Higgsfield hits a wall: Viral AI video unicorn faces fury over racist ads and shady billing

When an AI video startup turns into a $1 billion unicorn, clears $200 million in annual revenue, and amasses 15 million users in a flash, Silicon Valley usually celebrates another triumph.

With Higgsfield AI, however, the rapid ascent has quickly descended into a full-blown PR nightmare.

Forbes and international outlets recently reported that behind the platform's meteoric rise sits a dark trail of predatory billing tactics, locked accounts, and promotional materials loaded with racism, sexism, and unauthorized deepfakes.

Unlimited credits that suddenly dry up

Higgsfield was founded with headquarters in San Francisco and deep engineering roots in Kazakhstan, serving as a flagship member of the Astana Hub. By late 2025, it attained unicorn status, driven largely by performance marketers who make up 85% of its user base.

Yet trouble brewing beneath the surface erupted over subscription practices.

During recent Black Friday campaigns, Higgsfield slashed prices by up to 65%, selling monthly plans for $25 with promises of unlimited generations. Paying customers quickly discovered that rendering speeds plummeted to a crawl, rendering the tool practically useless unless they shelled out extra cash for credits.

Cancellations were nearly impossible to execute, leading to accusations of classic dark patterns.

Tensions flared in late December 2024 when scores of paying accounts were abruptly banned without warning, locking users out of both the service and their previously rendered files. Higgsfield later claimed the purge was aimed at blocking tens of thousands of automated bot accounts, adding that it eventually refunded around $1.35 million to legitimate customers.

Rage-baiting with unapproved deepfakes

The company’s marketing playbook drew even fiercer backlash.

Reports revealed that Higgsfield's promotional team directly supplied creators with ready-made video packages containing racially charged slurs and sexualized dialogue dubbed over well-known cartoon characters.

The materials also leaned heavily on unauthorized celebrity deepfakes, featuring sexualized likenesses of actors Sydney Sweeney and Zendaya, as well as caricatures of Donald Trump and Elon Musk. Critics called it textbook rage marketing—provoking disgust purely to capture attention.

Co-founder Alexander Mashrabov took to social media to apologize, conceding that the startup's rapid scaling caused operational breakdowns and failed to match company values.

Why it matters

We think Higgsfield's sudden crisis highlights the toxic shortcuts some generative AI startups take when prioritizing hypergrowth over basic product integrity.

For creators and marketers, this tool is hard to recommend right now. A platform that lures you in with unlimited tiers only to choke bandwidth, hides cancellation buttons, and locks down your generated assets is a major operational liability. Using assets derived from controversial and unauthorized deepfakes also exposes brands to serious legal blowback.

For the broader industry, this episode offers ammunition to regulators pushing for strict controls on deepfake generation and digital subscription billing. Growth hacking via rage bait may generate flashy top-line revenue, but it destroys enterprise credibility in seconds.