Short answer: Yes, it is now a buyer's market, and the evidence is structural rather than seasonal. China accounted for 35% of global manufacturing value added in 2025, up from 9% in 2000 (UNIDO), and installed 295,000 industrial robots in 2024 alone, 54% of the world total (IFR). For an overseas buyer the effect is straightforward: more capacity competing for your order, lower MOQs, shorter lead times, and more room to negotiate. The risk does not disappear. It moves. When supply was tight, your problem was finding anyone at all. When supply is abundant, your problem is telling a real manufacturer from a new reseller with a good logo, and telling a quoted price from a price that will hold at production stage. The discipline of buying from China shifts from sourcing to sorting.
For most of the last two decades, importing from China was a search problem. You had a product, you needed a factory that could make it, and the factory had more orders than it needed. Price was largely set by the factory, because the factory held the option of walking away.
That has changed. Capacity has been added faster than most buyers added demand, and Chinese manufacturers and sellers have pushed outward into overseas markets at the same time. If you buy from China today the constraints are almost inverted. Supply is not scarce. Attention and trust are.
This matters because the habits that worked in a seller's market now cost you money.
What actually changed, in numbers
Capacity grew faster than demand
UNIDO's Industrial Development Report 2026 puts China's share of global manufacturing value added at 35% in 2025, against 9% in 2000. Over the same period the combined share of upper-middle-income and high-income industrializing economies doubled from 20% to 44%, and almost all of that gain came from East Asia and the Pacific.
UNIDO's quarterly indicator data tell the same story more recently. In Q4 2025 China's manufacturing exports grew 4.2% year on year, and China's normalised manufacturing trade surplus was 26.8% of its total manufactured trade, the highest of any economy in the table. Fast export growth alongside a wide surplus is what a surplus capacity picture looks like from the outside.
The capacity is real, and it is increasingly automated
Automation is why this capacity did not come with matching wage and price inflation. According to IFR's World Robotics 2025 data, China installed 295,000 industrial robots in 2024, 54% of all new deployments worldwide, and its operational stock passed 2 million units. For the first time, Chinese robot makers outsold foreign suppliers at home, reaching 57% domestic share, up from about 28% a decade earlier.
China's own national statistics point the same way. In the first three quarters of 2026, China's manufacturing value added grew 5.6% year on year, equipment manufacturing 10.2%, high-tech manufacturing 14.3%. Output of industrial robots made in China grew 29.1%, lithium-ion batteries 39.5%, 3D printing equipment 62.5%.
None of this means every category has surplus capacity. It means the aggregate has, and that aggregate slack shows up as negotiating room in the categories closest to it.
Sellers are going out, not just factories
Demand is following supply outward, which is what creates the buyer market rather than merely a competitive one. TikTok and the Mexican Online Sales Association (AMVO) reported that between the February 2025 launch of TikTok Shop in Mexico and September 2026, daily GMV grew 133 times, the catalogue expanded 174 times, active sellers grew 59 times, and the contracted creator network grew 42 times. Those are multiples off a near-zero base and should be read that way. The structural point survives the discount:
What a buyer's market looks like in commercial terms
| Dimension | Seller's market (2015-2022, typical) | Buyer's market (2026, typical) |
|---|---|---|
| Who sets price | Factory quotes, limited room to move | Factory quotes, wide room to move; expect 10-20% negotiated on standard items |
| MOQ on standard items | Often 1,000-5,000 pcs | Frequently 300-500 pcs, and 100 pcs for simple SKUs |
| Lead time quoted | 45-60 days, firm | 25-40 days, with room to pull in for stocked items |
| Number of viable quotes for one RFQ | 2-4 | 8-15 |
| Main failure mode | Cannot find capacity at all | Wrong counterparty, or a quote that collapses at production |
| Where value comes from | Finding anyone | Sorting, and holding a supplier to the quoted terms |
The table above is an average pattern, not a promise. The categories where the shift is furthest along are standard, machine-produced items with many qualified factories. The categories where it has barely started are ones with high setup cost, proprietary process, or certification barriers. Knowing which side of that line your product sits on decides how much negotiating room you actually have.
Why more supply does not automatically mean lower risk
The entry wave is mostly new, not established
A factory that has run the same process for fifteen years and a company that registered a trading business last year both quote you. Both send samples. Both have a website, and increasingly both have a factory video, because the video is cheap and the visitor does not know what it is looking at.
In a buyer's market the ratio of new entrants to proven operators rises. When a factory was busy and rationing orders, its reputation was doing some of the screening for you. When twenty suppliers answer your RFQ, that screening is gone and every one of them needs checking.
The lowest quote is usually a different business model
In our experience the quote that comes back 30% under the others is rarely a more efficient factory. It is usually one of four different things:
None of these are dishonest. They are normal market structures. They are just not what you assumed you were buying.
| What the low quote actually is | How it shows up | The check that exposes it |
|---|---|---|
| Trading company with a factory relationship | Price does not move when you push back; the sub-contracting question is answered vaguely | Ask which processes run in-house and where; ask for the production entity's own business licence |
| Machine owner, not a process owner | Quote looks complete; sample quality drops relative to the price | Ask for a second sample built from the production route, not the showroom route |
| Cheaper component substituted | Spec sheet lists a grade or finish that does not match your drawing | Compare the quoted material against your drawing line by line before ordering samples |
| A quote that assumes you will not check | Full balance falls due after production is complete and packed | Move part of the balance to after inspection; a real manufacturer accepts, a quote-only seller resists |
The new entrants are not all weak
This needs saying clearly, because the buyer's market creates a bias in the other direction. Many of the fastest-growing Chinese exporters of the last three years are young companies with real factories, real equipment, and genuinely better prices. The problem for a buyer is not deciding that all new suppliers are suspect. It is deciding which of them are factories, and doing it with evidence rather than impression.
If you are screening candidates, the useful reference is a step-by-step method for verifying a Chinese supplier is a real factory rather than a middleman, which is designed for exactly this situation.
Distance between capacity and capability
Capacity statistics describe how much a country can produce. They say nothing about whether a given factory can produce your part to your tolerance on your delivery date. A plant with excellent equipment and no in-house quality function will still ship a batch with a cosmetic defect rate nobody measured.
This is where two IFR metrics diverge usefully. Robot density per 10,000 manufacturing employees in 2024 was 267 in Western Europe, 231 for the EU as a whole, and 204 in North America. China was at 166, up 17% year on year, ranked 22nd worldwide despite having the largest operational stock in the world. China's automation depth is still catching up to its automation scale. Large domestic manufacturers are heavily automated;
How to actually shop a buyer's market
Stop treating the RFQ as a request for one quote
In a seller market you sent one RFQ, got two answers, and moved on. In a buyer's market that is leaving most of your leverage on the table. The single most useful change is process, not price negotiation: run every RFQ as a structured comparison rather than a request.
Practically, that means:
• Send the same specification to every candidate, with the same tolerances and the same delivery terms. Differences in what vendors choose to quote against are themselves data.
• Ask for MOQ at each of three quantities, not just one. A supplier whose MOQ does not move between quantities is telling you something about how flexible the operation really is.
• Ask what changes if the quantity doubles. If the answer is "nothing", the first price was padded. If the answer is a specific line-item breakdown, you are talking to the person who controls the cost.
• Ask which part of the quoted scope is sub-contracted. Direct suppliers answer plainly. Traders need a beat.
This turns a single relationship decision into a comparison. You stop evaluating one vendor's persuasion and start evaluating eight vendors' consistency.
Read the price the way you would read a spec sheet
In a buyer's market, price differences between credible quotes are usually smaller than buyers expect, and the differences that matter are in the exclusions. What to check every time:
• Is tooling included, and is it amortised or one-off? A tooling charge hidden in a unit price never comes down.
• Is the quote EXW, FOB, or DDP? Two quotes differing by 25% often differ by nothing except delivery scope. Compare like with like or you will chase a phantom bargain.
• Is tax invoicing included? A factory that cannot issue the invoice you need costs you more than its unit price gap suggests, at your end.
• What is the payment structure? Deposit, balance before shipment, and balance after inspection mean very different things for your leverage.
• What is the lead time tied to? Tooling, sample approval, raw material availability, or a queue. A 30-day promise tied to an unconfirmed raw material position is a 45-day promise.
The questions map to a short list you can attach to any quote:
| Check | What a clean answer looks like | Red flag |
|---|---|---|
| Trading terms | Same basis across all vendors, stated explicitly | Mixed EXW, FOB and DDP quotes compared directly |
| Tooling | Stated as a one-off, with amortisation shown separately if amortised | Tooling folded into unit price, disappears at higher volume |
| Tax invoicing | Vendor states it can issue the invoice type you need | "No problem" with no confirmation, or a special-company invoice |
| Payment structure | Deposit, balance before shipment, balance after inspection | Balance due on production completion, before you have seen goods |
| Sub-contracting | Named openly, with which steps and why | "Everything in-house" plus reluctance to name the production entity |
| Lead time basis | Tied to named milestones you can verify | Single number with no breakdown |
Understanding whether a quote is a transparent original factory quote or a price with layers added is the difference between a real comparison and a comparison of markups.
Verify before you negotiate, not after
Verification is cheaper before the order. Once a deposit is paid, your options narrow sharply and your counterpart's options do not.
The checks that carry the most weight are the physical ones. Walking the floor tells you what no document will: whether the machines in the video are on the floor, whether the workers are on the line or standing idle, whether the goods in the warehouse are the same goods, whether QC exists as a function or as one person with a clipboard. If you cannot travel, that is what an on-site visit by someone who works in your category is for.
Keep your option to switch
The most expensive thing a buyer can do in a buyer's market is decide too early, because the price you agreed to in a rush is the price you will pay for months.
What we do differently when the market is this soft
When the market turns soft, most buyers respond by pressing harder on price. That is the wrong first move, because the soft market is also the moment when the marginal suppliers are most willing to take work they should decline, and when a quote that is too good is most likely to be structurally different from what you will receive.
What changed in our own process, after watching several buyers get burned by exactly this pattern:
We stop quoting from a supplier pool and start quoting from a verified one. Our standard rate is per factory, paid on the outcome, and the thing we are paid for is a factory that goes into your final list, not a search we ran. If every factory we bring you is rejected, that category is not charged. That structure exists so that screening quality is worth more to us than volume of names, and it changes what our people do on site.
We ask the disqualifying questions first. On a first call for a new buyer we ask about annual volume, target market, and certification requirements before asking for a quote. Certifications and which compliance obligations attach to your product and market can remove a third of the candidates before anyone has visited anyone. Better to know in week one than in month three.
We hold back the volume information deliberately. In a soft market, quoting your full three-year forecast to eight vendors invites eight vendors to price for a customer that does not exist yet. We ask suppliers for a price at your actual first order quantity, and we keep the forecast internal.
We check the invoice and the licence against the operation. A business licence covering a broad trading scope tells you nothing. What matters is whether the entity you are contracting with is the entity that will appear on the production documentation and the export declaration, because a mismatch there creates customs problems at the shipping stage that no amount of good factory work will fix.
The pattern across all four: we spend more time before an order and less time negotiating during one. That is the correct allocation of effort when supply is abundant and counterparty quality is the binding constraint.
The short version
A buyer's market is a gift that arrives with a bill attached. You get better prices, smaller MOQs, faster lead times, and more suppliers willing to talk. You give up the automatic screening that a factory's willingness to be booked was once doing for you.
If you take one action from this, make it the comparison structure rather than the price. Send one identical specification to eight qualified factories, ask each the same three quantity breaks and the same sub-contracting question, and look at which answers are consistent. In our experience the consistency of the answers predicts the consistency of the goods more reliably than price does.
Global manufacturing capacity is not going to shrink because buyers learned to negotiate. Buyers who build a real sorting capability now will keep it; the ones who treat this cycle as a temporary discount will be back to paying the scarcity price within a few years with none of the leverage.
Footnotes
- UNIDO, *Industrial Development Report 2026*, Figure 1.5 and Figure 1.6. https://www.unido.org/sites/default/files/unido-publications/2025-11/UNIDO%20IDR26.pdf
- UNIDO, *World Manufacturing Production and Trade Statistics, Q4 2025*, Tables A.1, A.2 and A.3. https://stat.unido.org/portal/storage/file/publications/qiip/World_Manufacturing_Production_2025_Q4.pdf
- International Federation of Robotics, *World Robotics 2025 - industrial robot installations*, press release, 25 September 2025. https://ifr.org/ifr-press-releases/news/global-robot-demand-in-factories-doubles-over-10-years
- International Federation of Robotics, *World Robotics 2025 - robot density*, press release, 8 April 2026. https://ifr.org/ifr-press-releases/news/robot-density-surges-in-europe-asia-and-americas
- National Bureau of Statistics of China, first three quarters 2026 economic performance release, 19 October 2026. https://www.stats.gov.cn/sj/zxfbhjd/202610/t20261019_1965647.html
- State Council of the People's Republic of China, *First three quarters 2026: economy holds steady, high-quality development makes progress*, 19 October 2026. https://www.gov.cn/lianbo/202610/content_7083647.htm
- Figures attributed to TikTok and the Mexican Online Sales Association (AMVO) in the seller-entry section are as reported in their joint Mexico market study covering February 2025 to September 2026, and are presented as multiples from a near-zero base.
