300 Standards by 2026: What “Meets GB Standards” Does and Doesn’t Prove to a CNC Buyer
- Executive Summary
- 1 · Background: a standards plan, not a certificate
- 2 · The six layers — and the one buyers should read
- 3 · Three test codes worth knowing by name
- 4 · Five things a standards claim does not prove
- 5 · Key data and credibility ratings
- 6 · Why this matters to buyers: seven asks
- 7 · Honest limits
- 8 · Source & credibility note
- FAQ
China’s national standardiser and its industry ministry have put a number on machine-tool standardisation: at least 300 standards drafted or revised by 2026, a 90% international conversion rate, and a full system by 2030. None of that certifies any individual machine — but it does change what you can legitimately ask a supplier to put in writing.
- The shift: Chinese policy has moved from funding and projects toward defining measurement. The layer that matters most to a buyer is 检验检测 — inspection and testing — because it governs how an accuracy number was produced, not just what it claims.
- What it enables: as Chinese makers increasingly cite GB/T and ISO test codes, the useful question moves from “does it meet a standard?” to “which standard, which edition, measured by whom, under what conditions, on which serial number?”
- What it does not do: a standards work programme is not a conformity mark, not CE or UKCA, and not a ranking. “300 standards drafted” is an input; “90% conversion” means adopting ISO methods, not indigenous development. A machine can pass a test code on acceptance day and still drift in service.
1 · Background: a standards plan, not a certificate
The Industrial Machine Tool High-Quality Standards System Development Plan (《工业母机高质量标准体系建设方案》) was issued jointly by China’s Standardization Administration and the Ministry of Industry and Information Technology. The signed document carries a date of 1 July 2025; it was publicised in early September 2025 and circulated to provincial authorities, and was also republished as a trade circular outside mainland China.
It sets two horizons. By 2026: the standards system is to be “basically established”, with no fewer than 300 standards drafted or revised, China leading the drafting or revision of no fewer than five related international standards, and an international standard conversion rate of 90%. By 2030: the system is to be “fully formed”, with subtractive and equal-material manufacturing standards reaching world-advanced level overall, and some additive-manufacturing standards reaching world-leading level.
Read plainly, this is a work programme: a commitment to write, revise and align specifications. It is not a product certification scheme, it assigns no mark, and it creates no independent body that signs off an individual machine. That distinction is the whole point of this briefing — because the phrase “meets GB standards” is starting to appear more often in export conversations, and buyers deserve to know exactly how much weight it carries.
2 · The six layers — and the one buyers should read
The plan organises machine-tool standardisation into six layers covering 34 major categories and 17 sub-categories, adjusted dynamically as the industry develops:
| Layer | What it covers | Relevance to an overseas buyer |
|---|---|---|
| 基础通用 Basic & general | Terminology, classification, safety fundamentals — described in the plan as the base of the system | Low direct use, but it is what makes Chinese and English spec sheets comparable |
| 设计配套 Design & supporting | Design rules and the supporting components that go into a machine | Useful when auditing which subsystems a builder makes versus buys |
| 制造装备 Manufacturing equipment | The machine itself — the plan calls this the core of the system, reflecting overall technical level | Medium: sets the frame for machine-level specs |
| 制造工艺 Manufacturing process | How parts are formed; the plan ties this to product quality and cost control | High if you care about process repeatability across batches |
| 检验检测 Inspection & testing | Explicitly “the basis for evaluating product quality” | Highest. This is the layer you can put into an acceptance clause |
| 行业应用 Industry application | Applying machine tools to other industries | Medium: emerging application standards (EV, die casting, humanoid robotics) |
The plan also names four priority tasks, two of which matter commercially. The first is accelerating standards where they are urgently needed — high-end compound machine tools, integrated die casting, precision forging, hybrid additive manufacturing, high-end injection moulding, high-grade CNC systems and high-performance functional components. The second is standards for cross-cutting technologies — big data, cloud computing, artificial intelligence, IoT and digital twins as they apply to machine tools.
There is one further line worth quoting, because it sets expectations honestly: the plan calls for some key standards to run moderately ahead of the industry’s average level. Standards here are intended to pull the industry forward, not merely to describe where it already is. That is a sensible policy design — and it is also a reminder that a standard can exist before the production capability behind it is common.
3 · Three test codes worth knowing by name
You do not need to read 300 documents. For buying decisions, three test codes do most of the work — and all three exist in both Chinese and international form, which is precisely what the 90% conversion target is about.
GB/T 17421.2 — positioning accuracy and repeatability (the Chinese adoption of ISO 230-2)
This is the standard test code for determining the positioning accuracy and repeatability of numerically controlled axes. It specifies the method: how to measure, in which directions, how many target positions, how to process the data. It does not specify an acceptable value. Two machines can both be “tested to” this standard and land in very different places — which is why the report matters and the label does not.
GB/T 39967 — the S-shaped test piece for five-axis machines
Five-axis capability shows up in the S-shaped test piece, not in a rotated cube. A machined S-specimen with measured form error tells you something real about simultaneous five-axis motion, servo matching and the controller’s behaviour through direction changes. If you are buying five-axis, ask for the specimen and the measurement, not a demonstration video.
Ballbar and circular tests — the quick diagnostic
A telescoping ballbar test is fast, cheap and repeatable, and it is the standard way to spot backlash, servo mismatch and squareness problems. It will not replace a laser measurement for acceptance, but it is excellent as a periodic health check — and therefore excellent as a re-test clause at a defined running hour.
4 · Five things a standards claim does not prove
- It is not a conformity certificate. The plan is a standards work programme. There is no mark, no third-party sign-off and no certificate attached to an individual machine. “Tested to GB/T 17421.2” means a measurement was performed a particular way — it does not mean anyone certified the outcome.
- It is not market access. CE (EU), UKCA (UK), UL/NFPA (North America) and local electrical and safety rules are separate regimes with their own technical file, guarding, EMC and documentation duties. A machine built for the domestic market is not automatically compliant at your door, and Chinese domestic standards do not substitute for any of it.
- “300 standards drafted or revised” is an input, not an outcome. The count measures standards work completed, not machines meeting them. It says nothing about the specific unit you are buying.
- A 90% conversion rate is adoption, not invention. It means China has taken the great majority of relevant international standards into its own GB/T system, so test methods are broadly aligned with ISO. That is genuinely useful for comparability. It is not evidence of indigenous technology development, and it should not be read as such in either direction.
- Passing a test code sets a floor, not a ranking. Test codes define method. Two suppliers can both pass and differ substantially in what the parts look like after three years of production use.
There is a sixth point, and it is the one most often missed: acceptance-day accuracy is not in-service accuracy. A machine can measure beautifully in a controlled acceptance test and still drift once it has thousands of running hours, thermal cycles and real cutting loads behind it. Chinese users and official-style media have described accuracy decline after extended running as a prominent problem; the practical response is not to distrust the acceptance number, but to contract for a second measurement later.
5 · Key data and credibility ratings
| Item | Figure | What it means for a buyer | Credibility |
|---|---|---|---|
| Document issued by | Standardization Administration + MIIT (国标委联〔2025〕35号); signed 1 July 2025, publicised early Sept 2025 | A policy work programme, not a certification | High (primary document, officially republished) |
| 2026 target — standards | No fewer than 300 drafted or revised | More test methods exist; ask which apply to your machine | High |
| 2026 target — international | Lead ≥5 international standards; 90% conversion rate | Chinese and ISO test methods broadly align — good for comparability | High |
| System structure | 6 layers, 34 major categories, 17 sub-categories | 检验检测 (inspection & testing) is the buyer-relevant layer | High |
| 2030 target | Subtractive & equal-material standards at world-advanced level overall | A direction of travel, not a present capability claim | High |
| Priority directions | Compound machine tools, integrated die casting, precision forging, hybrid additive, high-grade CNC systems, high-performance functional components | Where Chinese suppliers will have the most standard-backed material | High |
| Design intent | “Some key standards moderately ahead of industry average” | A standard can exist before capability behind it is common | High (stated in plan) |
| Industry commentary framing | Standards will compress the space for inflated peak-parameter claims | Ask for sustained-load figures, not only best-case numbers | Low–Medium (industry commentary, not audited) |
One comment from a named domestic manufacturer is worth reproducing because it shows how suppliers themselves read the plan. Qinchuan Machine Tool & Tool Group told a Chinese financial newspaper that it would focus its standards work on its metal-cutting core business and strengthen high-level standard supply across the six key links. That is a reasonable posture — and it is also a fair thing to ask your own supplier about: which layers are you actually working in?
6 · Why this matters to buyers: seven asks
Standards become useful the moment they are written into a contract. These seven work on any supplier, from any country, and none of them requires you to argue about quality — only about evidence.
- Name the standard, the edition and the clause. “Meets GB/T” is not a specification. “Positioning accuracy measured per GB/T 17421.2-2023 (≡ ISO 230-2), bidirectional, three runs” is.
- Ask for the full test report, not the summary. Instrument used (laser interferometer, ballbar), direction of approach, number of target positions, ambient temperature, warm-up procedure, payload on the table, and whether error compensation was active during measurement.
- Ask who measured — and on which serial number. In-house or accredited third party? Laboratory accreditation number? Test date? Critically: was the report produced on your machine, or on a demonstration unit?
- Define acceptance by a run-off on your parts. Cycle time, first-pass yield and Cpk on your workpiece, with measured form error. For five-axis, include an S-shaped test piece with the measured result.
- Add a re-test clause at a defined running hour. For example, a ballbar and axis re-measurement at 2,000 hours, with an agreed tolerance and a written statement of who pays for correction. This is the single most effective counter to accuracy decline in service.
- Settle the conformity file before contract. Which standards your CE or UKCA technical file maps to, who prepares it, who signs the declaration, and who pays if a unit fails local inspection.
- If you need data interfaces, specify them now. OPC UA or MTConnect support, what data is exposed, and who owns the firmware and licence on the controller — retrofitting these later is far more expensive than specifying them at order.
The commercial logic is straightforward. A supplier who has invested in standards work will usually have the documents to prove it, and will usually be willing to attach them. A supplier whose standards language is decorative will struggle with item two. The questions do the sorting for you.
7 · Honest limits
- This page is not a certification opinion. We are a distributor, not a notified body or an accredited laboratory. Nothing here certifies a machine; everything here is a question you can ask.
- Standards alignment is not capability parity. The 90% conversion target improves comparability of measurement. It does not close gaps in long-term accuracy retention, multi-batch consistency, or accumulated process know-how — those depend on time and production volume, not on documents.
- A 2030 target is not a 2026 fact. “World-advanced level overall” by 2030 is an objective stated in a policy document. Treat it as direction, not as present status.
- There are cases where you should buy elsewhere. Certified aerospace or medical qualification histories; tolerances that depend on a process database the builder has not yet accumulated; aero-engine titanium and nickel superalloy rotating parts; or any site where a two-hour service radius outweighs the price difference. Standards alignment does not change any of those.
- Figures here are policy figures. They describe government targets and document structure, not measured machine performance. See the credibility note below.
8 · Source & credibility note
Sources: 《工业母机高质量标准体系建设方案》(Industrial Machine Tool High-Quality Standards System Development Plan), issued by the Standardization Administration of China and the Ministry of Industry and Information Technology, document dated 1 July 2025, publicised 1–2 September 2025 — republished by a municipal industry and information technology bureau: gxj.qingdao.gov.cn; reported by Xinhua Finance (China Financial Information Network), 2 September 2025: cnfin.com; and by Economic Reference Daily, 2 September 2025: jjckb.cn. Also circulated as a trade circular: tid.gov.hk. Named-company response from Qinchuan Machine Tool & Tool Group via Securities Daily (证券日报), September 2025.
Credibility: High for the policy figures, Low–Medium for the interpretive framing. The targets, structure, dates and issuing bodies come from the primary document as republished by official channels and reported by Chinese state financial media — these are verifiable and stable. The “peak parameter versus sustained-load parameter” observation comes from Chinese industry commentary rather than from the plan itself, and is flagged as such. No figure on this page is independently audited by us, and no machine-specific performance claim is made. Verify any standard number, edition and test report directly with the supplier before signing.
Gree CNC Insights publishes translations and buyer-oriented analysis of Chinese machine-tool policy and commentary for overseas buyers. We do not adopt a source’s framing as our own, and we do not rate suppliers on the basis of commentary alone.
FAQ
Does meeting GB/T standards mean a Chinese machine tool is certified?
No. GB/T standards are specifications and test methods, not a conformity mark. Meeting one means a machine was measured a particular way, not that an independent body certified it. Ask for the actual test report, the standard edition, and who performed the measurement.
What is GB/T 17421.2 and why does it matter?
GB/T 17421.2 is the Chinese adoption of ISO 230-2, the standard test code for determining positioning accuracy and repeatability of numerically controlled machine axes. It defines the measurement method, not an acceptable value. Two machines can both be “tested to” it and produce very different results, so always ask for the measured figures and the test conditions.
Does China’s 90% international standards conversion rate mean Chinese machines match ISO?
It means China has adopted the great majority of relevant international standards into its own GB/T system, so test methods are broadly aligned. It says nothing about whether any individual machine passes them, and it is not evidence of indigenous development.
Is a standard-compliant machine automatically CE or UKCA compliant?
No. CE and UKCA are market-access regimes covering safety, guarding, electrical and EMC requirements, with their own technical file and declaration responsibilities. Chinese domestic standards do not substitute for them. Decide in writing who owns the conformity file before contract.
What should I ask for instead of a catalogue accuracy figure?
A full test report: laser interferometer measurement, bidirectional, per axis, with ambient temperature, warm-up procedure, payload and whether error compensation was active. Plus a run-off on your own part class with measured form error and cycle time, and a re-test clause at a defined running hour.
When should I still buy German or Japanese instead?
When your work depends on a certified aerospace or medical process history, on tolerances that rely on a process database the builder has not accumulated, or when a two-hour service radius matters more than the price difference. Standards alignment does not close those gaps.