Rules are set for both brain-computer interface and humanoid robots at the same time: the next patent war starts with standards.
The technology has not been fully finalized, yet the selection of standards has already begun to take shape.
On August 25, the Ministry of Industry and Information Technology of the People's Republic of China simultaneously released two documents for public comments: the "Guidelines for the Construction of the National Brain-Computer Interface Industry Standard System (2026 Edition)" and the "Guidelines for the Construction of the National Humanoid Robot Industry Standard System (2026 Edition)".
According to the two draft documents for soliciting public opinions, by 2028, China will develop and revise more than 40 standards in the brain-computer interface field, and take the lead in or participate in the formulation of more than 10 international standards; in the humanoid robot sector, at least 100 key standards will be completed. The former plan aims to promote the implementation of relevant standards by more than 100 enterprises, while the latter will cover more than 200 enterprises.
On the surface, this marks that these two future-oriented industries are beginning to "establish rules"; looking deeper, it means that the competition in brain-computer interfaces and humanoid robots has moved ahead of schedule from the stage of technology, product and financing competition to a phase where standards and patents are deeply intertwined.
The next patent war will most likely not start in court, but at standard-setting meetings.
I. Why formulate standards first when products are not yet mature?
Whether it is brain-computer interfaces or humanoid robots, both are currently in a stage of rapid evolution of technical routes, and the industrial chain has not yet been fully finalized.
Brain-computer interfaces involve electrodes, chips, signal acquisition, codec algorithms, data transmission, stimulus feedback, clinical evaluation and safety governance; humanoid robots involve brain-like computing, joints and actuators, sensors, control systems, operation platforms, complete machine evaluation, scenario application and ethical safety.
Different enterprises can develop their own products, but if interfaces cannot be connected, data cannot be interoperated, performance cannot be compared, and there is no bottom line for safety, it will be difficult for the industry to form a real large-scale market.
A brain signal acquisition device developed by one enterprise may not be compatible with the analysis software of another enterprise; a robot joint produced by one manufacturer may only access its own control system; different enterprises have inconsistent definitions of "load capacity", "motion accuracy" and "environmental adaptability", so purchasers cannot even judge which robot is better.
Therefore, the earlier a future-oriented industry is in its development stage, the more urgent it is to establish a common language as soon as possible.
The draft for soliciting opinions on humanoid robots proposes to establish standards around basic commonality, brain-like and intelligent computing, limbs and components, complete machines and systems, applications, safety ethics and other fields, and standardize capability grading, unique identifiers, coding rules, industrial environment operation capabilities and ethical impact assessment.
The brain-computer interface standard system covers seven aspects: basic commonality, hardware, software and algorithms, data and communication, products and systems, industry applications, and safety and governance.
The primary goal of these standards is to solve problems including whether products can be connected, how capabilities are evaluated, how data flows, and how risks are controlled. They are also the necessary infrastructure that must be completed before the industry moves from laboratories to factories, hospitals and consumer markets.
Without standards, enterprises can develop products on their own; only when standards are formed can products be integrated into the whole industry.
II. Standards select not only parameters, but also technical routes
A standard is never a neutral product specification.
When a standard stipulates an interface form, a set of communication protocols, a data format or a test method, it is actually making a choice between different technical solutions. Once a certain technology is written into the standard, it may evolve from the internal solution of one enterprise to the common solution adopted by the entire industrial chain.
This is exactly where standards and patents begin to interact.
Patents grant the right holder exclusive rights within a certain period of time, while standards require market participants to abide by common technical rules. When the technical solution covered by a patent becomes an unavoidable content in the implementation of a certain standard, it may become a standard essential patent (SEP).
In the past, standard essential patents were mainly concentrated in the fields of mobile communication, Wi-Fi and audio and video codec. Mobile phone manufacturers, chip enterprises, television enterprises and video platforms have long-term games around patent licensing royalty rates, licensing hierarchies and injunction relief.
Although brain-computer interfaces and humanoid robots are not traditional communication industries, they also have high demands for interconnection and compatibility.
The connection method between brain-computer interface devices, the data format of neural signals, the acquisition and transmission protocols, and the interoperability mechanism between devices and software platforms may all involve patented technologies. The component interfaces, perception fusion, motion control, operating systems, middleware and data communication in humanoid robots may also form technical solutions commonly adopted by the industry.
If these solutions are eventually incorporated into national standards, industry standards or international standards, the value logic of relevant patents will change.
What it protects may no longer be just one product, but a technical path adopted by the entire industry.
III. Standards do not equal standard essential patents
However, it needs to be specially emphasized that: The formulation of more than 140 standards does not mean that more than 140 groups of standard essential patents will definitely be generated in the future, nor does it mean that enterprises participating in standard formulation naturally have the right to charge the industry.
Terminology definitions, ethical requirements, safety bottom lines and some test evaluation standards do not necessarily need to adopt technical solutions protected by patents. Even if a certain standard involves patents, whether the patent is "essential" still depends on whether there is an alternative way to avoid it when implementing the standard.
A real standard essential patent needs to meet at least several conditions at the same time: the patent right is valid, the claims can cover the technical solution stipulated by the standard, and there is no feasible non-infringing alternative when implementing the standard.
Therefore, "participating in standard formulation", "submitting technical proposals", "having patents mentioned in standard documents" and "owning standard essential patents" are four different concepts that cannot be confused.
China's current rules also require that organizations or individuals participating in the formulation and revision of national standards shall disclose the essential patents they own and are aware of as early as possible; patent holders or applicants shall also make corresponding implementation licensing declarations. The "Guidelines for Invention Patent Applications Involving Standards" issued by the State Intellectual Property Office this year further proposes that enterprises should carry out patent layout following the standardization process: carry out initial layout at the standard proposal stage, carry out detailed layout at the drafting, deliberation and public comment stages, and timely adjust the claims of pending patents at the standard approval and release stage.
This shows that standards and patents can no longer be separately managed by two departments that do not communicate with each other.
If a patent application is filed too late, the technical solution may have lost its novelty due to public disclosure; if the application scope is too narrow, it may not be able to cover the final formal standard; if the application is filed too early, it may lose value due to changes in the standard route.
Standard essential patents are not "applied for out of nothing", but formed after the patent layout is continuously calibrated with the evolution of standards.
IV. What enterprises are really competing for is not just the number of standards
Facing the standard formulation work that will be launched intensively in the next few years, the most common misunderstanding that enterprises are prone to fall into is taking "how many standards they participate in" and "how many patents they submit" as the final achievement.
However, the licensing practices of the communication and audio-video industries over the past years have proved that what really determines the industry status is not the total number of patents, but the patent quality, standard coverage, global patent family layout and licensing operation capability.
One patent that can accurately cover key interfaces and core protocols may be more valuable than hundreds of peripheral patents; one patent that is only authorized in China can hardly support enterprises to participate in global licensing negotiations.
For enterprises engaged in brain-computer interfaces and humanoid robots, at least four tasks should be completed in advance now.
• First, promote the R&D route, patent layout and standard proposal synchronously. Engineers participating in standard work should not only be responsible for technical exchanges, and the intellectual property team should not wait until the standard is released to retrieve relevant patents.
• Second, establish the corresponding relationship between standard clauses and patent claims. Enterprises need to continuously judge whether their own patents can cover the draft standards, and whether competitors' patents may become unavoidable technical thresholds.
• Third, lay out overseas patents in core markets in advance. Once humanoid robot and brain-computer interface products enter the international market in the future, licensing and litigation will not only occur in China. Only by forming a stable global patent family can enterprises obtain real negotiating chips.
• Fourth, study the patent disclosure, licensing commitment and interest balance mechanism in advance. The purpose of incorporating patents into standards is not to form new technological monopolies, but to strike a balance between protecting innovation returns and reducing the industry's implementation costs.
V. The next competition is to turn technology into rules
The simultaneous launch of the standard system construction for brain-computer interfaces and humanoid robots indicates that these two industries are crossing the single-point technology verification stage and starting to build a common foundation for large-scale application.
For Chinese enterprises, this is a rare time window.
In the past, in some mature industries, when Chinese enterprises entered the market, international technical standards, core patent portfolios and licensing rules had basically taken shape. Even with manufacturing capabilities and market scale, enterprises often could only passively accept the existing licensing system.
Brain-computer interfaces and humanoid robots are different. The global technical routes have not yet fully converged, product forms have not been finally determined, and the standard system is still in the process of formation. Chinese enterprises not only have the opportunity to manufacture products, but also the opportunity to submit technical solutions, participate in standard formulation and build corresponding high-quality patent portfolios.
However, this competition cannot repeat the old path of "a large number of standards, a large number of patents, but not enough truly essential patents".
The value of standards does not lie in how many documents are released; the value of patents does not lie in how many certificates are obtained. What really matters is whether Chinese enterprises can turn original technologies into solutions commonly adopted by the whole industry, and transform technological advantages into rule-making capabilities that can sustainably participate in global competition.
Technology determines whether an enterprise can enter the track, standards determine where the track extends, and patents determine who has a longer-term right to speak on this track.
The next patent war for brain-computer interfaces and humanoid robots has already begun.
This article is from the WeChat official account "Zhichanli" (ID: zhichanli), written by Mao Liguo, and published with authorization from 36Kr.