, but with a catch: this agent was not permitted to make predictions about where the market was going. With our goal in place, we got to work designing our loop.

Across the global landscape, telecommunications operators have already deployed machine learning for predictive maintenance, customer care chatbots, and anomaly detection. However, as the industry transitions toward Autonomous Networks Level 4 (AN L4), where networks can make intent-driven, predictive decisions and perform closed-loop management with minimal human intervention, a fundamental architectural challenge has emerged.