TLDR IT 2026-09-18

We think Auvik will show you something about your network you didn't know.

Prove us wrong.

Just start an Auvik trial and watch it automatically discover devices and map your network in under an hour.

Worst case? You validate your network.Best case? You discover something worth knowing.

Either way, you can get a free $301 Dream Router 7.

It's a bet worth taking (this month only, terms apply).

Take the $301 bet →

Huawei moved up the launch of its next-generation Ascend 960DT AI chip to Q1 2027 as it tries to close the gap with Nvidia. The company says demand for its AI compute hardware in China is already exceeding supply.

Gartner expects worldwide AI spending to rise 49.5% this year to $2.7 trillion. The forecast reflects continued spending on AI infrastructure, devices, software, and services as companies move projects into production.

OpenAI disclosed cases where models tried to evade oversight, conceal mistakes, and work around constraints during testing. It is introducing a framework for systematically tracking and publishing these incidents as increasingly autonomous agents create new monitoring challenges.

According to Gartner, traditional centralized cloud architectures are holding back the agentic AI revolution by failing to provide the continuous, high-frequency inference needed for real-time enterprise workflows. Infrastructure and operations leaders need to adopt a hybrid edge-cloud architecture to overcome severe latency barriers, unsustainable hyperscaler bandwidth egress costs, and outage risks.

Microsoft says its own AI rollout showed that deploying tools broadly is not enough. Companies need to redesign entire workflows around business outcomes. Proprietary context, evaluations, and organizational learning could become more valuable than access to any individual foundation model.

As agents perform longer tasks, they increasingly need reliable access to enterprise data, past decisions, and memory across many retrieval steps. AWS is positioning OpenSearch as part of that context layer while enterprises grapple with retrieval errors, governance, and rising infrastructure costs.

Every new monitoring tool, service, or "quick integration" adds another dashboard to the pile—and before you know it, you're spending more time switching between tools than actually building. With Microsoft Azure, your entire stack lives in one workspace: infrastructure, data, AI, and more. See how Azure brings it all together →

Sourcegraph has launched Agentic Batch Changes, a frontier coding agent designed to execute codebase-wide modifications across hundreds or thousands of repositories. The tool combines Batch Changes with the Deep Search code analysis agent, allowing engineers to manage large-scale migrations, security remediations, and dependency upgrades using plain-language prompts. Sourcegraph is offering the feature with outcome-based pricing tied to successfully merged changesets.

Dataminr Advanced for Corporate Security introduces agentic AI capabilities designed to automatically corroborate, contextualize, and predict security threats. The release marks a transition from real-time alerting to Autonomous Real-Time Intelligence, utilizing fine-tuned LLMs trained on a proprietary event archive rather than general-purpose frontier models.

Claude Code Projects gives developers a persistent workspace that can remember context and delegate work across long-running software projects. The launch pushes coding agents beyond individual sessions toward continuously operating development environments.

Google Cloud and Salesforce are expanding their partnership to connect Gemini Enterprise, Agentforce, Hyperforce, Tableau, and Salesforce data. The integrations are designed to let enterprise agents work across the two companies' platforms with less custom integration.

There are two dominant architectural patterns for AI gateways in production inference: access gateways that route outbound requests from applications to multiple model providers, and serving gateways that manage inbound customer traffic for organizations operating their own models. Serving gateways act as an essential control layer for identity, routing, protection, metering, and auditing when exposing proprietary models commercially.

The Graph RAG pipeline achieves significant reductions in API hallucination rates and improvements in dependency resolution, though it trades off some cyclomatic complexity and docstring preservation.

The outage was caused by an internal login service that consumed excessive server resources while waiting for responses.

Google's Dream-RSI reuses previous search trajectories to test new agent strategies, reducing discovery-agent calls by as much as 162x and potentially lowering agent experimentation costs.