Kingdom Come: Deliverance 2 dev says he was "fired" and replaced with AI: "I feel incredibly betrayed by the management of the company I've come to care about"

Warhorse Studios informed a Czech-to-English translator on March 27, 2026 that his position (nearly four years with the company) will be made obsolete and replaced by AI-driven translations. The decision underscores a push to cut localization costs via AI, has provoked employee and community backlash, and represents modest reputational risk to the studio with limited direct financial or market impact.

Analysis

Replacing in-house translators with AI creates a classic unit-cost-versus-revenue-risk trade: annual headcount savings per language lead are on the order of tens of thousands, while a single high-profile localization failure on a narrative RPG can depress lifetime title revenue by an estimated 5–15% (for a $50–200m title, that’s $2.5m–$30m potentially at risk). Management incentives to cut fixed costs are immediate (quarterly P&L), but the economic friction — lost trust, refunds, review-score degradation — plays out over 3–12 months and is order-of-magnitude larger than the labor savings for most mid/AAA releases.

Second-order winners are the compute and cloud providers supplying the inference stack: sustained, production-level translation pipelines push recurring GPU and cloud spend rather than one-off model costs, supporting NVDA, MSFT Azure, and AMZN AWS revenue lines over the next 6–24 months. Conversely, pure-play localization labor suppliers face disintermediation; however, boutique human-in-the-loop QA vendors can reprice services (2–3x premium) if they position as quality insurance — a likely niche for incumbent translation firms to pivot into.

Regulatory, legal and community backlash constitute the main catalysts that could slow or reverse substitution: unionization drives, class actions for wrongful termination or misleading marketing, and consumer review storms can force rehiring or hybrid workflows within 3–18 months. The key near-term signal to watch is a meaningful decline in review aggregates (Metacritic/Steam rating drops >0.2 points) or a surge in refunds post-launch, which historically trigger rapid remediation spending and re-hires.

Net: short-term cost efficiency favors AI vendors and cloud infra, but medium-term economic value will accrue to hybrid providers and publishers who invest in quality-assurance layers. Monitor incremental cloud / GPU bookings, public translation providers’ margin expansions, and community sentiment metrics to time trades.

Key Decisions for Investors

• Long NVDA (6–12 months): buy NVDA calls or 1–2% portfolio exposure in stock to capture incremental inference GPU demand from productionized translation pipelines. Upside: continued AI deployment; Downside: near-term macro/GPU cycle pullbacks — implied vol on options is a natural hedge.
• Long MSFT or AMZN (6–12 months): add 0.5–1% position in MSFT or AMZN to play recurring cloud spend for model hosting and fine-tuning. Risk/Reward: modest upside if enterprises standardize on Azure/AWS; downside limited by diversified cloud franchises.
• Long RWS.L (3–9 months) or similar public language-service providers: buy equity or calls anticipating a re-rating for human-in-the-loop QA services as publishers pay premiums to avoid quality hits. Target return 30–50% if the market reprices; risk is sustained AI adoption that compresses services pricing.
• Tactical hedge: size a small protective put on narrative-heavy game publishers (select names where exposure is high) ahead of major releases where translations are being moved to AI. If a release triggers a review/refund wave within 30–90 days, these puts should materially appreciate.