CAE has the footprint, credibility, and installed base to lead at global scale ^[1][2]. Over 250 full-flight simulators, 50+ training locations, and relationships with major airlines worldwide give the company a deeply entrenched position in civil aviation ^[1].

The problem is that the next competitive layer is no longer about visual fidelity or motion platform quality. It is the software and evidence layer that connects training nodes, senses what changed in the operating environment, and helps the instructor decide what the trainee should do next ^[11][12].

The most underestimated risk is the ecosystem itself: local training bases, flight schools, and certification-linked operators that build domestic scale before a national champion consolidates the market ^[54]. Companies like Haite have real domestic training-centre scale and Level D operations, letting them shape the market through physical footprint rather than product branding alone.

China Simulation Sciences (CSS) is the most visible export signal. CSS delivered its first EASA Level D full-flight simulator to Simaero in Paris in late 2025 ^[53]. With official CAAC and EASA qualification claims, CSS operates not just as a Chinese simulator vendor, but as a bridge to the next low-altitude eVTOL training market.

CAE already has the high-end device estate, the training centers, the instructors, the records systems, and a serious software stack above the simulator ^[1][6][7].

Rise is already a serious event-intelligence system. The defense datasheet says CAE Rise has been used in more than 50,000 unique training sessions by over 7,100 pilots across sixteen aircraft types and more than 120 simulators, with live coaching, a synthetic Instructor Pilot, automated grading, replay, eGrading and Records Manager views, biometrics, and centralized training-plan management ^{[61][71][72][75]}.

Connect is already a serious operations layer. In the current business-aviation datasheet it handles reservations, user management, document exchange, and instructor access. In CAE's filings it is framed more broadly as the digital layer coordinating instructors, simulators, classrooms, and courseware across the network ^[1][70][74]. Training Management and Deployment System (TMDS) already covers training management, qualifications, scheduling, and administrative records ^[73].

CAE also already has adaptive training and connected synthetic environments in place through CORe training, TRAX Academy, distributed mission training, and synthetic environments ^{[11][12][73][75][77][78][79][80]}.

Those layers already do real work. Rise explains and coaches the event. Connect gets people and resources into the system. Training Management and Deployment System (TMDS) holds the formal record. CORe training and TRAX Academy adapt parts of the syllabus. Distributed mission training and synthetic environments network the environment itself ^{[11][12][70][73][74][75][77][78][79][80]}.

However, that still is not the same thing as one persistent training memory. The user still does not have one profile that follows them for years. The instructor still does not start every session with one clean carry-forward record. The organization still does not get one lasting human and team memory across devices, sites, and mission contexts.

The weakness becomes much more obvious once training stops being individual. The U.S. Naval Safety Command described repeated breakdowns in communication, situational awareness, and teamwork between bridge, combat information centre (CIC), and navigation teams ^[68]. The United Kingdom Defence Science and Technology Laboratory (Dstl) asked for persistent records across unrelated events and better visibility of individual performance inside collective exercises ^[69]. The U.S. Army and ST Engineering both point toward shared mission rehearsal environments rather than isolated trainers, while ASTi's SERA product shows the same logic in airspace training with multiple human pilots inside one simulated traffic environment ^{[13][14][64][67][76]}.

That is why the missing layer cannot just connect devices. It has to keep one live exercise coherent across roles, then preserve team context, mission memory, and what actually changed from one event to the next.

The gap is not that CAE has no software above the simulator. The gap is that those software layers still hand the problem off to one another ^{[11][12][70][73][74][75][77][78][79][80]}.

The real opening is the layer that keeps a real person, a real crew, and a real exercise coherent across time. That means one profile for the user, one shared state for the team while the event is running, and one carry-forward memory after the event ends ^{[68][69][75][76]}.

Not every important failure mode belongs at the extremes of CAE's device mix. The 7000XR remains the qualification anchor and XR devices are increasingly useful for access, familiarization, and repetition ^{[6][7][22][23]}. But sensory failure, vestibular mismatch, unusual-attitude recognition, and degraded-cue recovery still sit awkwardly between those tiers ^[39][44][81].

CAE's 700MR already points in the right direction for military helicopter training: a task-specific device tier built around mission conditions, not just a generic realism ladder ^[82]. That is why a compact electric-motion pod is believable as a first hardware shift. It is not the whole answer. It is a focused answer to a focused training problem ^{[6][39][44][82]}.