# DATALAND in Los Angeles: How the Museum of AI Arts Builds an Immersive, Responsive Space

> DATALAND is a museum of AI arts founded by Refik Anadol and Efsun Erkılıç that opened on 20 June 2026 at The Grand LA in downtown Los Angeles. It has five galleries in about 25,000 square feet. Its opening show, Machine Dreams: Rainforest, combines projection, sound and scent generated with an AI model trained on nature data, and the rooms respond to visitors' movement and, optionally, their body signals.

**Published:** 2026-11-09 - **By:** MLABART

## What is DATALAND?

DATALAND calls itself the world's first museum of AI arts. It was founded by the artist Refik Anadol and Efsun Erkılıç of Refik Anadol Studio, and it opened to the public on 20 June 2026 at The Grand LA, the Frank Gehry-designed complex at 100 South Grand Avenue in downtown Los Angeles. MLABART was not involved in the project; this article describes it from published sources and looks at what the format means for brand and exhibition spaces.

The museum has five galleries in roughly 25,000 square feet of exhibition space. Epson, which supplied the projection, says a further 10,000 square feet houses the technology behind the galleries, and that the space was designed with the architecture firm Gensler and engineered with Arup. The opening exhibition is Machine Dreams: Rainforest.

| Fact | Detail | Source |
| --- | --- | --- |
| Opened | 20 June 2026 | AFP; DATALAND website |
| Location | The Grand LA, 100 S Grand Ave, downtown Los Angeles | Wikipedia; Blooloop |
| Size | About 25,000 sq ft of exhibition space in five galleries, plus 10,000 sq ft for technology | Epson press release |
| Opening exhibition | Machine Dreams: Rainforest | MyNewsLA; Wikipedia |
| Visit | Timed entry in 30-minute slots; about 1.5-2 hours; tickets from US$49; closed Mondays | DATALAND website, read 9 November 2026 |

- [DATALAND -- official website](https://dataland.art/)
- [AFP via Dawn -- AI museum brings sights, sounds and smells of the rainforest (19 June 2026)](https://www.dawn.com/news/2009108)
- [MyNewsLA -- AI art museum DATALAND to open in downtown LA (23 April 2026)](https://mynewsla.com/arts-and-entertainment/2026/04/23/ai-art-museum-dataland-to-open-in-downtown-la/)
- [Wikipedia -- Dataland](https://en.wikipedia.org/wiki/Dataland)

## What makes it an AI space rather than a projection room?

Most immersive rooms play a finished film on a loop. At DATALAND the picture, sound and scent are produced with a generative model, and they change with what the room senses. Three kinds of input are described in published reports:

- Live nature data. Epson's release says the artworks are inspired by live data from 16 rainforest locations.
- Movement. AFP reports wall-mounted sensors that track visitors' movement through the galleries.
- Body signals. Guests can wear a watch-like device that reads heart rate and skin temperature, according to AFP. The museum's website says participation is entirely optional and that only guests aged 13 and over may wear the devices.

- ![Diagram, drawn by MLABART from published descriptions, of what goes into DATALAND's rooms and what comes out. Inputs: live data from 16 rainforest locations (Epson release), visitor movement tracked by wall-mounted sensors (AFP), and optional body signals from a watch-like wearable reading heart rate and skin temperature (AFP). These feed the Large Nature Model, a generative AI model trained on nature images and sound. Outputs: picture from 84 laser projectors in the largest gallery (Epson release), rainforest sound such as macaw calls (AFP), and scent from a portable diffuser each guest may carry (AFP). MLABART was not involved in DATALAND.](https://mlabart.com/insights/dataland-los-angeles-ai-art-museum-immersive-space/inputs-and-outputs.svg)

## What is the Large Nature Model?

The Large Nature Model is the generative AI model behind the exhibition. DATALAND describes it as open source and trained on an ethically collected dataset of the natural world, with a focus on the flora, fauna and fungi of rainforests. The museum's website shows the Natural History Museum, the Smithsonian, the Cornell Lab and Getty as partners. Wikipedia's entry gives the scale as more than 500 million nature images and over 50,000 audio recordings.

For anyone planning a similar space, the point is where the data comes from. A model trained on licensed or partner-supplied material can be shown in public without the copyright questions that follow models trained on scraped images, and the museum makes that sourcing part of its story.

- [DATALAND -- Large Nature Model](https://dataland.art/about/large-nature-model)

## How is the picture made?

The largest gallery, Gallery A or the Data Pavilion, is a projection space on an unusual scale. Epson's press release gives these figures:

- Another gallery, the Infinity Room, is a 12 x 12 foot cube with mirrored walls, floor and ceiling. Blooloop reports that the work was first built in 2014 and has since been shown in 35 cities.
- Scent is delivered per person. AFP describes a portable scent diffuser that each guest carries, so the smell follows the visitor and does not have to fill and then clear a whole room.

| Item | Figure (Epson press release) |
| --- | --- |
| Projectors | 84 synchronised Epson EB-PQ2220B 4K laser projectors |
| Brightness | 20,000 lumens each, over 1.5 million lumens in total |
| Resolution | Nearly 700 million pixels across the array |
| Alignment | Camera-assisted calibration and edge blending across all 84 projectors |

- [Epson -- Powering DATALAND (press release, PR Newswire)](https://www.prnewswire.com/news-releases/powering-dataland-epson-projection-helps-bring-the-worlds-first-museum-of-ai-arts-to-life-302893877.html)
- [Blooloop -- Refik Anadol's DATALAND (24 October 2025)](https://blooloop.com/refik-anadol-dataland-opening-2026/)

## What can a brand or exhibition space take from it?

Few projects have a museum's budget or 84 projectors, but the method scales down. These are general planning points, not DATALAND's own guidance:

- Decide what the room listens to. Movement, touch, sound, a live data feed or a wearable each give a different kind of response. One clear input that visitors understand works better than several they do not notice.
- Make the response visible within a second or two. If visitors cannot tell that the room reacted to them, the sensing is wasted.
- Keep sensing optional and say what is collected. DATALAND makes wearables opt-in and sets an age limit. Any space that reads faces or bodies needs the same clarity, and a way to enjoy the room without taking part.
- Plan generated content like live content. A generative system needs limits on what it can show, a fallback loop if it fails, and someone responsible for checking output before the public sees it.
- Budget the back of house. DATALAND gives roughly two square feet to technology for every five of gallery. Projector rooms, servers, cooling and access routes are part of the floor plan.
- Use scent sparingly. Scent delivered per person, or in a small enclosed zone, is easier to control than scent released into a large room.
- Design for timed entry. A two-hour visit in 30-minute entry slots is a capacity decision as much as a ticketing one; see our note on throughput in the World Expo article below.

- [MLABART -- Interactive installations at World Expos (capacity and throughput)](https://mlabart.com/insights/world-expo-pavilion-interactive-installations/)
- [MLABART -- Building an immersive projection space](https://mlabart.com/insights/immersive-projection-space-installation-guide/)

## What has MLABART built along these lines?

MLABART builds responsive rooms at brand and exhibition scale. Three projects use the same building blocks as the points above, on a much smaller footprint:

- A room that reads movement. At Armani/Privé's L'Exposition at the Shanghai Exhibition Center (10-26 February 2023), an Azure Kinect camera tracked visitors' movement and gestures, which appeared as flowing particle silhouettes on an LED wall. Visitors scanned a QR code to retrieve a video of their interaction.
- Scent made visible. For an Atelier Cologne fragrance pop-up, MLABART built a blue immersive space with a light-up scent-spray interaction and a printer that produced a personalised fragrance card.
- Blended projection on a physical surface. For Starbucks' "A Thousand Li of Rivers and Mountains" activation, three 1920 x 1080 laser projectors were aligned and edge-blended onto a fabricated terrain model, with scheduled daily startup and shutdown. The project ran for 45 days.

- [MLABART -- Armani/Privé L'Exposition case study](https://mlabart.com/work/armani-fragrance-art/)
- [MLABART -- Atelier Cologne case study](https://mlabart.com/work/atelier-cologne/)
- [MLABART -- Starbucks case study](https://mlabart.com/work/starbucks-a-thousand-li/)

- ![MLABART's motion-tracked particle installation at Armani/Privé L'Exposition in Shanghai: a visitor's movement appears as a flowing particle silhouette on an LED wall.](https://mlabart.com/cases/armani-fragrance-art/particle-motion.webp)
- ![MLABART's projection-mapped terrain model for Starbucks' "A Thousand Li of Rivers and Mountains" activation, lit by three edge-blended laser projectors.](https://mlabart.com/cases/starbucks-a-thousand-li.jpg)

## Sources

Facts about DATALAND come from the museum's website (read 9 November 2026), an Epson press release, AFP reporting carried by Dawn, MyNewsLA, Blooloop and Wikipedia, as cited in each section. Figures from the Epson release are the supplier's own. Planning points are general practice. MLABART projects are described from project records.


## FAQ

### What is DATALAND and where is it?

DATALAND is a museum of AI arts founded by Refik Anadol and Efsun Erkılıç. It opened on 20 June 2026 at The Grand LA, 100 South Grand Avenue in downtown Los Angeles, with five galleries in about 25,000 square feet of exhibition space.

### How does DATALAND respond to visitors?

Published reports describe three inputs: live data from 16 rainforest locations, wall-mounted sensors that track movement, and an optional watch-like wearable that reads heart rate and skin temperature. An AI model called the Large Nature Model uses them to shape picture, sound and scent.

### How many projectors does DATALAND use?

Epson says the largest gallery, the Data Pavilion, uses 84 synchronised EB-PQ2220B 4K laser projectors of 20,000 lumens each -- over 1.5 million lumens and nearly 700 million pixels in total -- aligned with camera-assisted calibration and edge blending.

### Can a brand build an AI-driven immersive space on a smaller budget?

Yes. The method scales down: pick one input the room listens to, make the response visible within a second or two, keep any body sensing optional, give generated content limits and a fallback, and plan back-of-house space. MLABART has built movement-reactive LED walls and edge-blended projection at brand-event scale.



## Related Pages

**See also:** [Immersive Art Exhibitions](https://mlabart.com/solutions/immersive-exhibitions/) | [AI Interactive Installations](https://mlabart.com/solutions/ai-interactive-installations/) | [Projection Mapping](https://mlabart.com/solutions/projection-mapping/) | [Motion & Body-Tracking Interactives](https://mlabart.com/solutions/motion-tracking-interactives/)

---

**MLABART** -- Interaction - Imagination - Innovation

MLABART is an interactive installation and interactive software development studio. We design, program and deliver interactive installations -- motion-sensing games, touch and AI experiences, interactive LED walls, projection mapping, naked-eye 3D and kinetic sculptures -- for mall and New Year's Eve countdown campaigns, trade-show booths and launches, and brand pop-ups, from concept and custom software through fabrication and on-site delivery.

- Web: https://mlabart.com
- Email: mark@touchworld-sh.com
- Phone / WhatsApp: +86 18917292695
- Canonical HTML version: https://mlabart.com/insights/dataland-los-angeles-ai-art-museum-immersive-space/