Key Takeaways

  • Survival forced customer focus: Higgsfield burned through $10 million of its $16 million seed round without traction before finding product-market fit on its final attempt.
  • Specific user feedback unlocked video generation: Interviews with eight creative directors revealed that AI video tools were unusable for storytelling because they lacked direct camera control.
  • Hyper-growth outpaced prior benchmarks: Higgsfield scaled from $1 million to $1 billion in annualized revenue in 18 months, beating Cursor (24 months) and trailing only OpenAI and Anthropic.
  • Self-serve accounts expanded into massive enterprise deals: One customer started on a $99 monthly subscription and expanded into a contract worth over $6 million within six months.
  • Revenue is tracked with strict monthly proration: The company calculates its run-rate strictly by taking live, prorated monthly revenue and multiplying it by 13.

The Last Attempt: From $10M Burn to Product-Market Fit

Most founders who burn $10 million of a $16 million seed round run out of cash and shut down. Alex Mashrabov faced that exact wall at Higgsfield. The company had spent most of its initial capital chasing general AI video generation without finding user traction.

“We burned more than 10 million out of 16 million raised in seed fundraising,” Mashrabov explained. “So we felt we have just one attempt left.”

Instead of building another generic model update, the team paused development and interviewed eight creative directors. The question was simple: what is stopping you from using AI video in production? The answer was unanimous.

“We spoke to eight creative directors about their experience with AI and what's simply missing,” Mashrabov recalled. “Everyone told us that camera control does not exist in AI and camera control is so important to tell a story.”

By narrowing their product scope from "generate any video" to "give creators exact camera control," Higgsfield turned their remaining runway into an engine that unlocked real creative workflows.

The $99-to-$6M Enterprise Expansion Loop

Fixing camera control triggered a rapid surge in growth. Higgsfield moved from $1 million to $1 billion in annualized run-rate revenue in 18 months. Mashrabov tracks this number conservatively: they count only live, prorated monthly revenue multiplied by 13.

“It took us 18 months from 1 million to 1 billion. For Cursor it took 24 months,” Mashrabov noted. “So we are probably the third after OpenAI and Anthropic.”

That scale requires heavy infrastructure, with internal model compute spending running over $4 million per month. An engineering hub in Kazakhstan builds alongside the core product team to support the workload. But the real driver of the revenue spike came from bottom-up expansion inside large accounts.

“One customer started six months ago spending just subscription $99 a month,” Mashrabov said. “And now we just signed a deal over 6 million.”

Creative teams tested the tool individually on personal credit cards, proved it solved shot-composition bottlenecks, and pushed corporate procurement to replace legacy software budgets with massive annual enterprise licenses.

What to Do With This

If your startup is burning cash without clear usage, stop building broad features. Book calls with eight practitioners in your target role this week and ask one question: "What single missing control makes our product impossible for you to use in production?" Find the specific bottleneck they all mention, strip out every other secondary roadmap item, and direct your remaining runway toward that single control point.