Higgsfield (Platform)
AI video and image generation platform operated by Higgsfield, focused on camera-motion video generation and editing tools.
Higgsfield develops AI tools for video creation, including products for marketing and filmmaking. Founded by Alex Mashrabov and colleagues, the company focuses on cinematic media generation and workflows that help users create and produce video content.

Founded 2024 · San Francisco, California, US
Official Website$80M · $1.3B valuation
The official release names participating investors, while media accounts differ over the lead role. Participation does not by itself establish lead status.
Source · reuters.com ↗$50M
The valuation comes from reporting rather than the company announcement.
Led by GFT Ventures. Participation included Menlo Ventures.
Source · prnewswire.com ↗$8M
Announced alongside the initial rollout of the Diffuse app.
Led by Menlo Ventures.
Source · markets.financialcontent.com ↗Alex Mashrabov (former Snap AI lead)
Alex Mashrabov co-founded AI Factory — the personalized-video startup Snap acquired for US$166M in 2020 — then ran Snap's generative-AI work as Head of Generative AI, shipping the MyAI chatbot (150M+ users) and viral GenAI AR effects. In September 2023 he left to found Higgsfield with Kazakh AI researcher Yerzat Dulat (CTO) and veteran operator Mahi de Silva, spending about a year in stealth before launching a browser-based 'click-to-video' platform in March 2025. The thesis: build for the attention economy, where social-media marketers and brands need cinematic, character-consistent video at scale, not just prompt-based clips.
AI video and image generation platform operated by Higgsfield, focused on camera-motion video generation and editing tools.
Higgsfield's trajectory compressed several years of growth into months. The platform went from US$0 to US$10M ARR within weeks of its March 2025 launch, to ~US$100M within five months, to US$200M annual run rate by January 2026 (doubling in about two months), and to US$700M annualized by the August 2026 Series B — a pace the company benchmarked against Lovable, Cursor, OpenAI, Slack and Zoom. Users grew from 11M (September 2025) to 15M+ (January 2026) to 30M+ across 200 countries, with professional social-media marketers driving roughly 85% of activity.
The Series B's US$5.4B price more than quadrupled the US$1.3B set in January 2026, itself up from a ~US$1B unicorn round five months earlier. The round's shape is the subplot: DST Global led, while a strategic tier of corporate and sovereign investors — Intel Capital, Goldman Sachs Alternatives' Growth Equity, Liberty Global Tech Ventures, Mirae Asset Capital and NTT DOCOMO Ventures — signalled that compute, connectivity and distribution giants see Higgsfield as infrastructure for the next era of visual media. CEO Mashrabov framed compute as the moat: one minute of video is like processing 60,000 words, making reserved capacity a competitive necessity against Synthesia and Runway.
Higgsfield was widely reported as Kazakhstan's first unicorn at the September 2025 Series A — a point of national and diaspora significance given Mashrabov and Dulat's origins. The Series B doubled down on the APAC path: Korea's Mirae Asset Capital and Japan's NTT DOCOMO Ventures joined as strategic investors, with proceeds earmarked for a dedicated Asia-Pacific go-to-market team, alongside the Higgsfield Academy and the Higgsfield For Good program (partnering with NGO YGA to put AI visual tools in the hands of 70,000 students and 13,000 educators).
Higgsfield's early notoriety came from celebrity viral videos — Madonna, Will Smith and Snoop Dogg were named users — but the company has worked to reposition as enterprise infrastructure, with Marketing Studio for advertisers and Cinema Studio for filmmakers, and agentic products (after the May 2026 Supercomputer rollout) that automate multi-scene production and grew 42x in three months. The durable question is the same one facing every consumer-born AI tool: whether it can consolidate into a platform with staying power before the next hot model arrives.
Because DST's entries in this dataset cluster at proven consumer-scale revenue: it joined Lovable only from the Series B onward and bought into Clay via a US$5B tender before arriving at Higgsfield's US$400M round with US$700M annualized — up from US$200M in January. The price it set, US$5.4B, more than quadrupled January's US$1.3B in eight months.
It is San Francisco-based with Kazakh roots: CEO Alex Mashrabov and CTO Yerzat Dulat are Kazakh, and media widely applied the unicorn label at the September 2025 Series A — a ~US$1B valuation reported rather than officially disclosed. The Series B then institutionalized the regional angle with Korea's Mirae Asset Capital and Japan's NTT DOCOMO Ventures, plus an APAC go-to-market team.
Because compute is the claimed moat: Mashrabov's ratio — one minute of video equals processing about 60,000 words — argues reserved capacity, not model quality alone, separates Higgsfield from Synthesia and Runway. The Series B tier — Intel Capital, Liberty Global, NTT DOCOMO, Mirae — is a compute-and-distribution coalition, echoing World Labs' NVIDIA-and-AMD double-down.
The company name-checks Lovable, Cursor, OpenAI, Slack and Zoom: US$0 to US$10M ARR within weeks of the March 2025 launch, ~US$100M by month five, US$200M by January 2026, US$700M annualized by August. Lovable needed eight months for its first US$100M — Higgsfield five — and Higgsfield's US$700M by month seventeen exceeds Lovable's last disclosed run-rate (~US$500M, June 2026).
Most of the thesis: Snap bought Mashrabov's prior startup AI Factory for US$166M in 2020, installed him as Head of Generative AI, and his MyAI chatbot reached 150M+ users before he left in September 2023. Converted to capital, that pedigree priced at a US$8M seed — AMI Labs, founded months later on LeCun's research fame, cleared US$1.03B: executive versus scientific credibility, 100x apart.
No. A creative-AI revenue or run-rate figure does not disclose inference costs, third-party model fees, refunds or gross margin. In Higgsfield's profile, product growth and fundraising are context; neither supplies a unit-economics breakdown for individual video generations.