Computing industry · Green computing · Zero-carbon computing · Compute-power synergy

The faster compute runs,the clearer power and carbon must be

MeetCarbon Cloud gives computing facilities one ledger for energy, carbon, compute and water. It connects green-facility applications, zero-carbon construction, green power sourcing, compute-power synergy and intelligent O&M, so you get on the list, stay on it, and cut your power bill.

Compute
Power
Carbon
Water
MeetCarbon CloudOne ledger for energy, carbon, compute and water

80%

Green power share for new data centers in national hub nodes, 2026

1.25 / 1.2

PUE cap for new large data centers / hub projects

Nov 16

Deadline for 2026 national green computing facility applications

Macro · Policy window

Policy requirements have moved from PUE to five dimensions

In 2026, compute-power synergy entered the national new-infrastructure program. The four-ministry action plan, green computing facility selection and zero-carbon factory construction followed. Green power share, carbon accounting, carbon and water efficiency, flexible load and dynamic list management are now assessed together.

0%

Green power share for new hub data centers

Set uniformly for all provinces in 2026

≤0.00

PUE for new large data centers

1.2 for national hub projects

≥0%

Rack utilization

Entry threshold for green computing facilities

0

Tasks in the four-ministry action plan

国能发科技〔2026〕34号

Key policy timeline

Click a milestone for highlights and the original text

  1. 2023-12
  2. 2024-07
  3. 2025-03
  4. 2025-07
  5. 2026-04
  6. 2026-05
  7. 2026-09
  8. 2026-12

Local policy · Nationwide

Hubs set the green floor, cities fund demand

Western hub clusters write green-power ratios and PUE into entry thresholds. Eastern and central cities cut the cost of compute with computing and training vouchers. Key published provisions by region.

14 regions

Pick a theme to highlight matching provisions

  • Inner Mongolia

    Horinger cluster

    Hub cluster

    ≥80%

    Green power ratio for cluster data centers

    • New data centers strictly held below PUE 1.2
    • Phase out legacy sites with annual PUE above 1.5
    • Direct green supply and grid-forming storage to cut delivered green power prices

    2024Opinions on supporting green computing in the Horinger cluster

    Policy source
  • Ningxia

    Zhongwei cluster

    Hub cluster

    RMB 10M

    Max award per company for new sites below PUE 1.2

    • Investment over RMB 100 million and measured annual PUE below 1.2: up to RMB 100 per m²
    • Using in-cluster computing services: 30% of contract value, up to RMB 200,000
    • First reaching PUE 1.2: up to RMB 300 per tonne of coal equivalent saved

    2022Policy opinions on building the Ningxia national computing hub

    Policy source
  • Gansu

    Qingyang cluster

    Hub cluster

    ≤RMB 0.4

    Delivered power price per kWh

    • Gigawatt-scale aggregated green supply for data centers; first 1 GW connected to the grid
    • About 55% of park consumption supplied by aggregated green power
    • A key pillar of the national zero-carbon park build-out

    2026Qingyang East-Data-West-Computing park green power aggregation pilot

    Policy source
  • Guizhou

    Gui’an cluster

    Hub cluster

    30%

    Incentive on valid contract value for AI computing

    • Up to RMB 2 million per buyer per year
    • RMB 140 million allocated for 2026, open to users nationwide
    • Training corpus purchases: 30%, up to RMB 1 million a year

    2026Guizhou computing voucher measures (effective April 1, 2026)

    Policy source
  • Hebei

    Zhangjiakou cluster

    Hub cluster

    1.19

    Average PUE of city computing facilities

    • 4 computing centers approved for direct green supply, cutting power cost by about RMB 0.1/kWh
    • Renewables make up 88.85% of installed power capacity
    • 53 computing centers in service with 1.07 million standard racks

    2026Zhangjiakou data center cluster plan · direct green supply projects

    Policy source
  • Sichuan

    Clean-energy regions

    Hub cluster

    ≤1.25

    Average PUE of clean-energy clusters by 2027

    • Wind and solar connected nearby through source-grid-load-storage supply
    • 10,000-GPU projects in constrained grid sections exempt from peak pricing on grid purchases
    • Retained and locally dispatched hydropower prioritized for compute-power projects

    2025Implementation opinions on accelerating compute-power integration

    Policy source
  • Guangdong

    Shaoguan cluster

    Hub cluster

    ≤1.25

    Average PUE of new projects in the hub cluster

    • Tiered power prices based on measured operating PUE
    • No new large or hyperscale facilities outside the Shaoguan cluster in principle
    • Targeting 80%+ green power for data centers by end-2026

    2025Opinions on strengthening data center layout

    Policy source
  • Beijing

    Beijing-Tianjin-Hebei

    Province / city

    ≤1.25

    PUE for new and expanded AI data centers

    • AI data centers above 30,000 tce a year: PUE no higher than 1.15
    • Build zero-carbon AI data centers on 100% green power
    • 45,000 PFLOPS of AI computing across the city in 2025

    2024Beijing computing infrastructure plan (2024–2027)

    Policy source
  • Shanghai

    Yangtze River Delta

    Province / city

    >20%

    Green energy share in AI data centers

    • New AI data centers below PUE 1.25; retrofits below 1.4
    • Liquid-cooled racks above 50%
    • AI computing capacity above 30 EFLOPS

    2024“Computing Pujiang” AI computing action plan (2024–2025)

    Policy source
  • Shenzhen

    Greater Bay Area

    Province / city

    RMB 10M

    Max training voucher per entity per year

    • Up to RMB 500 million in training vouchers a year, covering up to 50% of contracts
    • Startups funded at up to 60%
    • New data centers below PUE 1.25 with green rating 4A or higher

    2024Measures to build Shenzhen into an AI pioneer city

    Policy source
  • Hangzhou

    Yangtze River Delta

    Province / city

    RMB 250M

    Annual computing voucher budget

    • Usually 30% of contract value; SMEs with contracts under RMB 500,000 up to 50%
    • Up to RMB 8 million per company per year
    • Pilot smart vouchers stack with computing vouchers up to 60%

    2025Hangzhou plan to build an AI innovation highland (2025 edition)

    Policy source
  • Chengdu

    Chengdu-Chongqing

    Province / city

    RMB 5M

    Max annual deduction per entity

    • Up to RMB 100 million issued each year
    • Self-reliant AI computing: 50%; other AI computing: 30%
    • SMEs receive an extra 10 percentage points

    2026Chengdu computing voucher measures (revised)

    Policy source
  • Hubei

    Province-wide

    Province / city

    25 EFLOPS

    Provincial computing capacity by 2027

    • AI and supercomputing at least 90% of new facilities
    • New data centers investing over RMB 200 million: 10% subsidy, up to RMB 5 million
    • Source-grid-load-storage supply for integrated compute and power

    2025Hubei measures for coordinated computing, networking, storage and use

    Policy source
  • Wuhan

    Optics Valley

    Province / city

    RMB 3M

    Max annual subsidy per entity

    • Up to RMB 100 million in vouchers a year
    • Self-reliant computing: 30%; domestic LLM token services: 20%
    • Current service vouchers cover 50% of fees, up to RMB 200,000

    2026Wuhan computing voucher measures (draft for comment)

    Policy source
  • Map local provisions to your site, line by line

    Green power ratios, PUE thresholds, computing vouchers and power pricing differ by region. MeetCarbon maps the provisions that apply to your site and reviews them alongside the national green computing facility criteria.

    Book an eligibility review

Status quo · Five tests

One data hall must pass five tests at once

The five dimensions sit with different teams and systems. Hard to prove at application, harder to keep after selection.

01

Green power share

80% and provincial quotas

Certificates, green power trades, on-site solar and direct supply use different evidence. Certificate years drift from consumption years, and each report counts differently.

Procurement & trading

Root cause

Data is scattered across five systems

Four data types never align on the same time grain, and the same green power share differs from report to report.

BMS & DCIMPUE, temperature
Bills & metersTotal and IT load
Trading & certificatesGreen power evidence
Cluster schedulerCompute output
Annual spreadsheetsInventory, applications

One energy-carbon-compute-water ledger

Aligned on one time grain · every number traceable

Scattered data

Energy, bills, certificates, carbon and compute live apart

Mixed methods

433, 334, SJ/T 12141, quotas and ESG each calculate differently

Reactive year-end

Slippage surfaces only at annual filing

Energy disconnect

Green power, storage and demand response sit in different teams

Lost know-how

O&M and application methods depend on individuals

Industry chain · Collaboration

Covering every link of the computing value chain

From upstream power and equipment to midstream facilities, downstream users and the government, grid and finance around them. Click any role for pain points and collaboration scope.

Upstream

Midstream

Downstream

MeetCarbon Cloud

Enablers

Solution · Architecture

Four layers, one dataset for every method

Read-only access to existing systems builds energy, carbon, compute and water ledgers on one time grain. Business apps and the MeetCore layer turn that data into reports and actions.

MeetCore layer

Drafts reports, explains gaps, captures playbooks; writes and filings need human sign-off

L4
Application and disclosure draftingIndicator Q&A and gap analysisO&M knowledge base and playbooks

Business apps

One dataset reused across policy methods

L3
Green & zero-carbon computingPUE · CUE · WUECarbon & footprintGreen power & certificatesVPP & power tradingSmart O&MESG & green financeGovernment oversight

Unified ledger

Hourly, daily, monthly and annual alignment; every number traceable

L2
EnergyCarbonComputeWaterEvidence & workpapers

Access layer

Read-only; source systems keep control

L1
BMS & DCIMMeters & sub-meteringBills & settlementsCertificates & green power evidenceScheduler & compute outputSolar, storage & distribution

One dataset, many uses

Hover a row to see which methods the same base data serves

433 green computing334 zero-carbonSJ/T 12141QuotaESG & finance
Total and IT loadPUETotal energyCUE denominatorConsumptionEnergy disclosure
Purchased power splitGreen shareIndicators 1.2 / 1.3Purchased emissions80% checkGreen consumption
Fossil fuelsCUEIndicator 1.1Direct emissions—Scope 1
Water withdrawal & useWUE———Water disclosure
Racks & compute outputUtilization, compute efficiencyProduct footprint——Compute carbon efficiency

SaaS

Single owners and small halls, ready to use

On-premises

Operators, large AI centers and government platforms; data stays in your domain

Hybrid

Campus-side collection and edge computing, consolidated at HQ

Solution · Nine modules

Nine modules, from listing and efficiency to finance

Combine as needed on one shared ledger. Click a card for policy requirements, capabilities and deliverables.

01

Green computing list & zero-carbon facilities

Eligibility checks, self-assessment chapters, annual five-indicator review and zero-carbon core indicators from one dataset

  • Application profileFacility type, sector, recommendation channel, boundary, racks and location
  • Eligibility checkSix criteria judged one by one with a gap list
  • Self-assessment chaptersFigures pulled from the ledger; MeetCore drafts, people finalize
  • Annual reviewPUE, CUE, WUE, green share and utilization tracked monthly with alerts
  • Zero-carbon facilityIndicators 1.1, 1.2 and 1.3, yearly targets, project linkage and acceptance files
433334Annual reviewView details
02

PUE, CUE & WUE operations

PUE, CUE, WUE, compute efficiency and utilization aligned at four time grains, with diagnostics that raise work orders

PUECUEWUEView details
03

Carbon accounting & compute footprint

Organizational inventory and footprint per GPU-hour or per thousand tokens from the same source

InventoryFootprintFactorsView details
04

Green power & certificates

A four-way green power ledger that outputs 433, 334, quota and ESG methods at once

GECsGreen tradesDirect supplyView details
05

Compute-power synergy & flexible load

Compute load, storage and solar aggregated into flexibility for demand response and power markets

VPPDemand responseStorageView details
06

Green power engineering

Distributed solar, grid-forming storage, direct supply works and clean backup power

SolarStorageDirect supplyView details
07

Intelligent O&M & forward deployment

Device records, inspection tickets and the MeetCore knowledge base, running the O&M flywheel in 90 days

Smart O&MKnowledge baseOn-siteView details
08

ESG disclosure & green finance

Green power data in ESG reports; efficiency and carbon evidence for green loans, bonds and REITs

ESGGreen loansREITsView details
09

Government computing & synergy oversight

Regional monitoring of facility efficiency and carbon, dynamic list management and subsidy data verification

MonitoringList managementSubsidy checksView details

Value

Turning policy pressure into measurable business results

Get on the list, hold the indicators, secure green power, cut power costs and connect to finance.

01

Get listed

Eligibility checks, gap lists and self-assessment from one dataset

02

Stay listed

Five indicators tracked monthly with early alerts and work orders

03

Build zero-carbon

Yearly targets for core indicators linked to projects

04

Count carbon right

Inventory, CUE, compute carbon efficiency and footprints from one source

05

Secure green power

Four evidence types in one ledger, checked against 80% and quotas

06

Cut power costs

Storage and flexible load in demand response and power markets

07

Keep know-how

Troubleshooting and application methods as reusable playbooks

08

Connect to finance

Evidence for green loans, bonds and REITs on export

Value by role

Core need

Get on and stay on the national green computing list, build zero-carbon facilities, lower power costs and raise green share

Entry point

433 eligibility review and application pack

Solution mix

Listing, efficiency, carbon and green power first; synergy and engineering next

Expected outcomes

  • Complete application this year
  • Five review indicators visible monthly
  • Green power pathway and solar-storage plan ready for decision

Before / After

Before

  • Five teams each own a slice; data moves in spreadsheets
  • Year-end filing reveals slippage too late
  • The same indicator differs across reports
  • O&M and applications rely on individuals

After

  • One ledger on one time grain, five indicators every month
  • Automatic alerts and work orders on deviation
  • One dataset, many uses, with method differences labelled
  • Know-how captured as playbooks; MeetCore drafts the reports

Roadmap

Start with a 90-day application pack, reach zero-carbon computing in a year

Build the data foundation first, then green power and flexibility, then zero-carbon facilities and engineering.

  1. 1
  2. 2
  3. 3
  4. 4

Capabilities & advantages

Why MeetCarbon Cloud

From policy interpretation to data, reports, engineering and operations, one team delivers the full chain.

One ledger end to end

Energy, carbon, compute and water on one time grain; connect once, serve every app

Traceable methods

Every number traces to source, algorithm, factor version and owner, with methods shown side by side

Policy-to-report loop

Eligibility checks, gap lists, self-assessment and progress reports from the same data

MeetCore role agents

Drafting, gap analysis and playbooks, with human sign-off on key actions

Platform plus engineering

MeetCarbon delivers the platform; our partner builds solar, storage and direct supply

On-premises, data stays home

On-premises and hybrid options, read-only access, no control commands to production networks

MeetCore · The AI work hub that knows you better than you do

Captures computing O&M know-how, application methods and daily routines as reusable memory and playbooks.

Discover MeetCore

Engagement models

Combine as needed, in step with your business

Platform, services, engineering and revenue sharing, signed separately or combined.

Platform subscription

Unified ledger and selected apps, SaaS or on-premises license

By facility or rack scale, annually

Application & assessment services

433 application support, 334 plans and progress, inventory and footprint support

Per project

MeetCore & on-site teams

MeetCore seats, forward-deployed engineers and the 90-day O&M flywheel

Per seat per year / per quarter

Engineering & energy services

Solar, storage, distribution and retrofits via EPC or investment

By EPC contract or energy service agreement

VPP revenue sharing

Demand response and ancillary service revenue shared as agreed

By settled revenue

Government platform

Regional monitoring platform build and annual operations

Per project and annual operations

Pricing follows the formal quotation and contract.

Ways to work with us

1Direct contractFacility owners, operators and authorities
2Joint deliveryGreen data hall packages with server and cooling vendors
3Channel partnersConsultancies, design institutes and integrators by region
4EcosystemRenewable developers, power retailers and VPP operators
5White labelOperators and local platform companies under their own brand

Service boundaries

  • Does not replace BMS, DCIM, BMC or cluster schedulers
  • Does not replace official review or guarantee selection or acceptance
  • Trading filings and settlements are executed after customer confirmation, never automatically
  • Does not issue third-party verification reports or certificates

Next step

Start with one eligibility review

Share your entity, scale, PUE and green power data, and receive an initial assessment and gap list within a week.

  1. 1Needs discussion
  2. 2Quick eligibility review
  3. 3Proposal and quote
  4. 4Kick off the 90-day plan