Excelsa is a coffee species, Coffea dewevrei, native to Central Africa and now farmed in Uganda, South Sudan, India and parts of Southeast Asia. It was long treated as a variety of Liberica, and Kew’s 2025 genomic study restored it as its own species. Kew estimates Liberica and excelsa together at probably under 1,000 metric tons a year.
It is not native to Southeast Asia. Its wild range runs across the Republic of the Congo, Cameroon, the Democratic Republic of Congo, the Central African Republic, South Sudan and Uganda. Kew says that in Uganda, South Sudan and Guinea “attention is focused on excelsa owing to its ability to grow and produce commercially viable crops under higher temperatures and extended periods of low rainfall compared with robusta”, and that its production is also increasing in India and, “to supplement robusta and diversify coffee production”, in Vietnam and Indonesia. The Asian farms are where the “Southeast Asian origin” story comes from, and why it is wrong: they are places it was taken to, not where it came from.
Table of Contents
The taxonomy, in order
- Historically, excelsa circulated under the name Coffea excelsa.
- In 2006, C. excelsa, C. dewevrei and C. dybowskii were sunk into Coffea liberica, with excelsa treated as C. liberica var. dewevrei — a variety of Liberica, not a species. That is the name used in the title of the 2022 processing study described below.
- In 2025, Davis and colleagues at Kew sequenced 353 nuclear genes across 55 samples and found three distinct lineages. Their conclusion: “Genomic and other datasets support the division of Coffea liberica into three species: C. liberica, C. dewevrei (excelsa) and C. klainei” (Davis et al., Nature Plants 11: 1729–1738, 2025). Excelsa is a species again — but under the name Coffea dewevrei, not C. excelsa.
Older references, including plant databases that have not yet been updated, will still show C. excelsa as a synonym of C. liberica var. dewevrei; that reflects the 2006 treatment, not the current one.
Kew’s 2026 follow-up adds a complication. It found that Liberica and excelsa “have hybridized readily in cultivation and produced fertile progeny with a wide range of genomic admixture”, with “extensive genomic admixture in farmed accessions from Sarawak”, and that the hybrids show “intermediate characteristics and overlapping values for key agronomic traits, including seed size and parchment thickness”. The authors propose the name Coffea × libex for the hybrid (Davis et al., Scientific Reports 2026). Kew describes the hybrid as “rare in the global coffee value chain and currently only available in Sarawak, where it is currently sold as C. liberica”, and the six cultivated accessions it identified as excelsa showed only “low levels of admixture from C. liberica”, from 1 to 3%. In our reading, the hybrid question matters more for Liberica than for excelsa. Neither chemistry study behind the caffeine figures below reports genotyping its samples.
Where excelsa grows and why growers are planting more of it
The Kew study is specific about the ecology, and it explains the commercial interest better than any tasting note does. In the wild, C. dewevrei occupies mid-elevations — a mean of 654 m — while C. liberica and C. klainei sit in the lowlands at means of about 386 m and 273 m. In cultivation, Kew says excelsa is “principally farmed at mid-elevations (500–1,200 m)”, with a mean annual temperature of 22 to 25 °C and lower rainfall, 1,500 to 1,800 mm a year, and Kew writes that “field observations suggest that C. dewevrei exhibits a considerable measure of drought tolerance, particularly compared with C. canephora“.
For Liberica and excelsa together, Kew names two drivers: “climate challenges affecting Arabica and robusta, coupled with an increasing market demand”. It says their production is now being upscaled “particularly in Uganda, South Sudan, India, Vietnam, Malaysia, the Philippines, Indonesia and even the Pacific”. Our page on climate change and coffee farming covers the pressure driving that shift.
The bean itself is smaller than Liberica’s: Kew gives seed dimensions of 7.7–11.3 × 5.4–8 mm and fruit of 1.2–2.5 × 0.8–2.1 cm, describing the seeds as “similar in size and shape to Arabica”. That is a useful practical detail: Kew says the smaller seeds make excelsa “more amenable to existing post-harvest, and preconsumption (roasting, packaging and coffee making) processes”.
What does excelsa coffee taste like?
Nobody at this site has cupped excelsa, so this page will not publish tasting notes as though it had. What can be sourced is this: the 2022 International Journal of Food Science study cupped Indonesian excelsa prepared by several post-harvest processes and scored every sample above 80 points, with a top score of 83.83 for one natural-processed lot — specialty grade by Arabica scoring conventions. The authors summarised the character as “more similar with typical fruity aroma of Arabica”, and identified roasted, nutty pyrazines alongside phenolic and smoky guaiacol among the impact compounds. The same paper suggests the fruity side may owe something to isovaleric acid and its ester, which “are important volatile compounds in jackfruit” and “may contribute to jackfruit aroma, which are commonly detected in Excelsa coffee”; it notes that the acid alone gives a “rancid and cheese aroma” and that the ester is the fruity one. All six of its lots scored between 82.33 and 83.83 whichever process was used. The Kew paper adds only that “Liberica and excelsa have contrasting coffee flavour profiles”, which is why the market treats them separately.
How much caffeine does excelsa have?
Published figures disagree by more than a factor of three, so treat any single number as unsettled. A 2026 analysis of green beans grown on the Comoros measured excelsa at 0.89% caffeine by weight, against 1.52% for robusta from the same plantation on Ngazidja, harvested in July 2022 (Irchad et al., Metabolites 2026). A 2026 review of the Liberica group gives “approximately 1.2 to 3.8 g/100 g on a dry basis” for Liberica and excelsa green beans and concludes that “Liberica and Excelsa occupy an intermediate caffeine position, generally lower than Robusta and overlapping with or slightly higher than Arabica” (Kurniawan et al., Molecules 2026;31(9):1518).
| Source | Beans measured | Excelsa caffeine | For comparison |
|---|---|---|---|
| Irchad et al. 2026 | Green beans grown on the Comoros | 0.89% by weight | Robusta from the same plantation: 1.52% |
| Kurniawan et al. 2026, table 3, from the 2022 study | Wonosalam, East Java, Indonesia | 3.15 g/100 g, dry basis, as the review reports it | Arabica from Enrekang, Indonesia: 2.80 g/100 g |
| Kurniawan et al. 2026, text | Liberica and excelsa green beans, pooled | 1.2 to 3.8 g/100 g, dry basis | Ethiopian arabica: 0.78 to 1.55% |
The review says its range comes from “comparable brewing-based analyses”, and the 2022 study behind the second row measured caffeine in a brew made from the green beans, while the Comoros study analyzed the green beans themselves, so the rows are not directly comparable. The review’s own table also prints arabica at 2.80 g/100 g, above the dry-bean ranges in its text, which is part of why the figures disagree.
Does excelsa coffee have health benefits?
We found no study of excelsa’s health effects in people: a PubMed search on 6 October 2026 for excelsa or dewevrei with coffee or Coffea in the title or abstract returned 34 records, none of them a study of excelsa in people, and this page will not invent one to fill the gap. What exists is laboratory work on the bean’s chemistry. The newest, a 2026 analysis of Comoros-grown stock, measured less of the compounds usually cited for coffee’s antioxidant reputation than robusta grown on the same islands: total polyphenols of 49.08 ± 4.25 mg GAE/g against robusta’s 121.79, total flavonoids of 15.98 ± 3.26 mg QE/g against 29.43, and antioxidant activity of 41.74 ± 5.29 mM Trolox equivalent/g against 64.97 (Irchad et al., Metabolites 2026;16(5):303). The same study found excelsa’s mineral content — magnesium, calcium, sodium, zinc and copper — measurably higher than robusta’s, describing excelsa as offering “a nutritionally balanced profile with nutraceutical potential” rather than robusta’s higher raw antioxidant and stimulant load.
The 2026 review of the Liberica group is also explicit about what nobody has done yet: “the absence of integrative studies linking well-defined bioactive concentrations to sensory descriptors and clinically relevant health outcomes represents another critical gap.” In plain terms, no study lets this page say excelsa coffee does anything specific for the person drinking it — only what has been measured in the bean itself.
Buying and brewing it
Excelsa is a small, specialist crop, so the buying advice is mostly about traceability: look for a roast date, a named country and producer, and the processing method. Given the seed size sits close to Arabica, standard Arabica grind settings are a reasonable starting point rather than a guess.
Start at 55 g of coffee per 1 litre of water — about 1:18 by weight, the standard brew ratio in SCA Standard 310-2021, Home Coffee Brewers: “55 g of coffee per 1.000 kg of water”. For water temperature see our page on the best temperature for French press coffee. Filter brewing gives you the most information about an unfamiliar coffee; save espresso for the second bag.
Affiliate disclosure: the two links below are paid links. As an Amazon Associate I earn from qualifying purchases. Your price is the same either way. Neither brewer has been tested by us — they are the two standard filter devices, listed because filter brewing is what this section recommends.
For the species this one keeps being confused with, see our pages on Liberica coffee beans, Arabica coffee beans and Robusta coffee beans.
Corrections to earlier versions of this page
An earlier version of this page cited a study that does not exist.
The page said: “A study published in the Journal of Nutrition found that coffee consumption was associated with a reduced risk of all-cause mortality, cardiovascular disease, and several types of cancer” — and linked that sentence to PubMed Central record PMC9410941.
We opened that record. It is “Profile of Bioactive Compounds, Aromas, and Cup Quality of Excelsa Coffee (Coffea liberica var. dewevrei) Prepared from Diverse Postharvest Processes”, published in the International Journal of Food Science in 2022. Wrong journal, wrong title, wrong subject. The paper is a chemical and sensory analysis of processing methods and says nothing at all about mortality, cardiovascular disease or cancer. The health claim has been removed. So has a second claim on the same page — that antioxidants in Excelsa are “linked to a lower risk” of type 2 diabetes, certain cancers and neurodegenerative disease — which had no citation behind it either. This page makes no health claims and is not medical advice; if you have a reason to be careful about caffeine, that is a conversation with a clinician, not with a coffee site.
The mis-cited paper is relevant all the same: it is the 2022 processing study quoted on this page, and its title uses the 2006 name.
An earlier version also asserted three things that were deleted rather than corrected, because there was nothing to correct them against:
- Caffeine percentages. The page printed “1.9% – 2.4%” for excelsa, “1.2% – 1.5%” for Arabica and “2.7% – 4.0%” for Robusta with no source. The Kew study reports no caffeine figures for any of the three Liberica-group species. Rather than keep unsourced numbers, this page now gives only published measurements, set out in the caffeine section above.
- A ten-row comparison table rating excelsa, Robusta and Arabica on acidity, body, aroma, price range and market demand. None of it was measured or cited.
- “Discovered and cultivated in the late 19th century” in Vietnam and the Philippines. Wrong continent, as above.
Also corrected: an earlier version called excelsa and Liberica “two distinct coffee varieties”, using a superseded name and a rationale with no source; it said the 2022 processing study reports caffeine “only qualitatively”, although that study measured it and a 2026 review takes a figure from it; it said drought tolerance was “the whole story” of excelsa’s expansion, where Kew gives several reasons, among them climate pressure on arabica and robusta, market demand, and excelsa’s performance under higher temperatures and low rainfall; it gave the countries where Liberica and excelsa together are being upscaled as excelsa’s alone; it said the excelsa plant is large, which we could not source; and it said nobody had published research on excelsa’s health effects in humans without saying how that was checked.
