Check Point has launched the Maestro Hyperscale Orchestrator 200, a firewall system built to inspect AI traffic at data centre scale without slowing it down.
Key Takeaways
- The MHO 200 quadruples total fabric capacity to 12.8 Tbps and threat prevention throughput to 4.5 Tbps, up from the previous MHO 175.
- It adds 400GbE ports and keeps sessions live through failures with automatic failover across two sites, targeting 99.999% uptime.
- Check Point ranked first in the Miercom 2026 Hybrid Mesh Network Security Benchmark with 99.8% security effectiveness, for the fourth year running.
The problem: AI traffic is growing faster than firewalls can inspect it
Check Point Software Technologies says the new Maestro Hyperscale Orchestrator (MHO) 200 is meant to solve a specific trade-off: as AI training, inference and agentic workloads push more east-west traffic through data centres, much of it encrypted, many organisations end up switching off advanced threat prevention just to keep performance acceptable, or buying oversized appliances to absorb traffic spikes that leave capacity idle most of the time.
“AI is forcing a rethink of network capacity planning. Training, inference and agentic workloads are driving east-west traffic inside the data center to levels most security architectures were not designed for, and much of it is encrypted,” said Pete Finalle, Research Manager, Security and Trust at IDC.
Four times the capacity, same footprint
The MHO 200 combines firewalls of different models and generations into a single system that scales with demand, rather than being tied to dedicated chassis hardware. Compared with its predecessor, the MHO 175, it raises total fabric capacity from 3.2 Tbps to 12.8 Tbps and threat prevention throughput from 1.1 Tbps to 4.5 Tbps, in the same 1RU form factor. It can also be split into up to eight independent policy domains, each with its own rules.
“Too many organizations are still forced to choose between performance and prevention. With the MHO 200, they can inspect all their traffic at AI speed, grow capacity as demand rises, and keep their operations running through failures,” said Nataly Kremer, Chief Product Officer at Check Point Software Technologies.
Who it’s built for
Check Point is pitching the MHO 200 at neoclouds and GPU-as-a-service providers, colocation operators, telcos protecting their backbone, and government bodies building sovereign AI infrastructure, all of which are moving toward 400GbE as the new baseline for high-end data centres and AI factories. The system can discover and secure AI applications, large language models, agents and Model Context Protocol (MCP) servers directly from the firewall, and integrates with NVIDIA BlueField data processing units for private and sovereign AI projects.
The MHO 200 is available now on Check Point’s product catalogue for all regions.
