Starform keeps Metalstorm alive with 26 people and AI-native workflows

Starform launched Metalstorm in November 2023 as a 5v5 air-combat game available on mobile and PC. The game reached three million players by late 2024 when the studio raised $6 million from Bitkraft Ventures. CEO and co-founder Lou Fasulo now reports the player count at 15 million across the App Store, Google Play, and Epic Games Store.

At SLICE 2026 in Seattle, Fasulo discussed how a 26-person team maintains both a live game and early development on a new project. The key, he explained, lies in understanding what AI actually changes for game development.

"AI-native is more of a journey than it is a state," Fasulo said. "I wouldn't say that AI enables us to do something we couldn't already do." What shifted was throughput. "We've shipped way more features. We've addressed so much more feedback from players and ultimately we have a better player experience."

The studio uses AI to handle routine work, freeing the team for higher-order decisions. But Fasulo is explicit about its limits. Metalstorm's aircraft alone involve tens of thousands of design decisions: acceleration curves, energy bleed in hard turns, silhouette readability at distance. "When you don't specify one, the model just makes something up. That's fine for a prototype," he said, but shipping requires human specification across a full game's scope.

Fasulo pushes back on how the team-size question is typically framed. "The right size for a game is probably different from the right size for a company. What you're really drilling into is what's the team size that optimises for culture, shared vision and passion for that vision." He argues that larger teams struggle to process all feedback. "When the team is so large that you cannot process all the feedback from the 2,100 people that worked on the last Halo update, their voice and their contribution diminishes so much. And that's the cultural miss."

Metalstorm runs on both mobile and PC, but Fasulo's view is that those audiences don't want the same game. During early prototyping, the team tested five to seven-minute matches and noticed players wouldn't stop between rounds. "They just kept playing. We see them play multiple matches almost like eating chips out of a bag." That player behaviour, not platform strategy, proved the format could work across devices.

Making hard calls early defines Starform's approach. Metalstorm's initial build was simulation-heavy and appealed to early players. When thousands more arrived, the reception was different. "The easy thing would've been, oh, we'll make it 10% better. Instead we did a hard pivot on all of the mechanics and said we're nowhere close to success," Fasulo said. That decision required significant work, but he credits developing "the muscle memory to make the smaller calls so that the bigger ones are just more natural."

Starform hasn't raised funding since that $6 million round two years ago. Fasulo reframes how live games sustain themselves. "We're not the reason the game is still alive. It's the players that love it. Studios that don't have passionate players that are willing to invest their time and money become irrelevant." He argues the industry misdiagnoses why new games fail. "The new games aren't winning players from their hobby." Most venture funding for new studios never produces a game that finds an audience, he suggests.

Managing a live game and early development with 26 people is feasible, Fasulo explained, because the expensive phase has shifted. "That discovery and the R&D that you think of as innovation in gameplay, that is not expensive. What becomes expensive is taking it all the way into and scaling production to deliver it to players."