A stock photography platform's pivot to AI training data supply has produced one of the sector's more striking business transformations — $40 million in annualized revenue, a 20x jump in creator payouts, and now a $23 million round to push further.

---

When Wirestock launched in 2018, its pitch was straightforward: help photographers distribute their work across multiple stock image platforms with a single upload. By 2022, the platform had grown to more than 100,000 photographers and landed a partnership with Getty Images. Then the generative AI wave hit, and the company made a calculated bet that the real money was not in stock licensing — it was in becoming the connective tissue between the world's creative community and the frontier AI labs hungry for high-quality training data.

That bet is now paying off in earnest. On May 15, Wirestock announced the close of a $23 million Series A funding round led by Nava Ventures, with participation from SBVP — the fund co-founded by former Meta COO Sheryl Sandberg — alongside Formula VC and I2BF Global Ventures. The round brings the company's total capital raised to approximately $26 million and arrives as Wirestock reports an annual revenue run rate of $40 million, almost entirely driven by AI data licensing.

---

From Stock Library to Data Factory

The pivot, which began in earnest in 2023, required Wirestock to fundamentally reimagine what its platform was for. Rather than helping creators sell images to editorial clients, the company began sourcing custom datasets — images, videos, design assets, gaming content, and 3D models — for the training pipelines of the world's largest foundation model developers. Wirestock's CEO and co-founder, Mikayel Khachatryan, says the company now supplies six of the biggest foundation model makers, though he has declined to name them publicly.

The platform now functions more like a specialized freelance marketplace than a stock agency. Over 700,000 artists, photographers, videographers, graphic designers, 3D artists, filmmakers, and musicians have signed up as contributors. They complete data collection tasks — capturing specific footage, generating assets to detailed creative briefs, or annotating existing content — and submit work through a pipeline that uses a combination of AI and human review for quality control. Applicants must complete an unpaid trial task before being accepted, a filter designed to maintain the data quality that AI labs actually need.

"Initially, a lot of our deals were just selling what we had off the shelf, like our existing library," Khachatryan told TechCrunch. "But then it turned into a lot of custom requests for content and data, and that created new opportunities for creators, and the platform just took off."

The financial results back up the narrative. Wirestock has paid out $15 million to contributors since the pivot, and the move to an AI licensing model has driven a 20x year-over-year increase in creator payouts — a figure the company is leaning on heavily as it competes for a larger share of the creator economy against traditional stock platforms.

---

Why Investors Are Paying Attention

Freddie Martignetti, founder of Nava Ventures and the round's lead investor, said his fund was actively searching for a company innovating on data procurement and refinement before it encountered Wirestock.

"I think Wirestock has a deep understanding of what foundational models and hyperscalers need in terms of multimodal data to start creating more human-like systems," Martignetti told TechCrunch. "The cornerstone of our thesis was that multimodal data will be increasingly important, not just to create images or videos, but for models to complete real-world tasks."

That thesis is well-timed. The AI data supply sector has become one of the most contested segments of the broader AI infrastructure build-out. Scale AI has been valued at tens of billions of dollars. World Labs — Fei-Fei Li's visual world-model startup — closed a $1 billion round in February 2026. Rival AMI Labs raised $1.03 billion the following month. Meanwhile, a wave of newer entrants including Micro1, Human Archive, and Human Native AI are all competing for licensing and annotation contracts with top model developers. Wirestock's differentiation is its focus on creative and generative modalities — the kinds of rich, human-authored visual and audio assets that matter most for image generation, video synthesis, and computer use models — rather than general-purpose annotation work.

SiliconANGLE noted that Wirestock's catalog of over 50 million images and videos positions it specifically to serve the surging demand from world model developers, who require dense visual data for training systems designed to complete real-world tasks.

---

What the Capital Will Fund

Wirestock currently employs 60 people. The $23 million will be directed toward hiring in research, engineering, and product — with a particular focus on building enterprise collaboration software that allows AI labs to work more directly with Wirestock on dataset design and iteration. The company is also exploring expansion into audio and music as additional data modalities, a move that would extend its reach into one of the more contested new frontiers of multimodal AI training.

The broader competitive picture is one of rapidly accelerating demand. As foundation models push toward genuine multimodal capability — understanding and generating images, video, audio, 3D environments, and text as a unified system — the labs building those models increasingly need training data that reflects the full texture of human creative output. Wirestock's argument is that its community of 700,000 working creators, organized through a structured data supply platform, is uniquely positioned to meet that demand at scale and with the quality controls AI labs require.

Whether a 60-person company can hold its position as the largest players in the data supply chain continue to scale is the central question facing Wirestock's next chapter. But the numbers — $40 million in run-rate revenue, 20x payout growth, a roster of six frontier model clients — suggest the pivot has already succeeded. The Series A is a bet that the hard part is only beginning.

---