Midjourney's founder is telling users to "touch grass" while his profitable image generator bleeds relevance to Google's free alternative.

The Summary

  • Midjourney v8 has missed release dates from end of 2025 through March 2026, with founder David Holz publicly deflecting frustrated users on Discord
  • The 37-year-old CEO addresses 19.4 million Discord followers weekly as a blue dragon avatar with camera off, a telling posture for a company under pressure
  • Despite profitability, Midjourney faces an existential question: can a paid image generator survive when Google offers comparable quality for free

The Signal

Midjourney is profitable and stuck. That combination matters because it reveals something about the agent economy's early winners. The company charged $10-$60/month when image generation was hard and alternatives were bad. They built a moat around quality and a community around Discord. Now the moat is evaporating.

Google's image models, bundled into search and free to use, are closing the quality gap fast. Midjourney's response has been to chase hardware dreams (the piece's title telegraphs this pivot) while delaying the software update users actually want. Version 8 was supposed to ship in late 2025. It's now late March 2026. Holz appears weekly on Discord as an avatar, camera off, managing expectations downward.

The dragon avatar detail isn't just color. It's a tell. When you're winning, you show your face. When you're navigating a strategic pivot away from your core product while your users get impatient, you log in as a cartoon and tell people to touch grass. That's the posture of someone buying time.

Profitability gives Midjourney runway, but it also creates inertia. They're making money on the old model, which makes it harder to cannibalize themselves for whatever comes next. Meanwhile, Google doesn't need to make money on image generation. They need to keep you inside their ecosystem. Different incentives, different timelines, different endgames.

The Implication

Watch what Midjourney ships next. If it's v8 with incremental improvements, they're playing defense. If it's hardware or a pivot toward something agents can use programmatically (automated image workflows, API-first tools), they're trying to escape the consumer trap before Google fully commoditizes it. The companies that survive Web4 won't be the ones with the best models today. They'll be the ones who figure out what agents need to build tomorrow, then get there first.

Source: The Information