Translator On Kingdom Come 2 Claims He Was Replaced With AI

On March 27, 2026 Warhorse Studios informed English editor/voice-over director Max Hejtmánek that his position would become obsolete as the studio will use AI for all translations going forward, ending a role he held since 2022. The decision—backed by senior leadership public support for generative AI—creates reputational and workforce-risk issues for the studio but is unlikely to have material near-term financial impact.

Analysis

The operational impulse to replace human localization with generative models creates a classic cost-vs-quality trade-off that plays out on three time horizons. In the first 0–6 months you get visible savings on payroll and faster turnaround for builds, which reduces near-term OpEx but raises QA and customer-support costs as tonal/idiomatic errors surface; historical analogs (automated moderation, auto-captioning) show a 2–6% uplift in short-term margins offset by a 5–12% rise in rework and PR mitigation spend in the first year.

Over 6–24 months the second-order effect is vendor consolidation around inference platforms and transfer learning adaptation: studios that standardize on a single inference stack will concentrate GPU demand (benefitting high-end inference hardware and cloud GPU suppliers) while independent localization firms will either vertically integrate AI tooling or compress margins by 20–40%. This sets up a wave of tuck-in M&A where platform owners buy specialized language-model fine-tuning shops to reclaim quality control and IP over creative assets.

On a 2–5 year view, reputational and legal risks crystallize for companies that prioritize savings over authenticity—expect class-action or contractual warranty disputes in jurisdictions with strong moral-rights or employment protections, and greater unionization pressure in creative guilds. Conversely, firms that pair human-in-the-loop workflows with model inference can capture both the cost benefits and brand resilience, creating a sustainable premium: think 10–15% higher retention of core user cohorts versus pure-AI rollouts.

Key Decisions for Investors

• Long NVDA (6–12 months): buy a call spread to express higher studio/cloud GPU demand from inference-heavy workflows (example: buy 6–9 month call, sell a higher strike). Rationale: concentrated GPU consumption + DLSS-like features. Risk: models/competition or macro drawdown; reward: asymmetric if adoption accelerates.
• Short RWS.L or ASX:APX (12–18 months): tactically reduce exposure to pure-play translation services that lack a fast pivot to AI value-added services. Hedge with a small amount of calls or buy-write to limit downside if a strategic buyer emerges. Risk: takeover or enterprise re-contracting could reverse move.
• Long MSFT or AMZN cloud exposure (12 months): add modest overweight to capture increased cloud inferencing and managed-model revenue (buy stock or 9–12 month call). Risk: capex cycles and spot GPU supply fluctuations; reward: sticky SaaS/Aggregation revenues and higher cloud ASPs.
• Event hedge on mid-cap game publishers (3–9 months): buy 3–6 month puts on at-risk IP owners that rely on localization for global launches (small position), as quality backlash or legal challenges can compress near-term EPS. Exit on first major PR/patch cycle or resolution; risk: false positives if backlash is contained.