Arc Raiders boss promoted to Nexon chief backs AI push for "redesigning game development" after building a hit "at a fraction of the cost you'd expect for a AAA game"

A newly released capital markets briefing from publisher Nexon features extensive comments from Patrick Söderlund, CEO of Arc Raiders developer Embark Studios and newly appointed Nexon executive chairman. Söderlund discusses a renewed focus on efficiency, applying Embark's production lessons to the Nexon collective where possible, and, in his own words, "a topic you've heard about from every company this year: AI."

Beneath a slide titled, "Redesigning Game Development," remarks from Söderlund insist that "every company has a plan" for AI but "most will get it wrong."

"They're committing big investments in tools – but tools won't help because they've misread the challenge," he says. "AI may be a race, but the winners won't be the first movers – the winners will be the ones who understood the challenge."

What is that challenge? Söderlund suggests you "think of game development as auto mechanics," in that "the tools are available to everyone, but not everyone has the knowledge and experience to use them."

Nexon, he says, is in a unique position to get AI right thanks to 30 years of work that covers "billions of player sessions across some of the world's longest-running games," with MapleStory being a cornerstone. This experience provides the "context" necessary to steer AI pragmatically, Söderlund suggests, and to make AI "usable at speed, and at scale."

This builds on Embark's experience with Arc Raiders and The Finals, and "yes, some of that involves AI," Söderlund says, though he argues it's "really about encouraging people to use smarter processes, better tools, and to let go of habits that no longer serve them."

Meanwhile, Arc Raiders has notably been replacing its divisive AI voices, which were couched as a production save for a company now flush with cash, with actual human actors, earning praise from some AI critics like Baldur's Gate 3 Astarion actor Neil Newbon.

"At Embark, we started with a blank slate, questioning everything from: How do you get from an idea to a green light?" Söderlund recalls. "To what needs to be done by hand versus what a machine can do more efficiently?"

Embark's success with The Finals and then Arc Raiders "wasn't an accident," he stresses, "it was deliberate. And now we're bringing that thinking to the rest of Nexon."

"Two games, built with significantly fewer people, at a fraction of the cost you’d expect for a AAA game," he adds.

This really says aloud what Nexon leaders presumably had in mind when promoting Söderlund: hey, Arc Raiders really popped off, do you think you could make that happen with our other games? To that end, Söderlund outlines a plan to tackle rising development costs, tighten a "too wide" portfolio of products, and "figure out why" margins are shrinking and the "pop" from releasing new titles "doesn't stick." AI is, of course, a part of this plan.

Nexon's vision for AI crystallizes with "Mono Lake," an initiative described by president and CEO Junghun Lee as an "end-to-end step change in how we create and support our games."

"Nexon has used AI tools for some time and we're quickly moving past the tool level – to applying context to everything we do," Lee says.

Likewise highlighting how Nexon can succeed where others failed, Lee argues that, "without context, AI is a race to the arithmetic middle where everyone’s games look the same." But with Nexon's huge "base of information" at the helm, Mono Lake planning can avoid "a generic outcome."

He also heads off enduring concerns that AI will replace people or cannibalize creativity. Lee outright says, "Our methodology doesn't replace creative people, it frees them to create, with context." This does, of course, assume a readiness to offload work to AI, an idea which many game developers have rejected – as it happens, in response to older comments from Lee.

Context for AI is thrown around quite a lot in the briefing, but let's dig into a crucial bit of more traditional context: "How do the lessons of Arc Raiders factor into this?" as Lee asks.

Well, evidently, "beyond the breakaway commercial success, Arc Raiders is a Trojan Horse – a gift that contains a shift in the mindset about how technology frees developers and live service teams to spend more time thinking and less time typing. More time innovating; less time writing code."

I don't know that "Trojan Horse" was Embark's blue-sky vision for Arc Raiders, but Lee seems quite enthusiastic about it. (Amusingly, Arc Raiders is also described as "the first real proof that Nexon can build something that lands with a global audience" outside its Asian base.) Again bigging up Mono Lake as an extension of this approach, Lee says, "It changes how people work. The tools they use, how fast they can move, what they can accomplish."

As Nexon executives repeatedly insist that the company's approach to AI will not replace people or compromise their creative output, it can be tough to square that with the constant refrain of cutting costs, controlling margins, and improving efficiency. The bleedingly optimistic outlook would be AI-improved efficiency manifesting as faster work with less wasted labor, as game development costs are largely a function of people (and wages) over time. But, unignorably, we're already seeing companies accused of slashing people for AI, too.

"Man up": Former Microsoft and Blizzard exec calls out Crimson Desert devs for apologizing over AI art, because it'll "be in every single video game" someday.