The brain drain narrative is real but overstated. Austria retains nearly seven in ten AI career trajectories, and cohort drain rates are actually falling, from 20 % (2018 cohort) to 17 % (2022 cohort). But the risk concentrates precisely where it matters most: Build-tier drain hits 18 %, NLP / Generative AI reaches 26 %, and half of all departures happen in the first 12 months.
EXHIBIT 3.1
Retention overview — the big picture
67.4%
retained
Source: Revelio Labs via WRDS
Key findings
- 67.4% of AI trajectories are retained — above the European median for small open economies.
- Domestic churn (14.2%) is nearly as large as international outflow (13.1%) — Austria has a functioning internal AI labour market.
- Core AI and Full AI retention are near-identical (67.9% vs 67.4%) — no hidden leak in peripheral roles.
Source: Revelio Labs via WRDS · Chapter 3
EXHIBIT 3.2
Where they go — the geography of outflow
Source: Revelio Labs via WRDS
Key findings
- Germany dominates by volume (522, 23% of all brain drain) — DACH integration, not pay.
- US (188) and Switzerland (178) are the compensation magnets; harder to retain without matching.
- The long tail (UK, FR, ES, NL, IT) absorbs ~440 departures at modest or negative salary changes.
Source: Revelio Labs via WRDS · Chapter 3
EXHIBIT 3.3
The economics of leaving — salary by destination
Source: Revelio Labs via WRDS
Key findings
- Germany is an ecosystem pull, not a pay pull — +5% median uplift, only 53% get a raise.
- US and Switzerland are clear compensation-sensitive corridors — +33% each, 72–76% get raises.
- UK and France come with median pay cuts — career optionality, not money, drives those moves.
Source: Revelio Labs via WRDS · Chapter 3
EXHIBIT 3.4
Who leaves — the tier lens on brain drain
Source: Revelio Labs via WRDS
Key findings
- Build: 18.0% — the most strategically valuable cohort leaks at the highest rate.
- Enable: 10.2% — structurally anchored; infrastructure doesn't transfer across borders.
- Integrate: 14.7% — AI-business interface roles are internationally substitutable.
Source: Revelio Labs via WRDS · Chapter 3
EXHIBIT 3.5
Frontier vulnerability — drain by subcategory
Source: Revelio Labs via WRDS
Key findings
- NLP / Generative AI: 26.1% — the capability Austria can least afford to lose, on its smallest base (~120).
- Computer Vision (18.4%) and Core ML Research (17.7%) also well above the 13.1% average.
- Enterprise categories (BI, DS, Infrastructure) lose more people in absolute terms but at lower rates.
Source: Revelio Labs via WRDS · Chapter 3
EXHIBIT 3.6
The mid-career battleground — seniority of leavers
Source: Revelio Labs via WRDS
Key findings
- Analyst (30%) and Manager (20%) together account for half of all brain drain.
- Median brain-drain seniority is 2.0 (Analyst) vs 3.0 for stayers — a pre-anchoring phenomenon.
- Senior-leadership exits are smaller in volume (~20%) but each one removes institutional knowledge.
Source: Revelio Labs via WRDS · Chapter 3
EXHIBIT 3.7
The first-year window — when brain drain happens
Source: Revelio Labs via WRDS
Key findings
- 49.9% of brain drain happens within the first year — start retention on day one.
- Only 3.8% of movers stay 5+ years before leaving — if you hold them past year 3, they're yours.
- Build-tier median tenure before leaving is just 0.75 years — the frontier leaves before year one is over.
Source: Revelio Labs via WRDS · Chapter 3
Methodology
Revelio Labs workforce data via WRDS. Brain drain = next observed position outside Austria. Domestic churn = next position inside Austria, different employer. Coverage caveat applies (Revelio is not a census). 2025 transitions remain preliminary. Primary segment: austria_located (17,348 trajectories); Core AI sample (Build+Enable+Integrate) 12,350.