Quality takes time. One bean at a time.
ICB's data picture of India's specialty coffee ecosystem: 84 roasters, 1,009 coffees, price distribution, regional concentration, processing trends, and what the metadata gaps reveal.
India's coffee industry has extensive public documentation. The Coffee Board of India publishes annual production figures (363,500 metric tonnes in 2024-25), export volumes (388,925 MT valued at Rs 15,422 crore), and holdings data covering approximately 4.91 lakh hectares. The USDA Foreign Agricultural Service produces independent trade analyses updated annually. These are reliable, regularly updated, and widely cited.
None of them map the specialty retail layer. No official body counts active specialty-style roasters in India. No public dataset tracks D2C specialty coffee SKUs at scale, documents processing method distributions across the market, or shows what Indian specialty coffee actually costs across a broad sample. The official statistics describe India's coffee economy — export-oriented, Robusta-led, commodity-scale — not the specialty retail ecosystem that has developed alongside it.
ICB tracks 84 specialty-style roasters and 1,009 active and seasonal coffees. What follows is a catalog-level snapshot of that ecosystem: who is in it, where the coffees come from, how they are processed, what they cost, and where the data is still incomplete.
India's coffee sector has two distinct layers. The commodity layer — production, exports, wholesale trade — is documented by the Coffee Board and USDA. It is dominated by Robusta, anchored in Karnataka and Kerala, and oriented toward export markets. The specialty retail layer — roasters, D2C SKUs, provenance labels, processing details, flavor language, and pricing — is separate from this. ICB measures the second layer only.
Two terms need clarifying before reading the figures that follow.
The first is catalog metrics vs. sales metrics. ICB counts listed coffees and active roasters. It does not measure units sold, revenue, or market share. A roaster with 71 listed coffees may sell fewer bags per month than one with 12. Share of the catalog is not share of the national specialty market.
The second is ICB's tracked catalog vs. the full market. ICB tracks roasters who sell D2C online and meet its listing criteria. Some specialty-style roasters operate exclusively through cafes, retail distributors, or local channels without an online storefront. The figures here describe ICB's tracked D2C specialty catalog, not a certified count of every specialty roaster operating in India.
Note: The Coffee Board's specialty taxonomy — Monsooned Malabar, Mysore Nuggets Extra Bold, Robusta Kaapi Royale, high-grown estate coffees — is a different categorization from what most D2C specialty roasters foreground today: single origin, processing method, variety, and micro-lot storytelling. Both frameworks are valid; they describe different layers of the same industry.
The ICB catalog currently tracks 84 roasters and 1,009 active or seasonal coffees, of which 930 are in stock at time of this snapshot.
The catalog is fragmented rather than concentrated. The top 10 roasters by coffee count account for 30.7% of all listed coffees. The remaining 74 roasters share the other 69.3%. This is a long-tail structure: a small number of roasters with broader catalogs alongside a distributed base with smaller, more specialized offerings.
Roaster headquarters geography shows how the specialty retail market has developed. Karnataka leads with 30 roasters, consistent with the state's production base, followed by Maharashtra (14), Delhi (12), Tamil Nadu (6), and Telangana (4), with smaller clusters in other states. Production remains anchored in the south. Retail specialty brands are nationally distributed, with significant presence in cities that have no coffee-growing history. The specialty retail market has, to a meaningful degree, decoupled from the production geography.
That decoupling means there is no single retail hub for Indian specialty. Bengaluru, Delhi, and Mumbai all have significant roasting presence, but none dominates by count — a structure not visible from production data, and one that shapes how specialty coffee reaches buyers across the country.
Origin data exists for 499 of 1,009 active coffees — a coverage rate of 49.5%. All regional figures below are shares of that labeled subset, not the full catalog.
Among region-labeled coffees, Chikmagalur accounts for 229 entries — 44.7% of the labeled pool. The next four origins are broadly comparable: Kodagu/Coorg (41), Baba Budangiri (40), Shevaroy Hills (38), and Sakleshpur (37). Together they represent another 31% of labeled coffees. Smaller representation comes from Araku Valley, Wayanad, Koraput, and Nilgiris.
Origin storytelling in Indian specialty remains heavily Karnataka-anchored. The five top origins are all Karnataka districts or adjacent ranges. That concentration comes from both the actual production geography and where roasters have invested most heavily in traceable sourcing.
On the unlabeled half: 510 coffees in the ICB catalog carry no region data — 50.5% of the active total. This reflects current disclosure practice across D2C Indian specialty rather than a tracking gap. Many roasters source from multiple estates or sell blends without specifying a sub-region on the product page. The distribution above is accurate for the labeled subset; it does not extend to the full catalog.
Process data is the most complete metadata field in the ICB catalog, with 92.8% coverage. But the most important figure to note before reading the distribution is the size of the uncategorized bucket.
Of 1,009 active coffees, 308 — roughly 30.5% — carry a process label of "other" or are effectively uncategorized. This is not an anomaly; it reflects ongoing taxonomy normalization across Indian specialty retail. Many lots use hybrid or multi-stage processes that do not map cleanly to a single standard label. The "other" bucket is partly a taxonomy problem and partly a market-stage signal — Indian specialty roasters are experimenting with process more diversely than the available standard labels can currently describe.
Within clearly labeled coffees: washed (213) is the largest single category, followed by natural (115), anaerobic (72), experimental (58), honey (57), washed-natural (33), monsooned (26), carbonic maceration (23), and double-fermented (21).
On the perception gap: The marketing visibility of experimental processing — anaerobic, carbonic maceration, double-fermented — runs ahead of its catalog share. Washed remains the largest clearly labeled category. Anaerobic and carbonic maceration lots together account for fewer than 10% of the catalog before the "other" bucket is considered. The relative scarcity of fermentation-forward lots partly explains why they carry higher price points and generate more discussion than their volume might suggest.
Price data comes from 1,007 in-stock coffees with valid pricing normalized to a 250g equivalent.
The median price is Rs 673 per 250g. The 25th percentile sits at Rs 525 and the 75th at Rs 900. By price bucket: 18.7% of coffees are priced below Rs 500; 40.2% between Rs 500-749; 24.2% between Rs 750-999; 9.0% between Rs 1,000-1,249; and 7.9% above Rs 1,250. The majority of the ICB catalog sits between Rs 500-1,000 per 250g.
Processing method correlates with price. Among coffees with both valid price and process data:
| Process | n | Median (250g) |
|---|---|---|
| Anaerobic | 83 | Rs 900 |
| Carbonic Maceration | 23 | Rs 900 |
| Experimental | 66 | Rs 890 |
| Natural | 121 | Rs 749 |
| Honey | 67 | Rs 730 |
| Washed | 219 | Rs 615 |
| Monsooned | 24 | Rs 598 |
Fermentation-forward and experimental processes cluster Rs 200-285 above the washed median. Carbonic maceration and double-fermented sample sizes fall below 30, making those medians directional rather than conclusive. That the premium holds consistently across the three largest non-washed sample groups — natural (n=121), anaerobic (n=83), and experimental (n=66) — suggests buyers have accepted it, not just that roasters are asking for it.
Flavor-note data is the strongest interpretive metadata field in the ICB catalog. Of 1,009 active coffees, 792 carry flavor descriptors — a coverage rate of 78.5%.
The six most frequently used descriptors across flavor-labeled coffees: Dark Chocolate (117 coffees, 14.8% of the flavor-labeled pool), Nuts (111, 14.0%), Fruity (104, 13.1%), Caramel (102, 12.9%), Chocolate (101, 12.8%), and Citrus (92, 11.6%).
The vocabulary is anchored in approachable, widely recognized descriptors. Dark chocolate, nuts, caramel, and citrus are terms most coffee buyers encounter without specialist knowledge. This comes from both the cup character common to Indian arabica — typically in the structured-to-balanced range — and the communication strategy most Indian specialty roasters use when writing for a broad consumer audience.
The absence of highly specific or exotic descriptors in the top tier is informative. Whether it comes from the coffees themselves, the buyers roasters target, or deliberate communication choices is not something the catalog can resolve on its own. The vocabulary pattern points to a common approach among Indian specialty roasters: write for buyers who already know chocolate and caramel, rather than training buyers toward a more specialized sensory language. Where that line falls tells you something about the Indian specialty buyer base and where roasters are pitching to it.
The coverage gradient is the clearest maturity signal in the ICB catalog.
Two fields are nearly complete: roast level (94.1%) and process (92.8%). Almost every D2C specialty coffee in India tells you how it was roasted and how it was processed. These are the two most standardized parts of the specialty communication layer in this market.
Flavor notes (78.5%) and brew methods (59.9%) are present for most but not all coffees. Region coverage (49.5%) sits at the midpoint — roughly half the catalog includes a geographic origin. Estate coverage (37.1%) and variety coverage (34.7%) are the thinnest fields.
For more than six in ten coffees in the ICB catalog, the specific estate is not disclosed. A buyer who wants to track provenance below the regional level — to follow a particular farm across harvests or compare estates within a region — can currently do so for only a third of the market.
What this gradient reflects: High process and roast coverage emerged because consumers began asking for it. Low estate and variety coverage reflects multi-estate sourcing practices, seasonal lot variability, and the fact that Indian retail buyers have not yet consistently demanded this level of traceability at the bag level. The gradient is a record of where market-driven disclosure has normalized and where it hasn't.
Ratings are the thinnest data layer of all. Only 63 coffees in the current ICB snapshot carry any rating data, totaling 87 individual rating events. This falls well below the threshold for statistical comparisons, regional rankings, or processing-method conclusions drawn from scores. The culture of public cupping scores and consumer ratings is early-stage in India compared to markets like the US or Australia. What the existing rating data supports: directional observations, not rankings.
The coverage gradient is the most informative single figure in this snapshot. A market where process and roast are nearly universal, flavor notes are strong, and estate and variety are sparse has normalized what consumers first asked for and is still building the traceability layer that typically comes next.
A clean specialty SKU growth chart is not currently buildable from ICB catalog data. Coffee creation timestamps in the database reflect ingestion and backfill timing, not market-entry dates. Using those timestamps as a growth series would misrepresent what the data actually records.
For the 27 roasters with founding year data — 32.1% of the tracked set — the 2010s show the largest cluster of new brands entering the specialty market. But with only a third of roasters providing founding year information, this is a partial picture.
External signals provide more complete growth context, though they describe adjacent market layers rather than the D2C specialty catalog directly:
On interpreting these signals: Branded chain expansion and D2C specialty roaster catalog growth are related but distinct. Chains and independent D2C roasters operate different supply chains and sourcing models. These figures document expanding consumer exposure to specialty coffee in urban India. They do not directly measure the D2C catalog layer ICB tracks.
The ICB catalog documents an ecosystem that is real, functioning, and structurally fragmented. 84 roasters and 1,009 coffees constitute a market layer that had no public map before ICB.
The production map is concentrated in the south, particularly Karnataka. The retail map is nationally distributed. The catalog is long-tailed: no small group of roasters controls it.
Washed coffees dominate among labeled processing. Experimental methods are present but account for a smaller share of the catalog than their marketing prominence implies. The 308 uncategorized coffees show that taxonomy normalization is still in progress.
Pricing concentrates between Rs 500-1,000 per 250g, with fermentation-forward processes carrying a clear premium over washed.
The metadata coverage gradient is the clearest maturity signal in the data: nearly universal on process and roast, strong on flavor notes, thin on estate and variety. That pattern maps directly onto what Indian specialty consumers started asking for first and what the market is still building toward.
The data exists. The legibility is still being built.
