Maimai founder and CEO Lin Fan stated that from January to July 2026, the number of newly posted positions in embodied intelligence surged by 6 to 7 times year-on-year, yet the average salary for these roles is only two-thirds that of large model positions. Lin Fan disclosed this data during Maimai's "AI Roundtable Pi" program.
According to Maimai recruitment data, from January to July 2026, the number of domestic embodied intelligence positions grew by approximately 6 to 7 times, but the talent supply increased by only 3 to 4 times, making it difficult to meet corporate demand. Currently, talent in embodied intelligence primarily comes from five directions: native research talent in embodied intelligence, traditional internet talent, large model talent, autonomous driving and automotive talent, and smart hardware talent.
Lin Fan believes that the current state of embodied intelligence closely resembles autonomous driving a decade ago. The technical approaches have not yet converged, requiring simultaneous coordination among models, data, hardware, and the industry chain. The business model is still in the exploratory stage, leaving researchers and entrepreneurs greater room for direction judgment and technological innovation. This is a key reason why top-tier AI talent is beginning to shift toward the embodied intelligence field.
Lin Fan further noted that the success of model companies does not hinge on simply accumulating headcount, but on whether they possess core talent capable of determining strategic direction, and whether leadership can understand the technology, unify objectives, and reduce internal friction. "A truly successful core team for building models may only need one to two hundred people," Lin Fan remarked.
The most elite AI talent typically shares two common traits, according to Lin Fan: first, a solid mathematical foundation to grasp the fundamental principles behind models; and second, an interdisciplinary background that allows them to look beyond isolated technical components and evaluate technology roadmaps from a holistic perspective.
Lin Fan also revealed that Maimai has launched an AI talent tagging system to help companies more quickly identify "who is AI talent and which category of AI talent they belong to." Corporate HR personnel or headhunters can clearly see three types of AI talent tags—AI technical infrastructure talent, AI application development talent, and AI business efficiency talent—in the search bar, talent lists, and talent profiles within the "Talent Bank" backend.
Discussing the impact of AI on work, Lin Fan argued that CodingAgent's replacement of programmers' tasks does not necessarily mean large-scale unemployment for programmers, but rather triggers a reflow of talent. He proposed that while the past decade focused on moving servers and machines to the cloud, the next ten years may witness "human resources moving to the cloud." Companies may no longer rely solely on hiring permanent employees to acquire capabilities, but instead purchase specific outcomes from individuals or Agents in the cloud, fundamentally reshaping how organizations collaborate with talent.
"AI Roundtable Pi" is a high-end dialogue program launched by the Maimai community in November 2025, focusing on AI and workplace transformation. It invites founders, technology experts, scholars, and investors from the industry frontline to engage in in-depth discussions on topics such as technological evolution, industrial implementation, and talent mobility.