Findings summary

The hardware-first concerns_lantronix rule has now been executed across the dataset, not just drafted in source.

Final executed result:

• total companies reviewed: 125
• concerns_lantronix=true before: 60
• concerns_lantronix=true after executed rerun: 56
• net change: -4

This executed result supersedes the earlier pre-rerun estimate in the first reanalysis artifact.

What changed

The new rule is now active in the dataset:

• pure-play application software companies are not Lantronix relevant
• relevant companies must be hardware-related
• narrow exception allowed for lower-level / embedded / autonomy / navigation / payload-adjacent software

Confirmed flips to false include:

• DroneSense by Versaterm
• KBR
• RIIS, LLC.
• Zephyrs Ai, Inc
• TruWeather Solutions
• DroneDeploy
• Kespry
• PrecisionHawk

Confirmed true corrections include:

• Honeywell
• Northwest UAV
• Kratos Unmanned Aerial Systems, Inc.

Important note

The earlier reanalysis artifact captured the rule change correctly, but its after-count reflected the pre-execution estimate. The actual script rerun completed successfully afterward, and the final executed after-count is 56.

Recommendation

Use the executed 56-company result as the current baseline for Lantronix-relevant companies going forward. If Pete wants the exception narrowed further, the first borderline names to revisit are Flyhound, General Dynamics Mission Systems, and Technology Service Corporation.