One is the hottest "humanoid robotics stock" on the A-share market, the other is a pig farming enterprise. Comparing their R&D expense ratios side by side indeed presents an awkward picture.

Unitree Robotics reported full-year 2025 R&D expenses of 145 million yuan against revenue of 1.699 billion yuan, translating to an R&D intensity of 8.53%. In contrast, Muyuan Foods spent 1.648 billion yuan on R&D in 2025, with revenue around 144 billion yuan, resulting in an R&D intensity of just 1.14%. On a percentage basis, Unitree wins; its R&D ratio is over seven times that of the pig farmer, which might sound reasonable given the humanoid robotics sector.

But consider this from another angle: Muyuan's annual R&D spending of 1.648 billion yuan is eleven times greater than Unitree's 145 million yuan. A company in the pig farming business is outspending a robot manufacturer on research and development.

Why the Comparison Isn't Straightforward

Some might argue that comparing different industries is unfair—pig farming seemingly requires little R&D, while robotics is true hard tech. However, Muyuan's R&D expenditure has consistently risen over the past five years: 808 million yuan in 2021, 1.142 billion in 2022, 1.657 billion in 2023, 1.747 billion in 2024, and 1.648 billion in 2025. Cumulatively, that's over 7 billion yuan invested in R&D across five years.

What is Muyuan developing? Smart feeding systems, unmanned feed delivery vehicles, pig facial recognition, AI models for disease outbreak warnings, air-filtered pig houses, and waste-to-resource processing technologies. Qin Yinglin has transformed pig farming into a business of "industrialization plus digitalization." Muyuan's R&D isn't for show; it's aimed at lowering its "full cost." If the cost per kilogram of pork is one yuan lower than competitors', that translates into billions in annual profit differences.

Unitree's 2025 R&D investment of 145 million yuan doubled from 70 million yuan in 2024, but in absolute terms, it remains modest for the hard-tech arena. According to its prospectus, the R&D expense ratio for the first three quarters of 2025 was 7.73%, well below the industry peer average of 27.92%. This contrast highlights not who is superior, but a question of "stage."

Unitree's Strategy: Earning First, Spending Later

Unitree's current low R&D spending isn't due to a lack of need; it's a deliberate choice of "restraint" before its IPO. With full-year 2025 revenue of 1.699 billion yuan and net profit of 591 million yuan, the company has already crossed the breakeven point. Despite being profitable, allocating just 8.53% of revenue to R&D signals a strategy of "earning first, burning cash later."

The real "burn phase" arrives post-IPO. The prospectus reveals Unitree plans to raise 4.2 billion yuan from this IPO, with a hefty 85% earmarked specifically for R&D. The largest single allocation is 2 billion yuan directed toward large model development, representing 48% of total fundraising. To put that in perspective, Unitree's cumulative R&D over the past three years was only 265 million yuan (50 million in 2023, 70 million in 2024, and 145 million in 2025). The first R&D investment after the IPO is more than seven times the total of the previous three years combined.

Unitree has historically been a "hardware company," profiting from selling robots. The future vision is a "software-hardware integrated platform," where large models and embodied intelligence define the next generation of products. Transitioning from "selling bodies" to "building brains" requires an entirely different scale of R&D investment.

Two Divergent R&D Philosophies

Muyuan's R&D follows a "lean production" logic, extracting efficiency and reducing costs within a mature industry. Its 1.648 billion yuan in R&D, accounting for 1.14% of revenue, features a large absolute figure but a low relative ratio—a hallmark of "large industry, incremental innovation."

Unitree's R&D, conversely, follows a "from 0 to 1" logic, pursuing technological breakthroughs in an immature field. An 8.53% R&D intensity is indeed modest for the STAR Market, but if post-IPO the company can push this ratio above 30%, the 2 billion yuan large model investment becomes feasible.

One subtracts within an established framework to achieve lower costs and higher efficiency—Muyuan uses its 1.648 billion yuan R&D budget to turn pig farming into an "industrial product." The other adds in uncharted territory to make products smarter and more capable—Unitree uses its 145 million yuan to first sell robots, then leverages IPO capital to develop the "brain."

Neither path is right or wrong; they simply represent different stages. What remains certain is this: if Unitree fails to elevate its R&D investment scale post-IPO, or if that 2 billion yuan large model allocation turns out to be merely a PowerPoint promise, its "first humanoid robotics stock" narrative will eventually be repriced by the market. The capital market may accept "spending less now," but it won't forgive "not spending when it's necessary."

From "selling bodies" to "building brains," Unitree's R&D ledger has just opened a new chapter. Whether it's written well is too early to judge, but whether that "R&D leap curve" from 8.53% to 30% can be drawn will soon become evident.