Global Matcha & Tencha Market Report, Size & Forecast 2026–2033

Tencha, matcha, and the number everyone quotes

Matcha is not a plant. It is a process. Tencha is the leaf — Camellia sinensis grown under shade for the last few weeks before the first-flush pick, then steamed, dried and de-stemmed until only the leaf tissue remains. Matcha is that leaf milled to powder. The market this report measures is the leaf-to-powder chain at first sale: tencha at the auction or the mill gate, matcha at the mill gate. It is not the latte that ends up carrying it, and the distinction matters because two very different markets are routinely quoted under the same word.

Begin with the figure in circulation. Tencha prices rose 170 per cent. That number does not reproduce from the series it is drawn from.

¥8,235 against ¥4,862 is 69.4 per cent. ¥14,333 against ¥5,500 is 160.6 per cent. A 170 per cent increase requires a prior-year base of about ¥3,050, and no published series carries one. The honest sentence is a range: between roughly 69 and 161 per cent depending on grade and series, with Uji hand-picked leaf at 116.4 per cent and the national average at ¥8,562. Quoting the flat figure is not wrong in direction. It is wrong in precision, and in a market where a buyer’s entire margin sits in a two per cent swing it is the precision that gets used.

The same auction then moved again in 2026. The Uji market in Kyoto ran from 24 April to 31 July and turned over ¥12.94 billion against ¥9.629 billion the year before — up 34.4 per cent — on 595 tonnes of tencha, the largest volume since the market opened in 1974. Machine-harvested tencha alone generated ¥10.101 billion at an average of ¥17,250 a kilogram, the highest since records began in 1996. Divide one into the other: ¥10.101 billion at ¥17,250 is 585.6 tonnes, or 98.4 per cent of the volume traded. Hand-picked leaf is 9.4 tonnes — 1.6 per cent.

That division is the most useful calculation in this report. The most expensive tea in Japan, the tier every brand markets against, is 1.6 per cent of the volume traded at its own auction. Scarcity in the ceremonial tier is not a marketing claim. It is arithmetic, and it is why a supply response aimed at the premium end cannot scale: there is almost no volume there to scale.

The two official series cannot simply be added together

Japan reports the leaf. China reports the powder. MAFF puts 2025 tencha output at 6,278 tonnes, up 18 per cent. The China Tea Marketing Association puts world matcha output at 17,600 tonnes, up from 8,266 tonnes in 2020, of which China produced more than 12,000 — close to 70 per cent. Add the published figures as they stand — 6,278 plus 12,320 — and you get 18,598 tonnes against a world total of 17,600: 5.7 per cent more than the whole.

Something overlaps, and no official document resolves it. This report reconciles by taking the world total as the pool and the Japanese crop as leaf input, applying a 95 per cent milling yield. That yields 5,964 tonnes of Japanese matcha equivalent, 33.9 per cent of the pool, and 11,636 tonnes of Chinese-origin material, 66.1 per cent. The 95 per cent is a Pheonix assumption; everything else on the line is published.

Why Japanese matcha is worth 2.5 times Chinese matcha

Japan realises USD 57.08 a kilogram at the farm gate. Chinese matcha leaves the plant at an average of USD 22.79. The ratio is 2.51. Read it carefully in both directions, because it is the most misused comparison in the category.

The Chinese figures are output values for finished milled powder, so they already carry a milling and finishing margin. The Japanese figure is a farm-gate price for unmilled leaf and carries none. On a like-for-like basis the gap is wider than 2.51 times, not narrower — the opposite of what most origin comparisons assume. The Japanese premium is not a milling premium. It is payment for a cultivar and shading regime, for a grade structure buyers can specify against, and for the fact that Japanese supply cannot expand.

The export figure that cannot be a matcha price

Japan exported 8,718 tonnes of powdered green tea in 2025, up 71 per cent, worth 84 per cent of a ¥72.1 billion green tea export book — about ¥60.6 billion, or USD 404 million at ¥150. Divide: ¥6,947 a kilogram, USD 46.32. That is 18.9 per cent below the national tencha farm-gate average of ¥8,562. A finished powder cannot trade below the leaf it was milled from.

Two readings survive. Either the customs category carries cheaper powdered product that is not matcha, or Japan is re-exporting and blending imported leaf. Both are consistent with China at two thirds of world output. A third reading — that the tonnage is Japanese tencha — is not: 8,718 tonnes is 38.9 per cent larger than Japan’s entire tencha crop. What follows matters more than the puzzle. Japan’s export statistics cannot be used as a proxy for the Japanese matcha market, and any model that multiplies export tonnage by a matcha price overstates it.

The same export figure sets a floor under the pool. Japan’s declared export value alone is USD 404 million, more than the USD 340 million this report assigns to all Japanese matcha at farm gate. Adding it to the Chinese value gives a trade-level floor of USD 668.9 million against the USD 605.6 million ex-producer headline. This report quotes the conservative of the two and labels which is which.

The forecast, and why value lags volume

The base case is built origin by origin rather than as one line. Japanese leaf grows 3.5 per cent a year to 7,853 tonnes — conversion of sencha ground into tencha continuing, on an area that keeps shrinking — and Japanese price mean-reverts from the auction peak to USD 52.00 a kilogram by 2033. Chinese volume grows 11 per cent a year to 26,815 tonnes, with price rising 3.7 per cent a year to USD 27.68 as grades improve and compliance costs are absorbed. The consequence is in the mix, not the total: Japan’s share of pool value falls from 56.2 to 35.5 per cent and China’s share of tonnage rises from 66.1 to 77.4 per cent. Realised value per kilogram declines 3.5 per cent over eight years even as the pool nearly doubles.

A reader who takes the 2025 price spike as a growth rate will model a market this report does not describe. The spike is Japanese and it is a price event. The growth is Chinese and it is a volume event. Treating the first as the second inflates value and misses the mechanism that decides who earns the margin.

Market segmentation: five cuts, one of them published

Only one cut below is a published statistic — origin. Every other share is a Pheonix modelled allocation, calibrated so that each cut sums to the same 17,600 tonnes and USD 605.6 million. They are independent marginals, not a matrix: multiplying a grade share by an end-use share produces cells this report does not model, and no source publishes the joint distribution.

1. Origin

This is the only cut with a published denominator, and it is the one that carries the forecast. Two thirds of the tonnage generates 44 per cent of the value.

2. Quality tier — 3 per cent of the volume, 22 per cent of the value

The top tier realises 7.33 times the pool average on three per cent of the volume, and the bottom tier realises under a third of the average on a quarter of it. This is the cut that decides brand strategy, and it explains how a business can be a large buyer of matcha and never once touch the material whose price is moving. Note the difference from the auction: hand-picked leaf is 1.6 per cent of the Kyoto auction but the global ceremonial tier is 3 per cent of this pool, because the auction covers Japanese leaf only while the pool includes Chinese whisking-grade output.

3. End use — where the tonne is actually drunk

Fifty-six per cent of tonnage is sold as an input into someone else’s brand — food service and RTD combined. That decides how a price rise propagates: it arrives through a menu or a formulation cost, not through a shelf price, and both lag input costs by a season or more. It is also why the demand destruction a spike causes is invisible in retail scanner data, which is the series most analysts watch.

4. Certification

Organic realises 2.35 times conventional material and 2.13 times the blended pool on 8 per cent of the tonnage. Both shares are modelled, and the constraint behind them is not: certification runs on hectares and conversion time, so this cut is the only one in the report that cannot move in response to its own price.

5. Route to market

Forty-four per cent moves through a party that owns neither the leaf nor the brand. That is where margin is captured without capital, and it is also where provenance claims are most easily lost — which is the commercial reason the GI registration in July 2026 matters more to distributors than to growers.

What the cuts are for

Two of the five change decisions, and they say the same thing from opposite ends. In the quality cut, 3 per cent of tonnes carry 22 per cent of value. In the origin cut, 66 per cent of tonnes carry 44 per cent of value. Weight is not value in this market, and every strategy question — what to plant, what to mill, what to buy, what to put on a label — resolves to which side of that gap a business intends to sit on.

What would have to be true for the base case to be wrong

Three things. Japanese yield. 2026 first-flush crude tea rose 21 per cent to 24,100 tonnes on an area 1 per cent smaller — productivity, not hectares, is now the supply lever, and if that repeats the Japanese volume path is understated. Chinese compliance. The Positive List regime screens more than 800 residues at a uniform 0.01 ppm, and Chinese volume reaches Western brands only as fast as producers can demonstrate that. Western demand elasticity. Matcha in the United States and Europe is increasingly a café ingredient rather than a ritual purchase, and café ingredients get substituted when they get expensive.

Basis of estimate. The world output total, the Japanese leaf crop, the auction results, the farm-gate averages, the export values and the Chinese provincial output values are all published figures, linked where they appear. The 95 per cent milling yield, the 70:30 Chinese price blend, the organic shares and every forward path are Pheonix modelling, stated as modelling. Currency conversion is at ¥150 and RMB7.1 to the dollar throughout; a 10 per cent move in either rate moves the dollar pool by roughly that much and moves no conclusion.

Table of Contents

• Market overview: tencha, matcha, and the number everyone quotes
• Competitive landscape
• Value chain: where the kilo is repriced
• Investment activity
• Technology and innovation
• Market risk
• Regulatory landscape

Competitive Landscape

Competitive landscape

No single company dominates the material. That is not a diplomatic way of saying the market is competitive; it is a structural fact with a specific cause. The largest tea company in Japan is a beverage business, the reference names in the ceremonial tier are small family millers, and the largest single production asset in the world is a Chinese plant that most Western buyers have never heard of.

The auction is the only transparent price in the chain

Everything else in this market is a private negotiation. The Kyoto and Uji auctions are the exception, and that is why the 2026 result matters more than its size: ¥12.94 billion across 595 tonnes, with machine-harvested leaf at ¥17,250 a kilogram — the highest since records began in 1996. When the only public price prints a record, every private contract in the chain reopens. That is the mechanism by which an auction of 595 tonnes — 3.4 per cent of the world pool — sets the price of the other 96.6 per cent.

Concentration on the supply side, fragmentation on the demand side

The grower base has halved twice over. Japan counted 53,687 tea-growing farms in 2000 and 12,325 in 2020, a 77 per cent decline, and 62 per cent of core tea workers were 65 or older in 2020 against 53 per cent in 2005. Fragmentation at the farm gate is now a function of demography rather than of market structure: there are simply fewer sellers every year, and the ones who remain are older.

Against that, capacity in the substitute arrives in units larger than the incumbent’s entire crop. Tongren produced and sold about 2,500 tonnes in 2025 — 39.8 per cent of Japan’s national tencha output — and exported 1,300 tonnes, the largest export volume of any Chinese city, with Guizhou matcha first entering the Japanese market in February 2025. A single Chinese city now ships a volume that would take Japan years to add.

The umbrella is sixty per cent wide

Chinese material at USD 22.79 a kilogram is 39.9 per cent of the Japanese farm-gate price. A Chinese supplier could discount 60 per cent against Japanese leaf and still hold a margin Japanese growers cannot match at any volume, because the Japanese constraint is land and labour rather than price. That is not a competitive threat in the ordinary sense. It is an umbrella under which an industrial competitor can win share indefinitely without ever bidding for the premium tier.

And the umbrella has one pole doing most of the work. The Guitea plant at Tongren, 4,000 tonnes a year of capacity, is 34.4 per cent of the Chinese-origin pool and 22.7 per cent of the world pool at a single address. Tongren’s 2025 output and sales of about 2,500 tonnes already equal 39.8 per cent of Japan’s entire tencha crop. Concentration on the supply side is therefore not a forecast risk. It is the current structure, and it sits entirely on the side of the market that can still expand.

What the buyer’s leverage actually is

In a market where one origin cannot expand and the other can, leverage belongs to whoever can hold two origins in the specification. That is a sourcing decision, not a price negotiation, and it is the same logic that has already reordered supply in high-protein dairy, where ingredient origin became a formulation constraint rather than a purchasing one. A buyer single-sourced in Japan in 2026 is not paying a premium for quality. They are paying a scarcity rent with no substitute lined up.

No announced merger or acquisition was identified in this market for the period covered. Consolidation here is happening through capacity construction rather than through transaction, which is why the competitive map changes slower than the price does.

Value Chain

Value chain: where the kilo is repriced

The chain has five hands: the field, the co-operative or auction, the miller, the brand or blender, and the café or retailer that sells the cup. Four of those five have a published price. The fifth does not, and that absence is why matcha is so often mispriced by buyers.

The ladder, read from the top

Hand-picked Uji leaf trades at 5.06 times the national farm-gate average and 2.01 times the machine-harvested Uji average from the following season. That multiple is not a milling return and not a brand return. It is payment for the one input that cannot be mechanised, on 9.4 tonnes of volume — 1.6 per cent of the auction. Any buyer who treats the hand-picked price as the price of matcha has mistaken the top of a very thin pyramid for its base.

Read from the bottom, the customs figure is the odd one. ¥6,947 a kilogram of finished powder is 18.9 per cent below the average price of the unmilled leaf it would have to be made from. The only readings that survive are that the category contains cheaper powder which is not matcha, or that Japan is blending in imported leaf before re-export. Either way, the export line is a trade statistic, not a price series, and using it to value a matcha tonne is a category error repeated in a great deal of published work.

Where the margin actually sits

Scarcity is priced at the leaf, because that is where the constraint is. Milling and finishing are volume businesses earning a conversion spread — they scale with throughput, not with grade. The brand is where the retail multiple is earned, and this report does not measure it: a finished ceremonial tin reaches the consumer at a multiple of the leaf price that is set by packaging, story and shelf position, not by the auction.

That split has a practical consequence for anyone buying. The leaf price you cannot influence; the grade specification you can. Demanding a named cultivar and a shading standard converts a commodity negotiation into a specification negotiation, and specification is the only lever that survives a seller’s market. The downstream demand that sets the ceiling is visible elsewhere in this research library — in ready-to-drink beverages, where matcha competes as a flavour and a functional claim, and in premium chocolate, where it is a colour and a price point as much as a taste.

What the five rungs earn, priced in dollars

Convert the ladder into dollars per kilogram of finished powder and an anomaly becomes an inference. Japanese leaf at USD 57.08 a kilogram, milled at a 95 per cent yield, costs USD 60.08 per kilogram of powder before a single yen of milling, packing or freight. The customs category — all Japanese powdered green tea exports — realises USD 46.32. That is 77.1 per cent of the input cost: a 22.9 per cent shortfall before any conversion margin at all.

Run the same calculation against Chinese material and the sign flips. Chinese matcha at USD 22.79 milled at 95 per cent is USD 23.99 a kilogram, so the export unit value recovers 193.1 per cent of it — a 93.1 per cent margin that comfortably funds milling, freight and re-export. Only one of those two readings is compatible with a functioning trade. Japan’s declared powdered green tea book is therefore priced far closer to Chinese material than to Japanese leaf, which is independent support for the blending and re-export reading reached earlier from volume alone.

Every thousand tonnes is worth USD 34.3 million. The gap between origins is USD 34.29 a kilogram, so a thousand tonnes moved from Japanese to Chinese supply releases that much pool value at constant tonnage. Across the forecast that is the entire mechanism: the pool adds 17,059 tonnes and China supplies 15,170 of them — 88.9 per cent of the growth, at 40 per cent of the price. Realised value per kilogram falls because of what is being added, not because anyone is discounting.

The asymmetry that decides the forecast

Japan’s cost base is labour and land, both shrinking. China’s is land and capital, and neither is scarce at the volumes involved. That is why the same product trades at USD 57.08 and USD 22.79 a kilogram and why the gap does not close from the Chinese side — a Chinese producer competing on price has no reason to chase a Japanese cost structure, and every reason to hold the spread.

Investment Activity

Investment activity

Capital in this market is not following the price. It is following the land and the labour, and that is why it is flowing to China while the price signal is strongest in Japan.

Where the capacity is being built

The five-year payback is the whole investment case

New tencha fields take roughly five years to pay back. That single number decides the shape of this market for the rest of the decade. A conversion decision taken in 2026 — the best price environment since records began in 1996 — does not add a kilo of supply before 2031. Which means the 2025–26 deficit is not a temporary dislocation that investment will close. It is a scheduled outcome, already committed, and the useful question for an investor is not whether Japanese supply will respond but what gets priced while it cannot.

That window is where the premium sits. Between now and the first meaningful Japanese supply response, the only incremental tonnes available are Chinese, and they arrive at USD 22.79 a kilogram against a Japanese farm gate of USD 57.08. An investor who wants exposure to the price has to own Japanese leaf. An investor who wants exposure to the volume has to own Chinese capacity. The two are not the same trade, and the market quote conflates them.

What a repeat of the 2026 yield is worth

The 2026 season delivered 21 per cent more first-flush crude tea on an area 1 per cent smaller. That is neither a price event nor a demand event; it is a productivity event, and it is worth more than any plausible price move in the same year. Repeat it once: 6,278 tonnes of leaf becomes 7,596, which at a 95 per cent milling yield is 7,217 tonnes of matcha equivalent — 1,253 tonnes above the 2025 pool, or USD 65.1 million at the mean-reverted USD 52.00. That is 10.7 per cent of today’s pool added by agronomy rather than by capital.

Set it against the price sensitivity. A 20 per cent fall in the Japanese realised price — USD 57.08 to USD 45.66 — removes USD 68.1 million, 11.2 per cent of the pool. The two numbers are almost the same size, which means an investor holding Japanese leaf is running a position where a good harvest and a bad price exert roughly equal force. The hedge is not more leaf. It is a grade specification that survives a falling market, which is why the tier cut in the segmentation chapter carries more weight than the volume line in the forecast.

For completeness, the origin split used throughout resolves like this: Japanese leaf compounds at 3.5 per cent from 5,964 tonnes to 7,853 tonnes, which at USD 52.00 is USD 408.4 million — the 35.5 per cent value share the forecast quotes. China absorbs the remainder of the 34,659 tonne pool.

What is genuinely investable

Three assets, none of them a mill. Certified land, because the organic constraint is hectares and conversion time, not capital. Grade specification — the ability to sell against a named cultivar and shading standard rather than a generic “ceremonial” claim, which is the only defensible position in a market where the top tier is 1.6 per cent of volume. The brand licence, where the retail multiple is earned and where provenance protection now has a legal instrument behind it.

Capital intensity is medium rather than low: shading structures, processing lines and certification are real fixed costs, but they are an order of magnitude smaller than the working capital required to hold inventory through a 170 per cent price move. The capital risk in this market is not the plant. It is buying the leaf at the top of an auction cycle and discovering that the marginal Japanese garden can switch back to sencha.

The adjacent category where this arithmetic has already played out is herbal tea, where supply responded to demand quickly because the agronomy permitted it. Matcha’s constraint is that the agronomy does not.

Technology & Innovation

Technology and innovation

Tencha is a process specification before it is a product. The plant is shaded for the final weeks before the first-flush pick to raise theanine and chlorophyll and suppress catechins; the leaf is steamed to fix the colour, dried, and de-stemmed so that only the leaf tissue survives. Matcha is that tissue milled. Grade is therefore decided in the field and at the steamer, not at the mill — a fact that reorganises where technology matters in this market.

Productivity, not area, is now the supply lever

The 2025 season was poor: low minimum temperatures in Shizuoka from early April to early May suppressed bud development, and every volume line fell. The 2026 season reversed it. First-flush crude tea across the major prefectures rose 21 per cent to 24,100 tonnes on an area 1 per cent smaller, with Shizuoka up 32 per cent to 10,700 tonnes and Kagoshima up 14 per cent to 9,610 tonnes. Same land, more tea. In a market where the area has fallen to 33,400 hectares — roughly a quarter down in a decade — yield per hectare is the only supply variable still under management.

Mechanisation has already happened at the top of the volume curve

Machine-harvested leaf is 98.4 per cent of the tencha traded at the Uji auction. Hand-picking survives as a 9.4 tonne craft segment at ¥43,330 a kilogram, which is a cultural and commercial asset but not a supply source. The technology question in this market is therefore not “can we harvest at scale”. It is whether grade can be held at machine speed, and whether a converted sencha garden can be brought to tencha standard without losing the shading and cultivar characteristics that justify the Japanese premium.

What mechanisation did to the price of the top tier

The auction separates the two harvesting methods cleanly enough to price the craft. Machine-harvested leaf is 98.4 per cent of the volume and generated ¥10.101 billion of the ¥12.94 billion turnover. Hand-picked leaf is 9.4 tonnes, and at the 2025 hand-picked average of ¥43,330 it accounts for roughly ¥407 million — 3.1 per cent of turnover on 1.6 per cent of volume. The craft tier earns double its tonnage share and still cannot move the market, because 3.1 per cent of an auction is not a supply source.

The technical question the numbers leave open is whether grade survives speed. Machine-harvested leaf averaged ¥17,250 against ¥43,330 for hand-picked in the preceding season — a 2.51 times gap on the same kind of plot. Closing even part of it is the only technological objective in the chain with volume behind it: the machine tier is 585.6 tonnes against 9.4, so a one per cent improvement in machine grade touches 62 times the volume a perfect hand-picked season does.

Why the frontier is agronomy, not engineering

The processing side is mature: steaming, drying, de-stemming and milling are settled industrial operations with low patent activity. The unresolved problems are biological and demographic. New tencha fields take about five years to pay back, which means a planting decision made this year reaches the market around 2031 — outside the early half of the forecast window. Meanwhile the people who know how to make the leaf are ageing out, and the cultivars being replaced are being replaced by other teas, not by nothing.

That is why the innovation story with actual money behind it is happening where land and labour are not binding. Chinese capacity is being built at a scale that treats process know-how as an engineering procurement exercise rather than a craft inheritance. The volume that results will not be ceremonial grade. It will be the food-service and ingredient volume that already sets the demand curve — the same volume that fills functional beverage formulations where matcha appears as a claim rather than a taste.

Market Risk

Market risk

Four risks matter, and they are ranked below by how long their effect lasts rather than by how likely they are. A short, sharp price reversal is easier to trade around than a slow permanent shift in who supplies the market.

1. The price is the risk, and it is already in the model

This report assumes Japan’s realised price mean-reverts from USD 57.08 to USD 52.00 a kilogram by 2033, a decline of 8.9 per cent. That is deliberately mild. The upside case for Japanese growers is that a record auction is the new floor, not a spike, because the supply constraint is structural and the demand is not going away. The downside case is that the constraint is only structural for as long as demand holds, and demand for a café ingredient is exactly the kind that thins when the price of the latte rises.

What makes the downside sharper than usual is the switchability of the supply. A Japanese garden producing tencha in 2026 was, in most cases, a garden capable of producing sencha. Switching back is a decision taken over one season, not five years. That asymmetry — fast exit, slow entry — is the mechanism behind every agricultural price spike, and it is fully present here.

2. The composition shift is certain even if the total is not

Whatever the volume, the origin mix moves. China’s tonnage rises from 11,636 to 26,815 tonnes in the base case, a 2.3 times increase, while Japan’s share of pool value falls from 56.2 to 35.5 per cent. This is the least uncertain element of the forecast because it is driven by capacity that already exists and by a Japanese grower base that is contracting by demographics rather than by choice.

3. The premium tier has no buffer

Purchase limits of one canister per customer in Tokyo specialist retail are not a marketing device. They are what allocation looks like when a category has 9.4 tonnes of hand-picked leaf in its principal auction. The risk in the ceremonial tier is not that demand falls; it is that the tier cannot serve the demand it has already created, and that buyers diverted into machine-harvested higher grades never return to the top of the range.

4. Organic is a price risk, not a volume risk

Pheonix models organic rising from 8 to 11 per cent of tonnage and from 17.0 to 25 per cent of value, with the premium moving from 2.13 to 2.27 times the blended pool. If certification lags, those numbers move in an awkward direction: the share of tonnage grows slower than modelled while the premium grows faster. A buyer hedged on volume is therefore not hedged on price, and an organic contract struck on a volume assumption is exposed to exactly the variable it was meant to avoid.

The risk that reads as a loss and isn’t

Japan’s share of pool value falls 20.7 points across the window, from 56.2 to 35.5 per cent. That is routinely read as decline, and it is not one. Japanese value rises from USD 340.4 million to USD 408.4 million — 20.0 per cent growth in absolute terms — on volume that grows from 5,964 to 7,853 tonnes at a mean-reverted USD 52.00. Japan loses share because China grows faster, not because Japan contracts. Any strategy built on the share number alone will under-invest in the only origin that still earns a 2.51 times price.

The risk that is genuinely mispriced is the joint one. Separately, a repeat of the 2026 yield adds 1,253 tonnes, worth USD 65.1 million and 10.7 per cent of the pool, while a 20 per cent price fall costs USD 68.1 million, 11.2 per cent. Together they are not additive: a larger Japanese crop is exactly what would force the price down, so the two land in the same year on the same line. That combination is what turns the 15,034 tonne spread between the slow and accelerated cases from a band into a step — and it is the only scenario in which the 2025 auction stops being a floor.

The gap between cases is the honest measure of uncertainty

The slow case reaches 27,010.5 tonnes and USD 894.7 million; the accelerated case reaches 42,044.7 tonnes and USD 1,395.5 million. The distance between them is 15,034.2 tonnes and USD 500.8 million. That is a wide band, and it is wide for a reason: the forecast is not uncertain about demand, which has been strong in every scenario. It is uncertain about how fast a Chinese industrial base can clear compliance and how quickly Japanese yield recovers.

The risk this report has not priced is that both happen at once. A Japanese crop that repeats 2026 — crude tea up 21 per cent on a smaller area — would reset the auction faster than demand can absorb, at the same moment Chinese volume clears testing. Neither event is far-fetched. In that combination, the 2025 price is not a new floor, and the substitution that already reshaped other functional beverage categories arrives in matcha in a single season.

Regulatory Landscape

Regulatory landscape

Compliance in this market is not a single gate. It is four separate regimes that bind at different points in the chain, and they are the main reason the quality of the Chinese supply response is uneven rather than simply large.

The Positive List is the binding constraint on Chinese volume

A uniform 0.01 ppm limit across more than 800 residues is among the strictest food safety standards in the world, and it is applied without grade distinction. A producer either meets it or does not, at producer level and shipment level. That is why compliance, not capacity, is the rate limiter on Chinese-origin matcha reaching Japanese and Western buyers: the plants exist, the testing regimes have to be built and evidenced per producer, and third-party pre-shipment verification through an accredited laboratory has become a prerequisite rather than a differentiator.

The GI decision is about counterfeits, not tariffs

The registration of Japanese tea under the GI Protection System in July 2026 was a response to a specific problem that was becoming commercially material: products made outside Japan marketed under Japanese regional names, including “Uji matcha”. For a category whose entire premium rests on provenance that cannot be verified by taste at the point of sale, a protected name is the closest thing to an enforceable quality guarantee in the chain. Expect enforcement activity, and expect it to be slow, cross-border and therefore only partially effective.

Organic is a certification problem wearing a price label

The organic portion of the market is small because certification is slow, not because demand is thin. A garden must complete a conversion period under an approved certifying body before its output can be sold as organic, and during that period it carries the cost of the regime with none of the return. Combine that with generally lower yields under organic management and per-SKU certification cost for the brand, and the result is a certified area that responds to price with a delay measured in years.

This is the mechanism behind the scarcity that buyers describe. Pheonix models organic at 8 per cent of tonnage and 17.0 per cent of value in 2025, rising to 11 and 25 per cent by 2033; those are modelled shares and are labelled as such. What is not modelled is the arithmetic of the constraint: within a forecast window of eight years, a price rise in organic matcha cannot summon certified hectares, because the hectares in conversion do not count yet.

What the protected name is actually protecting

The GI registration of Japanese tea in July 2026 is usually described as a provenance measure. Priced, it is a defensive position around a specific sum. The Japanese realised price of USD 57.08 against the Chinese USD 22.79 is a gap of USD 34.29 a kilogram, and on 5,964 tonnes that is USD 204.5 million — 33.8 per cent of the entire pool. That is the asset a regional name defends, and it explains why enforcement will be contested commercially rather than politely: the counterfeited margin is a third of the market.

The Positive List is the reciprocal instrument on the other side. A uniform 0.01 ppm limit across more than 800 residues, applied without grade distinction, means the Chinese tonnes that would compete hardest for ingredient and food-service volume are the ones that must clear the highest evidentiary bar to land. Regulatory friction and provenance protection therefore push in the same direction — they slow the substitution of Chinese for Japanese material without altering its direction. What they change is the timing on which the 77.4 per cent Chinese tonnage share in 2033 depends.

Compliance cost is per SKU, per shipment and per origin

Every additional origin, grade and format multiplies testing, labelling and documentary obligations. That penalises precisely the strategy that would otherwise be most attractive in a supply-constrained market — a wide, multi-origin range. The practical result is that small importers consolidate into one origin, and large ones build compliance departments. Neither of those is a quality decision. Both are consequences of the paperwork.