Handshake pivots to $1.1B labeling marketplace

This is a supply side pivot disguised as a product pivot. Handshake already owned a hard to replicate pool of students, graduates, postdocs, and PhDs, then redirected that network from helping employers hire full time talent to helping AI labs buy hourly expert judgment. The result is that a once high margin campus software business is now economically driven by marketplace volume, contractor payouts, and frontier lab demand for credentialed human feedback.1 2 3 4 5

• The old business sold software to career centers and recruiting access to employers. The new business sells completed annotation and evaluation work to AI labs. That is why gross revenue can surge much faster than net revenue. In April 2026, Handshake was at $1.1B gross annualized revenue, but about $450M net after paying contractors.1 2
• The inversion happened because the recruiting engine stalled while the expert labor engine exploded. Handshake data showed postings down 15% year over year and applications per posting up 30%, a sign of a weaker entry level hiring market. At the same time, Handshake AI opened trainer roles beyond students and tapped a much broader expert base.1 4 5
• Compared with Mercor and Invisible, Handshake won by starting with proprietary academic supply instead of building supply from scratch. Mercor was at a $500M annualized run rate in August 2025 and Invisible did $134M in 2024, while interviews across the category point to a shift away from generic crowdwork toward verified experts in fields like math, law, and science.3 6 7

The next phase is Handshake becoming less like a campus recruiting vendor and more like core infrastructure for frontier model evaluation. As labs demand more auditable, domain specific human input, the company can keep moving up from sourcing workers to owning workflows, quality control, and research tooling around expert data.2 3 7