Phase Transitions: The 8 Big Shifts
Structural changes where the old way of doing something becomes obsolete and a new paradigm takes over. These are not incremental improvements -- they are phase transitions.
These Shifts Interact as a System
No shift operates in isolation. The Bug Harvest accelerates Agent Cyber and Custom Software. CCA Mass depends on Lights-Out Manufacturing and AI Decision Authority. Delays in one shift cascade to others. Understanding these connections is essential to prioritizing investment.
5
3
0
0
16
Shift Network Diagram
Hover to highlight connections. Click a node to see full details below. Arrow direction shows causal flow. Line weight indicates connection strength.
Shift 1: Cyber Goes Agent-First
2025 - 2030
Vulnerability discovery, exploitation, and defense all become agent-driven. Massive bug harvest in legacy systems, then equilibrium as both sides operate at machine speed.
Key Metrics
- CVE discovery rate
- Time-to-patch
- SOC automation %
Observable Indicators
- CVE spike 5-10x
- New-code CVE rate drops
- Machine-speed SOCs operational
Related Tripwires
Attribution in threat intelligence of AI-discovered vulnerability
Connections
Enables:
Depends on:
Analyst Assessments
Offense currently finds faster than defense patches. Asymmetry period is 2-4 years.
Legacy systems become operational liabilities on accelerated timeline.
Temporal Overlap
All 8 shifts shown as horizontal time bars. The red dashed line marks the current year. Multiple shifts overlap in the 2028-2030 period -- the window of maximum simultaneous change.
Cascade Analysis
Most Connected Shifts
Total inbound + outbound links
Gating Shifts
Enable the most downstream shifts
Delays in gating shifts cascade to all dependent shifts.
Most Dependent Shifts
Rely on the most upstream enablers
Dependent shifts are at highest risk of cascading delays.
All Shifts
Data Backing
Key database metrics supporting each shift. Data from platform economics, manufacturing AI cases, and decision AI economics tables.
Shift 1: Agent Cyber
No matching database records.
Shift 2: Custom SW
No matching database records.
Shift 3: Bug Harvest
- CVE data tracked externally
- AI-discovered vulnerability ratio increasing
Shift 4: AI Decides
- AI Logistics Optimization: out-of-loop (0x speed)
- JADC2 AI Decision Support: in-loop (0x speed)
- Project Maven / AI-enabled ISR: in-loop (0x speed)
- Autonomous Counter-UAS: out-of-loop (0x speed)
Shift 5: Pred Maint
- Boeing Apache Attack Helicopter Predictive Maintenance: 52.9% ROI
Shift 6: CCA Mass
- XQ-67A OBSS: $3M unit cost (0.04x ratio)
- XQ-58A Valkyrie: $15M unit cost (0.11x ratio)
- XQ-58 Valkyrie: $15M unit cost (0.04x ratio)
- FH-97 (China): $20M unit cost
Shift 7: AI Intel
- Project Maven / AI-enabled ISR: 0% accuracy gain
- Project Maven AWCFT: 35% accuracy gain
- Autonomous BDA System: 25% accuracy gain
Shift 8: Lights-Out
- Boeing Apache Attack Helicopter Predictive Maintenance: 52.9% ROI, 23mo payback
- L3Harris Electronic Warfare Systems AI-Enabled Test: 51.6% ROI, 23mo payback
- Kratos Defense XQ-58 Valkyrie Rapid Production Line: 50% ROI, 24mo payback
- Pratt & Whitney Advanced Coating Inspection AI: 50% ROI, 24mo payback
Shift Summary
All 8 phase transitions at a glance
| # | Shift | Timeline | Status | Connections | Tripwires | Key Metric |
|---|---|---|---|---|---|---|
| 1 | Agent Cyber | 2025-2030 | underway | 2 | 1 | CVE discovery rate |
| 2 | Custom SW | 2026-2032 | emerging | 2 | 2 | SW delivery velocity |
| 3 | Bug Harvest | 2025-2028 | underway | 2 | 1 | CVE count by year |
| 4 | AI Decides | 2027-2035 | emerging | 2 | 3 | Oversight level distribution |
| 5 | Pred Maint | 2027-2032 | emerging | 2 | 2 | MCR delta (AI vs traditional) |
| 6 | CCA Mass | 2028-2035 | emerging | 2 | 5 | CCA inventory |
| 7 | AI Intel | 2025-2030 | underway | 2 | 1 | Analyst-to-collection ratio |
| 8 | Lights-Out | 2028-2035 | emerging | 2 | 4 | Labor hours per unit |
The Window of Maximum Simultaneous Change: 2028-2030
Six of eight shifts overlap in the 2028-2030 timeframe. This is when the cumulative effect of agent-driven cyber, custom software, AI decision-making, predictive maintenance, CCA mass production, and lights-out manufacturing could compound into structural transformation of defense. Decisions made now determine whether the US is ready.