AI-sorted wine barrels enter Spanish cellar trials

What the pilot is about

Researchers at the University of Valladolid have placed 48 wine barrels sorted with artificial intelligence into cellar trials in Spain. The project’s central question is whether classifying wood by oxygen permeability could make barrel ageing more predictable.

Oxygen permeability simply means the extent to which oxygen can pass through the wood. The idea behind the pilot is that, if barrels can be grouped in advance according to this property, wineries may be able to choose them more deliberately for different ageing objectives.

The available material does not name the participating wineries, the types of wine, the origin of the wood or the length of the trial. It also provides no results. This should therefore be viewed as an early cellar test rather than a proven commercial solution.

How it works

The confirmed workflow has three basic elements: assess the wood’s oxygen permeability, use artificial intelligence to classify the barrels, and test 48 of them under cellar conditions. In this context, AI means a computer method that uses data to sort or group items.

The source does not explain how oxygen permeability is measured, what data are provided to the algorithm, how many categories are used or how the accuracy of the classification is checked. It is also unclear whether finished barrels were measured directly or whether they were sorted using characteristics recorded from the wood.

Importantly, the AI does not age the wine itself, and the supplied information does not suggest that it actively controls oxygen entry. Its stated role is classification. The cellar trial can then show whether those classifications are useful under practical winemaking conditions.

Why it matters for vineyards and wineries

If the classification proves reliable, winemakers could potentially select barrels more consistently for a desired ageing objective. This might support barrel allocation in the cellar, ageing plans and the management of separate wine lots. From a business perspective, oxygen permeability could also become a useful, comparable specification when wineries purchase barrels.

No direct change in vineyard operations is indicated by the available information; the potential benefit lies mainly in cellar work and procurement. There are also important limitations. Forty-eight barrels represent a restricted trial, and no evidence has yet been reported showing that the categories produce more predictable winemaking outcomes. There is no information about costs, equipment requirements or commercial availability either.

The trial also does not establish that one permeability category is universally better than another. Its possible value is in matching a known barrel characteristic to a winery’s intended use, provided that the classification is accurate and repeatable.

What to watch next

The most important next signal will be the release of cellar results, particularly how closely the barrels’ actual behaviour matches their AI-assigned categories. It will also matter whether the method can deliver repeatable classifications across different wines, wood materials and ageing periods.

For wineries, future adoption would likely depend on measurement accuracy, categories that cellar teams can understand, implementation costs and whether barrel suppliers can apply the system consistently. Until those details and trial results are available, the project is best treated as a promising classification pilot rather than a ready-made technology standard.

Source: material published by Bing News – Wine AI about the University of Valladolid research programme.