The games industry has shed thousands of positions since January 2025, with major publishers and independent studios alike reducing headcount. Microsoft Gaming, Sony, Twitch, and Riot Games have all announced significant cuts. The exact numbers shift weekly as more announcements arrive, but the trend is consistent: the post-pandemic expansion has reversed, and teams are shrinking.
Wildlight, developer of Highguard, laid off most of their staff in February after stating publicly that the game did not need a massive player base to remain sustainable. The reversal came within weeks. Other studios have faced similar sudden closures: projects cancelled, teams dismissed, sometimes with the announcement of layoffs preceding the actual communication to affected employees.
The sales data tells a different story than the employment figures. US video game sales climbed 3% year-over-year in January 2025, according to market tracking data. Players are buying games. The revenue is there. The layoffs are not a direct result of market contraction—they are a result of business models that no longer pencil out, of acquisitions that loaded companies with debt, and of a correction after years of over-hiring during the pandemic boom.
The AI Question
Artificial intelligence remains the variable that no one has solved for. Tooling has improved: procedural generation for environments, AI-assisted animation, voice synthesis for placeholder dialogue, and generative systems for concept art. These are real capabilities that studios use daily now. What remains unclear is the structural impact. Will AI reduce team sizes by automating asset production? Will it shift the economics toward smaller teams with bigger ambitions? Or will it simply become another tool that every artist and designer is expected to learn, adding to the skill requirements without changing the fundamental cost structure?
The investment money that flooded into AI gaming startups in 2023 and 2024 has not yet produced clear success stories. Most AI-native games remain experimental—technically interesting, commercially marginal. The established publishers are integrating AI quietly, behind the scenes, for production efficiency rather than player-facing features. No one has made the killer AI game yet, and no one has figured out how to make AI reduce costs without reducing quality.
Technical Infrastructure: The Quiet Progress
While the business side sorts itself out, technical standards continue to evolve. Digital Audio Denmark has released firmware updates that bring SMPTE ST 2110-30 compatibility to their Thunder | Core, Penta, AX32 units, and Dante expansion cards. The updates require DADman version 5.8.2.2 and corresponding firmware installations.
SMPTE 2110-30 is the standard for professional audio over IP, defining how uncompressed PCM audio travels across networks with the timing precision required for broadcast and post-production workflows. Compatibility with this standard means DAD hardware can now slot into broadcast infrastructures that have adopted IP-based signal routing rather than traditional SDI or MADI.
This matters for game audio because the same IP audio infrastructure is increasingly used for both game development and broadcast post-production. A facility that mixes game trailers can now use the same Dante-to-2110 bridge for broadcast deliverables. The distinction between game audio and linear audio workflows continues to blur at the infrastructure level, even as the business models diverge.
What This Means for Audio Professionals
The immediate impact of layoffs on audio departments varies by studio. Some have cut audio headcount proportionally. Others have retained audio staff while reducing other disciplines, treating sound as a fixed cost that cannot be reduced without damaging the product. The pattern is not consistent enough to draw industry-wide conclusions.
For freelance audio contractors, the market is more competitive. The pool of available talent has grown as full-time positions disappear. Rates have not collapsed, but negotiating leverage has shifted toward the employers. The studios that are hiring—there are some, particularly in mobile and live-service games—can afford to be selective.
The AI tools most relevant to game audio right now are not the generative systems that generate sounds from text prompts. They are the established tools: AI-assisted noise reduction for field recordings, dialogue cleanup algorithms that can rescue location audio, and procedural systems for creating variations of existing sounds. These save time. They do not replace sound designers.
Looking Forward
2025 is likely to be a year of consolidation. The studios that survive this period will emerge leaner, with lower overhead and more conservative projections. The AI question will not be resolved this year, but the experiments will continue. Someone will ship a commercially successful game that uses AI in a player-facing way, and the industry will study it closely. Until then, the tools will keep improving in the background, and the audio professionals who understand both the creative and technical sides will remain employable regardless of the broader employment trends.
The work itself has not changed. Good audio still requires good ears, good taste, and the patience to iterate. The context around the work has become more difficult. That is the current state of the industry.