Google unveils a public preview of Google Cloud Armor Adaptive Protection, a ML-powered service for Google Cloud apps and sites to protect against DDoS attacks
Context & Ripple Effects
This preview extends a line Google started years earlier, when it announced DDoS protections and new security dashboards for Google Cloud Platform and G Suite in its 2018 security push. What changed with Adaptive Protection is the mechanism: instead of static defenses, Google is applying machine learning trained on its traffic to detect attacks on customer apps and sites automatically.
The move also reads as an early template for where Google took cloud security afterward — folding learned detection into managed services, as seen later in Google Threat Intelligence built on Gemini and the eventual Unified Security bundle that pulls cloud security, threat intelligence, and Mandiant expertise into one offering.
First-order effects
- Google Cloud customers running apps and sites gain automatic, ML-based DDoS detection inside Cloud Armor during the public preview, rather than having to tune protection rules themselves.
Second-order effects
- Embedding learned detection into a managed edge service raises the baseline rivals' cloud security offerings must match, pushing DDoS defense from a configurable feature toward a default platform capability.
- The same ML-detection pattern gives Google reusable machinery it later productized across its portfolio — from the Gemini-based threat intelligence work to broader bundled offerings.
Third-order effects
- If the pattern holds, cloud providers consolidate around AI-native security layers sold as integrated suites rather than point products — the direction Google itself followed from Adaptive Protection through Unified Security to the AI security agents debuted with Wiz at Next '26.
- Attackers automating exploits and defenders automating detection push both sides toward machine-speed engagement, making the quality of a provider's traffic and threat data the durable differentiator in cloud security.
The trend: Cloud security is steadily shifting from manually configured rules to machine-learned detection embedded directly into the platform provider's own services.