When Electronic Arts used generative AI to help them model over 11,000 players likenesses and 150 stadiums for their recently released and acclaimed title, College Football 25, it heralded a new strategic direction for the company: investing heavily into AI.
If EA’s CEO Andew Wilson has his way, AI will become a fundamental part of development at the games studio in the hopes of fueling gangbuster growth and launching the company past its competition.
Artificial intelligence technology has long been a part of video games – characters which can act without input from the player are a perfect example – but generative AI offers the promise of greater efficiency in game development by saving time and cutting costs.
The AI push could allow EA to stand out from competitors in the world of video games, which has struggled to keep the momentum it gained from pandemic lockdowns. It could be particularly helpful for the California-based company, which has been middle-of-the pack in the industry, with a set of franchises that are steady but raise questions about growth.
“Right now we have over 100 active novel AI projects across three strategic categories: efficiency, expansion, and transformation,” said Wilson. “This does not just mean cost savings today. Efficiency is doing what we’re doing today faster, cheaper, and at a higher quality.”
AI is a risky gamble for any company. EA is early to the generative AI investment party among game studios, and the technology is not yet fully developed. Like other sectors of tech, there are concerns that the increases in efficiency brought by AI will not justify the huge investments into it.
EA has already poured billions into AI research and development and promised its effectiveness to investors. AI could impact as much as 60% of development processes at EA, Wilson said at the investor day.
The payoff could take a while.
“We think commercialization of the generative AI opportunity is still a ways away,” said Benjamin Soff, an analyst at Deutsche Bank.
Joost van Dreunen, an adjunct professor of video game business at NYU Stern, said it is “not immediately obvious” there will be an AI transformation in the videogame industry in the short term. But “in the long term, yes.”
It “makes things more efficient,” he said.
EA has strong fundamentals, with high cash flow and no bad debt. Revenue is projected to reach $7.7 billion dollars by the end of fiscal year 2025, which began in April for the company, and its net income is expected to come in at $2.05 billion.
But EA has faced criticism for its repetitive releases of the same franchises each year and reliance on live service revenue.
“EA has been a longtime proponent of annualizing games, especially sport ones,” said Michael Hodel, an analyst at Morningstar. “The annual release schedule induces gamer fatigue and leads to lower sales for franchises.”
EA depends heavily on live service – that sector of their business accounted for nearly three-quarters of total revenue in fiscal year 2024 – and live service revenue can only be maintained by keeping players engaged with titles that regularly offer something new.
When playing EA’s games, that three quarters figure begins to make sense. The company’s sports titles and franchises such as Apex Legends, Battlefield and The Sims all have store pages where players can purchase additional content like cosmetics or playable characters.
This business model is unpopular with many consumers, and a recent survey from the Game Developer Collective showed that around 63% of developers are worried about players losing interest in live-service games, but the strategy is key to EA’s success.
EA shares have risen roughly 20% over the last 12 months, which is comparable performance to competitors like Take Two Interactive and Capcom. But all three companies have underperformed against the S&P 500, which is up more than 30% over that period.
For the company to sustain and accelerate growth, it will need to do more than extracting greater revenue from players through live service. It will need to keep things fresh.
“As the games kept coming out year over year, you could tell that the fundamentals of the games were essentially copy pasted with the exception of a few new features, which to me weren’t really significant,” said Alec Shoup Herrera, a Van Leeuwen employee and American football fan who stopped purchasing and playing Madden games years ago due to their repetitive nature.
AI could allow EA to leverage increases in efficiency to make better games and capture new audiences, the company hopes. If the technology delivers on its promise of saving time and cutting costs, that energy could be reinvested into developing new game mechanics, higher quality graphics and less buggy releases, which could draw more players and strengthen long-term prospects by keeping them interested and spending on full game releases and live service content.
Even one of EA’s most lauded games, Jedi: Survior, could have benefitted from the more efficient development processes promised by AI. The game suffered from bad optimization on PC which made it run poorly on high-end machines and made it practically unplayable on mid to low range set-ups, according to reports from PC Gamer and other game review publications.
The result was a negative review bombardment on PC’s primary digital games distribution platform, Steam. Immediately after release, reviews for the game on the platform were “mostly negative,” in contrast to reviews from IGN, Metacritic and GameSpot which all gave the game a score of at least 8 out of 10. This situation could have been avoided if developers had more time and resources to optimize the game properly in time for its launch on PC.
Today, EA is already using generative AI in their development processes. The company likes to highlight the way it used the technology to help them model far more players and stadiums for EA Sports College Football 25 than they had for any prior Madden game. The franchises are especially comparable because they are both American Football simulation games.
AI helped developers model more than 3 times as many stadiums and nearly 6 times as many player likenesses for College Football 25 than the most recent Madden game. This saved time and labor, but employees still had to manually touch up the AI-created models.
EA has ambitions for further applications of the technology in development and gameplay. For example, they want to create an AI-assisted directory which would help developers sort through the over 100 million assets the company has access to. In regards to gameplay, EA wants to use AI to make the worlds and characters in their games feel more dynamic and immersive. But there is no guarantee that they will be able to accomplish these and other goals for AI’s applications.
AI is a logical bid for EA, as efficiency is the name of the game in the video game industry today, which has struggled because of rapidly rising development costs over the past decade. 2024 already set an industry record for layoffs in one year. As of September, at least 13,000 employees were let go across the industry. In March, EA fired 670 employees, or 5% of their total workforce.
The quest for ever greater efficiency in the games industry has its critics. Marcelo Viana Neto, an assistant professor of game design at CUNY’s Hostos Community College, sees the AI bid as superfluous. “It’s not like cutting costs needs to happen at the software level for game companies,” he said. “They can just fire people, and that has always been the case.”