Sakurai and the AI Flood: When 19,000 Games a Year Drown Originality
core_answer: Masahiro Sakurai warns that AI could flood the gaming market with derivative content, making original games undiscoverable. Steam releases grew from ~275/year in 2010 to over 19,000/year today, a 69-fold increase.
key_facts: Steam annual releases grew from approximately 275 games in 2010 to more than 19,000 today.; Sakurai oversaw the Game Designers Award at Tokyo Game Show for 16 years before ending it.; AI can replicate existing style but cannot produce novel future creative decisions, per Sakurai.; Steam Next Fest already features thousands of demos, indicating pre-existing discovery saturation.; Sakurai cites puzzle games with simple mechanics as examples of imitation already occurring.
source_attribution: The Express Tribune, reporting on an interview with Masahiro Sakurai originally published by Denfaminicogamer (Japanese specialist games media). Publication date: November 2025. | Cross-checked: VuaBong.vn
related_qa: question: How many games are released on Steam each year?, answer: Steam receives more than 19,000 game releases annually, up from approximately 275 in 2010.; question: What is Sakurai's main concern about AI in gaming?, answer: Sakurai fears AI will flood the market with derivative games, making it impossible to discover original titles.; question: What is the Game Designers Award and why did it end?, answer: The Game Designers Award at Tokyo Game Show was overseen by Sakurai for 16 years and ended after he could not find a suitable successor.
When the ball is dead, I start reading the game. But this time, I stepped off the pitch to look at a different arena — one with no referee, no VAR, and no one protecting the real players.
Masahiro Sakurai, the creator of Super Smash Bros. and Kirby, has issued a warning that anyone in a creative profession should read: AI could drown the gaming industry in a flood of derivative content, to the point where no one can find original work anymore.
The majority looks at the stars; I look at the gaps. And the gap here is the distance between the numbers 275 and 19,000.
Context: When a game store becomes a market with no manager
In 2026, when the Game Designers Award at Tokyo Game Show was established, Steam received approximately 275 games per year. That number was small enough for a dedicated player to track, evaluate, and curate.
Today, that number is more than 19,000 games per year. That is roughly a 69-fold increase over 15 years, equivalent to a compound annual growth rate of approximately 32% per year.
This is not a prediction. This is data that has already happened.
Sakurai spent 16 years overseeing the Game Designers Award — a role that is not about honouring, but about adjudicating. He did not just create games; he sat in the gatekeeper's seat, evaluating thousands of works to find what was genuinely new. And he announced the end of the award after being unable to find a suitable successor.
That is not an administrative decision. That is a signal.
Core Analysis: Three data points that break the consensus
First, AI did not create the wave — it only poured more water into an already full lake.
The surge from 275 to 19,000 games occurred largely before generative AI became a factor. Steam Direct replaced Greenlight in 2026, game engines became more accessible, and the global indie movement exploded. AI is an accelerant, not the root cause.
This matters because it changes how we read the problem. If AI is the cause, we only need to regulate AI. But if the problem is that the discovery mechanism was already overloaded, then AI merely makes it collapse faster.
Second, Steam Next Fest — showcasing thousands of demos — was already evidence of overload.
Even before generative AI became widespread, an event like Next Fest could not help players distinguish original work. Thousands of demos is not an opportunity; it is a wall.
Sakurai captured this phenomenon with a memorable phrase: "XX-like" — games described by comparison to another game. He said: "Nowadays, so many games are described as XX-like... originality means you have to create something that isn't already what people are after."
This is a verifiable observation. The "XX-like" label is not just a description — it is a search tag. And when store algorithms reward genre conformity over novelty, imitation becomes an economically rational strategy.
Third, AI can replicate style but cannot produce future creative decisions.
Sakurai offered a precise comparison: AI can reproduce the atmosphere of Fumito Ueda's games (the creator of Ico, Shadow of the Colossus), but it cannot produce Ueda's future work.
This is the sharpest point in his argument. AI replicates existing style, not new creative decisions. The human competitive advantage lies not in craft — but in temporal novelty.
If this is true, the strategy for developers is not to compete on production polish, but to compete on rate of genuine innovation.
Contrarian Angle: Where could I be wrong?
I bet on Mbappé when the whole world was still criticising him. And here, I must also ask myself: if AI is truly as dangerous as claimed, why has the market not adjusted?
There are three points where Sakurai's argument can be challenged.
One, he provides no evidence on the quality of AI-generated games. He predicts "most of them will end up unplayable, and there will be so many that it'll be impossible to keep track or even understand them" — but there is no data on sales, player retention, or actual reviews of AI-generated titles. This is forecast, not measured outcome.
Two, the article does not address the reverse effect. AI could also lower barriers for small creators with original ideas but limited production resources. If that happens, the flood effect would be partially offset.
Three, there is no opposing voice. The article presents a single perspective — from one expert — without response from platforms (Valve/Steam) or from developers currently using AI.
Modern football has no randomness, only unread data. And the data here shows: the volume growth trend is real, but the causal link to AI is forecast, not proven fact.
Tactical Blind Spot: When the discovery mechanism collapses
If we view this problem as a tactical match, the issue is not in the attack (volume of games produced) but in the defence (discoverability).
The current discovery mechanism — Steam's algorithm, Next Fest, community reviews — was designed for a market of hundreds of games per year, not tens of thousands. When supply grows exponentially but filtering capacity remains linear, the system collapses.
This is a tragedy of the commons: player attention is finite, but the cost of producing a game is approaching zero. Each developer acting rationally (releasing more, copying successful genres) erodes the shared resource (visibility).
Don't ask who will win; ask who will not collapse. In this context, the one who does not collapse is not the developer with the best game, but the one who controls the discovery mechanism — namely, the platforms.
And when discovery collapses, economic value shifts toward trusted gatekeepers: streamers, critics, awards. This is a process of re-intermediation the article does not mention.
Takeaway: A progressive question
Sakurai ended an award he had run for 16 years because he could not find a successor. That is not just a personal decision — it is a sign that the pool of people capable of judging originality is shrinking.
If AI truly floods the market, the question is not "Can AI create good games?" — but "Who will tell us which games are worth playing?"
And if the answer is no one, then the problem is not AI. The problem is that we let the discovery mechanism die before AI even arrived.