China Plans $295B AI Grid, Locking Out Nvidia

China has drafted a 2 trillion yuan plan — $295 billion at current exchange rates — to build a unified national AI computing grid by 2028. The mandate requires that at least 80% of the AI accelerator chips powering the infrastructure come from domestic suppliers. That requirement, if enforced, effectively terminates Nvidia and AMD's access to China's AI build-out market and positions Huawei, Alibaba Cloud, Biren Technology, and Moore Threads as the primary hardware beneficiaries of the largest state-directed AI infrastructure investment in history.

For context: the United States' Stargate initiative announced in January 2026 committed $500 billion over four years in private investment. China's plan is $295 billion over five years in government-directed spending. The scale is comparable. The architecture is different. And the constraint it places on domestic chip producers may be the biggest challenge the plan faces.

What China Is Actually Building

The plan, drafted by China's National Development and Reform Commission, calls for connecting thousands of data centers into a single unified computing grid. The facilities will be operated primarily by state carriers China Mobile and China Telecom, which provide the network fabric to stitch regional data centers into a nationally accessible compute resource.

The target is 2028 for full interconnection. The individual data centers are already under construction in several phases across Beijing, Shanghai, Chengdu, and the designated western computing zones that China has been developing since 2021 under the "East Data West Computing" initiative.

The scale numbers: the grid is designed to provide AI compute as a national resource, accessible to domestic companies, research institutions, and government agencies through a unified API layer. Think of it as a national-scale cloud, owned by the state, built on domestic silicon, and interoperable across carriers.

Funding flows primarily through sovereign debt and ultra-long special government bonds — a financing mechanism China has used for infrastructure at scale in previous programs. When power-grid integration is included (the electrical infrastructure required to run facilities at this density), the total projected investment reaches at least 5 trillion yuan, approximately $740 billion.

The 80% Domestic Chip Mandate

The most consequential element of the plan is the chip sourcing requirement. At least 80% of AI accelerator hardware in the national grid must come from domestic suppliers. The remaining 20% can come from international sources, but the mandate effectively structures the entire build-out around Chinese silicon.

This is not a new direction — it is the acceleration of a policy framework that has been building since 2022. In May 2026, Beijing formally approved nine categories of domestically developed AI chips for deployment across government and security-sensitive sectors. That approval effectively cleared Huawei's Ascend series, Alibaba's Hanguang 800, Biren Technology's BR100 and BR104, and Moore Threads' MTT S80 for procurement under government frameworks.

The mandate does three things simultaneously:

It eliminates Nvidia and AMD from the primary tender process for the national grid. Both companies are US-headquartered, and their flagship AI accelerators (Nvidia H100, H200, B100; AMD MI300X) remain subject to US export controls that prohibit their sale into China for AI applications above specific capability thresholds. Even if the mandate did not exist, those export controls already restrict the highest-performing Nvidia and AMD chips from the Chinese market.

It creates guaranteed procurement volume for domestic chip producers at a scale they have never had before. Biren Technology, founded in 2019, and Moore Threads, founded in 2020, are pre-revenue in meaningful volume at the enterprise level. A national grid mandate changes their market structure overnight.

It accelerates the timeline for Chinese chips to close the performance gap through volume production experience. Chip performance is partly a function of the number of wafers manufactured — more production means better yields, better quality control, and faster iteration. State procurement at this scale is an industrial policy for semiconductor maturation as much as it is an infrastructure program.

Who Wins: The Domestic Chip Stack

Huawei is the most advanced domestic AI chip producer. Its Ascend 910B chip, produced on SMIC's 7nm process, is the current benchmark for domestically produced AI accelerators. Benchmarks published by domestic researchers in 2025 showed Ascend 910B at roughly 60-70% of H100 performance on standard transformer training workloads — a meaningful gap, but a smaller one than existed two years earlier.

The constraint on Huawei's production is HBM (High Bandwidth Memory). High-bandwidth memory is the stacked memory that sits alongside the AI accelerator die and is critical to performance on attention-heavy workloads. SK Hynix, Micron, and Samsung produce the world's HBM supply, and all three are US-allied companies subject to export control frameworks. China's domestic HBM capacity is minimal. Huawei can design accelerators; it cannot easily fill the memory requirement without imported HBM, and that dependency constrains how many Ascend chips it can assemble.

Biren Technology holds the BR100, a 7nm chip targeting HPC and AI workloads. Early benchmarks suggest competitive performance with older Nvidia A100-class hardware. Biren received government security clearance in May 2026 and is now eligible for national grid procurement.

Moore Threads focuses on graphics and AI inference rather than training. Its MTT S80 chips are positioned primarily for inference deployment, which matters because China's build-out requires both training infrastructure (building models) and inference infrastructure (serving them). The national grid will need both layers.

Alibaba Cloud produces the Hanguang 800, optimized for inference on Alibaba's specific workload profile. It is less of a general-purpose AI accelerator and more of a tailored inference chip. Its relevance to the national grid depends on whether inference workloads are contracted to Alibaba's cloud infrastructure.

The Constraint: HBM and the 5-10 Year Gap

Chinese chip executives have publicly conceded the country trails the leading edge in AI data center silicon by five to ten years. That gap is not primarily in chip design — it is in the manufacturing process and in the memory stack.

TSMC at 3nm and Samsung and SK Hynix at advanced HBM3/HBM3e represent production capabilities China cannot currently replicate domestically. SMIC, the leading Chinese foundry, produces at 7nm and is expanding to 5nm-equivalent through process refinement. That is meaningful progress, but not catch-up.

The $295 billion mandate accelerates procurement of what China can make now. It does not eliminate the underlying technology gap. A national grid built on 7nm-class silicon and constrained HBM supply will provide substantially less compute per dollar than the equivalent US infrastructure running Nvidia Blackwell chips with full HBM3e memory stacks.

The workaround that China has demonstrated at the model level is inference efficiency. DeepSeek's R2 series achieved competitive performance on reasoning benchmarks against GPT-5 while using substantially less compute per inference call. If Chinese model development continues to optimize for hardware-constrained inference, the gap in compute capacity matters less for the applications that the national grid serves.

How This Compares to US AI Investment

The comparison is straightforward in dollar terms and complex in structure:

The structural difference: US AI investment is private capital chasing returns. China's investment is state capital pursuing strategic capacity. Both are building compute at scale. Only one of them is doing it with hardware constraints built into the mandate from day one.

Developer Impact: Two Compute Universes

For developers and companies building AI applications, the China national grid plan formalizes something that has been true informally since 2022: there are now two separate global AI infrastructure stacks, and they are diverging.

Applications built for US and allied-market deployment run on Nvidia hardware, use US-headquartered API providers, and operate under export control frameworks that assume chip access. Applications built for deployment in China run on domestic silicon, connect to state-operated compute resources, and operate under data localization requirements that assume state access.

These two stacks are not interoperable in any meaningful way. A model trained on Nvidia hardware and served through OpenAI's API cannot simply be ported to the national grid. A model trained on Ascend hardware and optimized for domestic inference chips will require re-work to run at equivalent performance on Nvidia infrastructure.

For companies operating in both markets, this means maintaining two parallel technical stacks — a cost that small teams cannot absorb and that even large teams find operationally complex. The $295 billion plan accelerates the divergence. It is not just an infrastructure investment. It is a structural separation of the global AI market into two compute blocs.

The pricing implications for the domestic Chinese market: state-subsidized compute at scale will likely be offered to Chinese companies at below-market rates, creating an asymmetric competitive environment for AI startups in China versus those in the US or EU. Track pricing through LLM API pricing comparisons as Chinese model providers begin offering internationally accessible inference endpoints.

Our Analysis

The $295 billion China AI grid plan is being covered primarily as a number — $295 billion is large, so it gets attention. The more important element is the 80% domestic chip mandate.

That mandate is simultaneously an industrial policy and an admission. It is an industrial policy because it creates guaranteed demand for an industry that cannot yet compete on performance. It is an admission because it reflects the reality that China cannot access the chips it would prefer — Nvidia Blackwell-class hardware — due to US export controls, so it is structuring the entire build-out around what it can manufacture.

DeepSeek proved at the model level that compute-constrained inference can produce competitive results. The national grid is a bet that the same logic applies at the infrastructure level — that building a large, interconnected compute grid on constrained hardware is better than waiting for the hardware constraint to resolve.

That bet may be correct on its own terms. A national grid with 80% domestic silicon and constrained HBM is still more AI compute capacity than China has today. It creates the volume production experience that accelerates chip maturation. It reduces China's vulnerability to further export control escalation.

The bet may not be correct on the global competition terms. If Nvidia's Blackwell B200 and the successor generation widen the performance gap faster than Chinese domestic chips can close it through volume production, the national grid is a large investment in second-tier compute. The $740 billion total investment (including power grid) could fund a significant amount of chip R&D if redirected — but infrastructure and chip design are separate industrial problems, and neither can fully substitute for the other.

The developer takeaway: the two-stack world is now a planned infrastructure reality, not just a trade policy effect. Build accordingly.

Key Takeaways

• $295 billion over 5 years — China's national AI data center grid plan, funded via sovereign debt; power grid integration brings total to $740 billion
• 80% domestic chip mandate — effectively locks Nvidia and AMD out of the primary procurement process for the largest state AI infrastructure build in history
• 2028 interconnection target — China Mobile and China Telecom operate the facilities; NDRC directs the build-out
• Domestic winners — Huawei Ascend series, Biren BR100, Moore Threads MTT S80, Alibaba Hanguang 800 all cleared for national grid procurement after May 2026 government approval
• The HBM constraint — China trails leading-edge AI silicon by 5-10 years; limited domestic HBM supply is the primary bottleneck on Huawei Ascend production volume
• Two compute universes — the plan formalizes the separation of global AI infrastructure into US-aligned and China-aligned stacks; developers in both markets now face parallel technical requirements

Sources

• Tom's Hardware: China $295B national AI data center grid plan
• TechTimes: China AI Data Center Grid Locks Out Nvidia With $295 Billion Mandate
• MLQ News: China $295B domestic chip mandate plan
• TechRadar: China $295B AI grid and domestic silicon strategy