MIIT Sets a 2030 Network Deadline, but 6G Is Only Part of the Plan

China's Ministry of Industry and Information Technology, or MIIT, set a 2030 deadline for building a nationally integrated next-generation communications network. The September 7 plan promises broad coverage and leading performance across terrestrial, satellite, computing, and industrial systems. Its central conflict is already visible: infrastructure deployment can move quickly, but global standards, viable applications, and operating economics develop on different schedules.

The announcement is not simply a promise to replace 5G with 6G. MIIT is defining a much larger network that combines 5G-Advanced, future 6G services, fiber, satellite links, computing infrastructure, and specialized industrial networks. It also brings sensing, computing, artificial intelligence, and security into a system once judged mainly by connection speed.

That scope puts pressure on China's state-controlled carriers, equipment suppliers, cloud operators, and industrial customers. They must turn an ambitious construction program into services that enterprises will actually adopt. Meanwhile, international standards bodies still need to finish the technical rules that make interoperable 6G equipment possible.

The 2030 Plan Sets Measurable Targets

MIIT has converted its network vision into a construction and adoption program with explicit national targets.

The ministry released its information and communications industry plan for the 2026 to 2030 period on September 7, 2026. That date is confirmed by a same-day dispatch describing the five-year blueprint, not merely by the hot-list headline that surfaced the story.

The plan contains 13 major indicators and 26 priority tasks across six policy areas. Those areas cover infrastructure, the communications industry, governance, network and data security, broader applications, and international development.

By the end of the period, MIIT targets 4.1 trillion yuan in annual information and communications industry revenue. It also calls for 3.8 trillion yuan in cumulative information infrastructure investment and average annual telecom business-volume growth of 7 percent.

The adoption targets are equally aggressive. MIIT wants 5G and 5G-A penetration to reach 95 percent, with 50 compatible base stations for every 10,000 people. The plan also specifies 500,000 newly built 5G-A base stations.

5G-A, also known as 5G-Advanced, is the enhanced phase of 5G that adds greater capacity, positioning precision, reliability, and support for specialized devices. It serves as a bridge between current mobile networks and future 6G systems.

The computing target reveals how far the definition of a communications network has expanded. MIIT wants China's intelligent computing capacity to reach 9,800 EFLOPS. One EFLOPS represents one quintillion floating-point operations per second, although reported capacity can vary with numerical precision and measurement methodology.

The plan proposes a multi-level computing architecture spanning national hubs, regional facilities, and edge infrastructure. Large clusters containing 10,000 or even 100,000 accelerators would support AI training. More distributed facilities would place inference capacity closer to factories, applications, and users.

Other tasks include broader international internet capacity, integrated satellite and terrestrial coverage, direct-to-phone satellite services, embedded SIM adoption, and communications that work without conventional network access. MIIT also calls for an identification system that allows AI agents to locate and interact with one another across networks.

This combination matters because none of the targets stands alone. A factory using machine vision needs radio coverage, predictable latency, local computing, secure data movement, and software integration. A satellite-connected handset needs compatible spectrum, devices, gateways, billing systems, and regulatory approval.

The plan therefore treats connectivity as a coordinated stack. Its success depends less on announcing an individual radio technology than on making several infrastructure layers function together.

AI Has Changed What a Communications Network Must Do

The plan responds to an AI workload problem as much as it responds to demand for faster mobile broadband.

Traditional telecom networks moved voice, messages, and internet traffic between people and servers. MIIT's next-generation model extends the set of participants to machines, vehicles, robots, sensors, and embodied AI systems.

That changes the required service. A video stream can tolerate small changes in network performance because software buffers incoming data. A production robot or remotely controlled machine often needs more stable latency and predictable availability.

MIIT describes this shift as a move from simple information transmission toward a combination of communication, sensing, computing, intelligence, and security. Sensing means network signals can help determine an object's location, movement, or surrounding conditions without relying only on a separate sensor system.

The ministry had outlined this broader definition before the September release. At a July briefing, officials said next-generation networks would expand beyond people, extend from terrestrial coverage into air, space, land, and sea, and provide more deterministic connections.

That earlier network strategy also placed 6G, ultra-fast optical communications, satellite systems, industrial networks, and computing coordination within one infrastructure program. The new plan gives that strategy targets and a deadline.

AI creates demand at both ends of the network. Training clusters require high-capacity links between accelerators, storage, and data centers. Inference systems need to move requests and results while managing latency, cost, privacy, and power consumption.

The plan's proposed network for AI agents pushes the idea further. Software agents will need identities, discovery mechanisms, authorization rules, and reliable ways to exchange tasks. Those functions resemble internet protocols, but they must also address the risks of autonomous software making decisions across organizational boundaries.

For enterprises, the immediate change will not arrive as a single 6G upgrade. It will appear through denser 5G-A coverage, faster fiber access, edge computing, satellite availability, and more integrated carrier services.

Manufacturing is a leading test case. China already uses private 5G networks in mines, ports, factories, and power facilities. A next-generation architecture could connect mobile robots, inspection cameras, digital twins, and production-control systems through shared infrastructure.

Healthcare, transportation, emergency response, and agriculture present similar opportunities. However, each sector brings different requirements for availability, liability, privacy, and equipment life cycles. A carrier cannot treat those customers like smartphone subscribers.

The shift will also create a documentation problem inside companies. Network diagrams, device records, security policies, vendor manuals, and incident notes often sit across separate systems. An engineering knowledge base can help teams connect those records when infrastructure becomes more complex.

The larger point is that AI infrastructure does not end at the data-center door. Models need networks that can move workloads, coordinate machines, and preserve service quality under real conditions. MIIT is betting that communications policy must now cover that entire path.

The Real Contest Is Ambition Versus Deployability

The primary tension is not China versus another country; it is the gap between a coordinated national blueprint and commercially deployable services.

China has already shown that national infrastructure targets can mobilize carriers, local governments, equipment companies, and construction resources. Its 5G rollout produced extensive base-station coverage and brought service to cities, towns, and most administrative villages.

That experience gives the new program a credible implementation mechanism. MIIT can coordinate spectrum, technical trials, investment priorities, operator obligations, and local pilot programs. State-backed carriers can make long-term investments that might appear unattractive under a shorter financial horizon.

The 2030 network plan also avoids depending on a single uncertain technology. If 6G standardization or device availability slips, 5G-A, fiber, satellite services, and computing networks can still deliver measurable progress.

That flexibility is a strength, but it also makes success harder to judge. A broad collection of upgraded systems can satisfy the phrase "next-generation network" without proving that the components work as one coherent platform.

The investment target presents the same ambiguity. Infrastructure spending can measure activity, but it does not automatically measure productive use. Operators still need revenue from consumers, enterprises, government agencies, or wholesale customers to sustain new facilities.

5G provided an important warning. Better coverage and higher peak speeds did not instantly produce a consumer application comparable to the smartphone transition triggered by 4G. Much of 5G's strategic value has emerged through industrial networks and network capacity rather than a single mass-market service.

6G faces an even larger expectation burden. Proposed capabilities include ubiquitous coverage, integrated sensing, AI-native operations, immersive communications, and extremely reliable low-latency links. Each one has different hardware, spectrum, power, and business requirements.

Global coordination remains essential. The International Telecommunication Union uses the name IMT-2030 for the family of standards commonly called 6G. In March 2026, experts agreed on draft technical requirements covering 20 performance measures and six usage scenarios.

Those requirements provide a basis for evaluating future radio technologies. They do not guarantee real-world performance, assign national spectrum, or establish a complete commercial standard.

Formal approval of the requirements remained scheduled for December 2026 when the ITU published its March update. Candidate technologies, evaluation work, detailed industry specifications, equipment testing, and device integration must follow.

China has started creating room for that work. In May, authorities approved 6 GHz trials for selected regions and the IMT-2030 promotion group. The trials focus on research and validation against internationally defined scenarios and indicators.

Trials are necessary, but they are not commercial networks. Laboratory throughput can fall when radio signals encounter walls, weather, interference, mobility, crowded cells, or power limits. Satellite links introduce additional questions involving orbital capacity, terminals, handoffs, and spectrum coordination.

The plan's central promise will therefore be tested by integration. Operators must show that terrestrial radio, fiber, satellites, computing centers, and industrial systems can share management and deliver consistent services.

A network can be technically advanced but economically underused. It can be nationally extensive yet fragmented between vendors and industries. It can also meet average coverage targets while leaving expensive gaps in remote areas or indoor environments.

The blueprint creates a strong direction of travel. Deployability will determine whether that direction produces durable infrastructure or a collection of impressive but uneven projects.

Carriers and Suppliers Now Carry the Execution Risk

The policy shifts pressure from central planning toward the organizations that must finance, install, operate, and sell the network.

China Mobile, China Telecom, and China Unicom will bear much of the implementation load. They must continue expanding 5G-A while preparing for 6G, upgrading fiber, adding computing capacity, supporting satellite integration, and maintaining existing networks.

These investments compete for capital and operational attention. A new base station requires more than radio equipment. It needs a site, power, backhaul, maintenance, software, security controls, and integration with the core network.

Dense urban areas can justify upgrades through large customer volumes. Rural, maritime, airborne, and remote industrial coverage often has weaker commercial returns. Satellite integration can close some gaps, but it adds another infrastructure layer rather than eliminating cost.

Equipment suppliers such as Huawei and ZTE face a different pressure. They must turn research into products while standards remain under development. Early investment can build technical influence, but premature designs can require costly revision.

Domestic chip and optical-component suppliers also sit inside the plan's security objective. MIIT wants an advanced and controllable communications industry, language that signals continued emphasis on local technology capacity and supply-chain resilience.

That goal interacts with export controls and restricted access to some advanced semiconductor technologies. Building domestic alternatives can reduce dependency, but performance, manufacturing yield, software compatibility, and energy efficiency all affect practical deployment.

Cloud and data-center operators must address the computing target. Installing accelerator clusters is only the first step. Operators need electricity, cooling, network bandwidth, software scheduling, customers, and sufficient utilization to support ongoing costs.

The planned hub, regional, and edge hierarchy adds another challenge. Moving an AI task to the right facility requires orchestration across networks and computing resources. Poor scheduling can leave one cluster congested while another remains underused.

Enterprises face their own adoption decision. A factory will not replace stable industrial equipment simply because a carrier offers a newer network. It will demand evidence that the system improves output, safety, maintenance, flexibility, or total operating cost.

Many industrial assets remain in service for years. Networking upgrades must coexist with legacy control systems, proprietary protocols, and strict production schedules. Downtime during migration can erase the expected benefit.

Security becomes more complicated as the network gains functions. A communication system that also senses objects, coordinates AI agents, and controls machines creates more valuable data and more potential attack paths.

MIIT's plan addresses network and data security, critical infrastructure protection, fraud prevention, and emergency communications. Yet the effectiveness of those measures will depend on implementation across thousands of organizations.

AI agents make identity and authorization especially important. A device or software process must prove what it is, what it can access, and which actions it can perform. Compromised credentials could allow an attacker to move beyond stealing data and trigger physical or operational consequences.

Satellite-terrestrial integration introduces further boundaries. Operators must secure terminals, gateways, orbital links, ground networks, and handoff procedures. Failures can involve several providers with different responsibilities.

There is also an energy tradeoff. More base stations, computing clusters, edge facilities, and satellite gateways consume electricity. AI optimization can improve network efficiency, but it does not erase the energy required by additional capacity.

These constraints do not invalidate the plan. They define its real work. Construction targets can drive supply, while enterprise adoption, utilization, reliability, and operating efficiency reveal whether the supply has lasting value.

6G Is a Milestone, Not the Whole Network

Treating this policy as a 6G launch story would obscure the technologies expected to carry most traffic during the transition.

The plan says China will begin 6G commercial use at an appropriate time. That wording preserves flexibility because commercial readiness depends on global standards, available spectrum, tested equipment, compatible devices, and viable services.

MIIT officials have described 6G as a top priority for the five-year period. Their stated work includes core technology research, technical trials, standard development, and preparation for commercialization.

That sequence matters. Research explores what is technically possible. Trials test candidate systems. Standards create common specifications. Commercialization requires equipment and services that customers can buy and operate.

The ITU framework includes immersive communication, hyper-reliable low-latency communication, massive communication, ubiquitous connectivity, AI and communications integration, and integrated sensing. These categories describe desired outcomes, not finished products.

Integrated sensing illustrates the opportunity and the risk. A radio network could detect motion or location while carrying communications traffic. This might support transportation, industrial safety, logistics, and emergency services.

The same capability raises privacy and governance questions. Organizations will need rules covering when sensing occurs, what data it creates, who can access that data, and how long records remain available.

Ubiquitous connectivity also sounds simpler than it is. Extending service across land, sea, air, and space requires cooperation between terrestrial carriers, satellite operators, device makers, aviation and maritime authorities, and spectrum regulators.

The plan's satellite provisions include broader global coverage and direct connections between satellites and ordinary mobile devices. These services could improve emergency access and fill coverage gaps without requiring a terrestrial tower everywhere.

However, direct-to-device satellite service usually offers less capacity than a dense ground network. It is best viewed as a complementary coverage layer, especially during its early deployment.

5G-A will therefore do much of the near-term work. Its deployment can improve carrier aggregation, reduced-capability device support, positioning, industrial reliability, and other functions without waiting for full 6G specifications.

Fiber remains even more fundamental. Mobile base stations, satellite gateways, computing clusters, and enterprise facilities need high-capacity backhaul. Faster radio access cannot solve congestion deeper in the network.

The computing network adds the third major layer. AI services depend on moving data and tasks between centralized training clusters, regional facilities, and edge systems. The best location depends on latency, privacy, energy, capacity, and cost.

A factory inspection model might run near the production line because delays disrupt operations. A large model-training job can run at a distant national hub because it values accelerator scale over immediate response.

MIIT's approach tries to coordinate these decisions. Its wider AI and communications policy calls for research spanning 5G-A, 6G, optical networks, IPv6-based systems, industrial internet technologies, space computing, and AI-driven network architecture.

The practical advantage of this layered strategy is continuity. Operators can deploy useful infrastructure before 6G arrives, then reuse fiber, data centers, sites, security systems, and operating experience during later upgrades.

The risk is branding ordinary upgrades as proof of a new network. Faster fiber, more accelerators, and denser 5G-A coverage are meaningful, but they do not independently demonstrate an integrated architecture.

The policy should be judged by end-to-end results. Can a service move between satellite and terrestrial coverage without disruption? Can computing workloads shift between national and edge facilities? Can industrial systems receive guaranteed performance under congestion?

Those questions separate an infrastructure inventory from a functioning network. The 6G label attracts attention, but integration will determine value.

Three Signals Will Show Whether the Deadline Holds

The next evidence should come from standards, deployment quality, and enterprise use, not from another high-level announcement.

The first signal is the formal progress of IMT-2030 requirements and subsequent global standardization. International agreement will narrow the range of possible radio designs and help equipment companies avoid incompatible investments.

If approval and evaluation work proceed on schedule, MIIT's commercialization language gains credibility. Delays or major disagreements over spectrum and technical requirements would weaken the prospect of coordinated deployment near the deadline.

The second signal is how operators deploy the promised 500,000 new 5G-A base stations. The headline count matters less than geographic distribution, traffic growth, service quality, and utilization.

A dense network that carries industrial traffic and supports measurable improvements would strengthen the plan's economic case. Low utilization or concentration in showcase districts would suggest that construction is moving faster than demand.

The third signal is enterprise adoption across manufacturing, transportation, energy, healthcare, and emergency communications. Useful indicators include connected production systems, paid contracts, renewal rates, workload volume, and documented reliability.

Successful pilots are not sufficient. Enterprises must move from demonstrations into recurring operational use. They also need evidence that benefits exceed integration, security, maintenance, and equipment costs.

The computing target deserves similar scrutiny. Reported EFLOPS capacity should come with information about precision, availability, energy use, network access, and utilization. Without those details, a large aggregate number reveals little about effective AI capacity.

Satellite services should be measured through device compatibility, geographic availability, connection success, capacity, and emergency performance. Announced coverage does not always equal a dependable user experience.

Security incidents will provide another indirect test. A network connecting machines and AI agents must contain compromised devices, protect sensitive data, and recover without cascading disruption. Resilience is an operational property, not a planning phrase.

The MIIT 2030 network plan has a plausible foundation because it builds on existing 5G, fiber, industrial, and computing infrastructure. Its deadline still combines several programs that mature at different speeds.

Watch what gets standardized, what operators deploy outside demonstrations, and what enterprises repeatedly pay to use. Those signals will show whether China is building an integrated network for the AI era or simply assembling its components under one national target.