Funding Brief | Higgsfield Series B $400M

🌏 中文版

Funding Details

What the Company Does

Higgsfield builds AI video and image generation tools that let brands, agencies, and studios produce commercial-grade marketing videos from text and reference assets, without needing a production studio or post-production team.

The core product doesn't train a single proprietary model. Instead, it routes multiple external generation models (Sora, Kling, Veo, etc.) through a unified workflow, enabling users to iteratively edit and re-render a video rather than treating generation as a one-shot process. This was a deliberate differentiation choice by founder Alex Mashrabov (formerly head of generative AI at Snap). The Supercomputer feature launched in May automates complex multi-scene, multi-shot video production. The agentic product line driving this capability grew usage 42x in three months, producing over 20 million pieces of content per month.

The platform now has over 30 million users across 238 countries and territories, serving 390 Fortune 500 companies across advertising, media & entertainment, fashion retail, financial services, and pharmaceuticals. Enterprise customers have flipped from under a quarter of revenue in January to a majority today, signaling a shift from consumer creator tool to enterprise marketing infrastructure.

Signals from This Round

What It Means for the Agent Ecosystem

The three stated uses of funds -- enterprise product, security (a prerequisite for enterprise adoption), and compute -- all point in the same direction: Higgsfield is transitioning from "fun AI video generator" to infrastructure that enterprise marketing departments depend on. The 42x growth in the agentic product line is the key signal: customers don't want a single image or clip -- they want automated pipelines that break scripts into multi-scene, multi-shot batch production. This is what "agent" concretely looks like in content production. Compute is explicitly called out as the biggest bottleneck, since video generation costs far exceed text-based applications, burning capital several times faster than text-focused agents.

What Investors Are Betting On

DST Global founder Yuri Milner's thesis is straightforward: betting that the Higgsfield team can keep building the next generation of visual creation AI tools. More telling is Tribe Capital partner Boris Revsin's framing -- "we look for true outlier metrics" -- pointing to Higgsfield's revenue growth rate and customer retention both being exceptional, not just user count vanity metrics. The investor mix is also revealing: Goldman Sachs Growth Equity, Intel Capital, and Liberty Global entering simultaneously map to compute, connectivity, and distribution respectively, suggesting that large infrastructure and media companies now view "AI video generation platform" as a strategic supply-chain position to lock in, not just a financial bet.

Numbers Worth Watching

• Annualized revenue grew from $20M a year ago to $700M now -- 35x growth -- at a valuation multiple of roughly 7.7x ARR, comparable to competitor Runway ($315M Series B, $5.3B valuation), though Higgsfield's revenue is growing significantly faster;
• Valuation jumped from $1.3B to $5.4B in 8 months (over 4x), while the most direct Western competitor Runway has pivoted toward world models for robotics and healthcare, effectively ceding the marketing video battlefield;
• Chinese competitor Kuaishou's Kling raised nearly $3B in the same period at an $18B valuation -- 3.3x Higgsfield's -- showing that AI video generation capital intensity is a global race, not a US-only market.

Watchlist Status

Higgsfield is not yet on the watchlist. Recommend adding to section C6 (multimodal AI generation tools), tracked alongside Runway, Pika, Kling, Sora, and Veo. Watch for: agentic multi-scene video production pipeline + enterprise marketing customer conversion, $400M Series B, $5.4B valuation.

Takeaway

I had assumed the AI video generation space was largely captured by model makers' own products like Sora and Veo, but Higgsfield's approach proves the opposite: once base models become commoditized resources anyone can plug into, the real value sits in the workflow layer -- the ability to chain multiple models into a repeatable, editable, batch-capable production pipeline rather than training yet another video model. That's why the agentic product line (automated multi-scene production) grew 42x, rather than raw generation volume alone.

References

• Higgsfield Raises $400 Million Series B Financing at $5.4 Billion Valuation with Annualized Revenue Reaching $700 Million — PR Newswire
• Higgsfield raises $400M from Goldman Sachs, DST Global at $5.4B valuation — Tech Funding News
• AI video startup Higgsfield raised $400 million, quadrupling its valuation in about eight months — Quartz