Embodied intelligence has been highlighted once again in the 2026 Government Work Report, which calls for cultivating future industries such as next-generation energy, quantum technology, embodied intelligence, brain-computer interfaces, and 6G. This marks the second consecutive year the concept has been included, following its initial appearance in the 2025 report.
The field of embodied intelligence experienced significant activity in February, with companies making appearances in high-profile events and securing funding, indicating intensifying competition.
"Even the best wine needs a sign," remarked Gu Shitao, co-founder of Magic Atom, explaining the company's decision to participate in a major televised event. In a recent interview, Gu revealed that Magic Atom will officially announce a new round of financing next week, with additional funds amounting to hundreds of millions of yuan queued for completion.
She described the televised event as a "catalyst" that complemented the company's ongoing efforts, offering insights into Magic Atom's strategic thinking at this juncture: why the appearance served as a technical validation, and how the company views the sudden influx of capital into the industry.
When asked about the tangible outcomes from the event, Gu noted significant benefits. "In an industry with many embodied intelligence and humanoid robot companies, demonstrating strong stage performance helped validate our technical capabilities," she said. The exposure also attracted interest from top talent at other firms, strengthening Magic Atom's position in the competition for skilled professionals and resources.
Regarding recent funding disclosures by several embodied intelligence companies, Gu characterized the event as synergistic rather than solely responsible. Magic Atom had been preparing its Series A financing beforehand and has now secured at least 500 million yuan, with additional term sheets still growing.
Commenting on the accelerated investment pace since late last year, Gu pointed to broader trends. "Many are calling this the second wave of AI, the first year of AI Agents or commercialization," she observed. While last year focused on basic assistant applications, current developments involve multiple agents tackling complex problems, with potential to transcend time and space constraints and boost productivity across sectors.
"The revolutionary impact of AI on productivity is already clear," Gu stated. She explained that unlike a decade ago when investors prioritized certainty through financial metrics, today's AI investments target sector-wide beta opportunities despite individual company uncertainties. "Everyone knows an inflection point is coming, similar to the next industrial revolution. While specific competitive landscapes and business models remain fluid, the sector's overall direction carries high certainty."
Gu emphasized that China's increasingly market-oriented investors are now betting on确定性 sectors while backing different company types and approaches, waiting to see which emerge successful.
Addressing why the industry requires substantial funding, Gu highlighted the distinction between embodied intelligence and conventional robotics: AI-driven systems require physical embodiment, necessitating simultaneous solutions to both software and hardware challenges. Significant upfront investment in software, computing power, data acquisition, and talent is essential before scaling becomes possible. "Only with ample resources can technology achieve exponential leaps and 'ChatGPT Moments' occur," she added.
Gu noted a visible increase in investor patience domestically. The ChatGPT breakthrough in the U.S. benefited from greater tolerance for software companies and massive resource allocation to firms like OpenAI. Now, Chinese investors show similar patience and recognition of AI's potential, feeding resources not only to Magic Atom but also to other teams exploring various software, data, and product directions.
Magic Atom prefers investors from industrial backgrounds, Gu shared, noting that many partners become shareholders after recognizing the company's potential during collaborations. Selection criteria focus on shared belief in robotics and AI, alignment with corporate values and long-term strategy, and steadfast support through development stages.
On revenue composition, Gu acknowledged that current industry patterns—heavily reliant on research scenarios and academic procurement—generally apply to Magic Atom. Confirmed 2026 orders are split evenly between humanoid robots and quadruped robots. Quadrupeds see mature applications in inspection, security, and emergency response with direct enterprise deliveries, while humanoids generate near-term revenue through entertainment, leasing, and guided services—a growing market across Asia, Europe, the Middle East, and the U.S.
"Robot adoption requires public brand recognition and interaction, making them feel integrated into daily life—conditions that now exist," Gu said. Performances serve not merely as commercial activities but as keys unlocking embodied intelligence's potential across industries, serendipitously revealing additional commercialization opportunities.
When asked if current revenues can support large-scale industrial adoption, Gu pointed to varying corporate philosophies. Conservative R&D spending aligns with available funds, underscoring the importance of capital market timing—securing external resources during high-need phases, then creating employment and shareholder returns upon commercialization.
Regarding industrial benchmarks—70% human labor replacement efficiency and over 95% success rates—Gu reported achievement in semi-structured specific scenarios. However, she clarified that single-scenario success represents phased progress rather than an endpoint, with generalization as the next challenge. Customer ROI must be calculable; while high-risk environments like high-voltage workshops justify single-scenario deployment, flexible production lines demand strong generalization capabilities for tasks like pick-and-place with variable object attributes—an area still under exploration.
On cost structures, Gu projected that key modules would see price reductions at scale. Hardware constitutes 70% of humanoid robot costs, offering significant reduction potential. Current dual-chip configurations suffice, though edge-computing needs may drive further chip cost optimization—an area where Magic Atom is pursuing strategic partnerships.
Regarding high-dexterity hands, Gu described cost management as a systemic challenge, with multiple players driving innovation. Magic Atom maintains an in-house dexterous hand team while evaluating external suppliers, balancing internal technical reserves with partner-supported mass production cost reductions.
Predicting industry consolidation, Gu acknowledged objective patterns resembling smile curves, driven by business model and technology convergence. Current diversity reflects genuine technical and product variations, with resource concentration expected as paths validate. Magic Atom positions itself as a permanent player, focusing on application acceleration and integrated hardware-software development.
Supply chain limitations appear manageable, Gu suggested, noting rapid industry growth exceeding expectations. Aggregate industry shipment volumes now attract major component suppliers, including listed companies with dedicated production lines. 2026 should see supply chain maturation benefiting both startups and upstream players.
For 2026 industry trends, Gu forecasted three developments: increased commercial activity and scenario integration as companies recognize application deployment as essential for physical AI viability; continued AI emergence dependent on algorithmic ceilings and integrated system design rather than isolated software or hardware advances, with capital influencing which approaches succeed; and commercial application shifts beyond large government orders, potentially reshuffling players.
She anticipated impressive model breakthroughs in late 2026, with leadership potentially rotating among developers. Market-wise, more companies beyond established players should demonstrate viable sales, robust hardware, and motion control capabilities—areas where Magic Atom has built solid teams, hoping to amplify its market presence this year.