Arduino VENTUNO Q edge AI computer gets DigiKey demo webinar

Board NewsBy November 18, 2026Fact-checked by Hugo Lindgren
Arduino VENTUNO Q edge AI computer gets DigiKey demo webinar

By Fact-checked by Hugo Lindgren Published

CONFIRMED

DigiKey is set to host a webinar demonstrating Arduino’s VENTUNO Q edge AI computer, according to a PR Newswire release headline. The Arduino Blog has also published a showcase of Moorcheh’s agentic memory and multimodal search running on the board. Together they give builders two places to see what a Qualcomm Dragonwing IQ8 board paired with a real-time microcontroller can do at the edge, though the webinar’s date, time and presenters are still unverified.

HardwareCONFIRMED A Arduino announced Dragonwing IQ8 plus a real-time MCU, sold via DigiKey and others
Launch priceCONFIRMED B CNX Software reports $299 introductory pre-order price
Moorcheh showcaseCONFIRMED A Arduino Blog post on agentic memory and multimodal search
Price rise to $399RUMOR C CNX Software article update, not verified against Arduino's store
Sub-27 ms retrieval and joint webinarsRUMOR D single Moorcheh LinkedIn post, vendor self-report

What we know

Arduino confirmed that VENTUNO Q pairs a Qualcomm Dragonwing IQ8 series processor with a real-time microcontroller. In its announcement, republished by Edge AI Vision, Arduino said the board would be sold through the Arduino Store and through resellers including DigiKey, Farnell, Macfos, Mouser and RS. A DigiKey demo fits that picture, since DigiKey is one of the named sales channels.

CNX Software reports that pre-orders opened at an introductory price of $299. It also pins down the silicon: a Qualcomm Dragonwing IQ8 (IQ-8275) SoC alongside an STM32H5 microcontroller. That split matters to anyone who has fought jitter on a Linux board. The Qualcomm side handles the operating system and AI workloads. The STM32H5 is there for timing-critical I/O.

The fuller spec sheet comes from CNX Software, Arduino’s store and other outlets. It lists an octa-core Kryo CPU, a Hexagon NPU rated up to 40 TOPS, an STM32H5F5 Cortex-M33 microcontroller, 16 GB of LPDDR5, 64 GB of eMMC and an M.2 NVMe slot. The board reportedly runs Ubuntu with App Lab. Treat that list as reported and check Arduino’s store page for the final figures before you order.

The schedule has already moved. Arduino announced VENTUNO Q in March 2026 and said it would be available in Q2 2026. Pre-orders opened on 2026-08-25, so the Q2 target slipped.

The Arduino Blog published a post on building agentic memory at the edge with Moorcheh on VENTUNO Q. The post covers a multimodal search engine scaled to millions of records. In practice, agentic memory is the store an AI agent queries to recall earlier context. Running that locally, without a cloud round trip, is the pitch.

Three similar-sounding events are easy to confuse:

  • DigiKey and Arduino ran a webinar called “From Board to Build: Using UNO Q and App Lab” on 2026-02-12. It covered UNO Q, not VENTUNO Q.
  • element14 lists its own VENTUNO Q and edge AI webinar, presented by Andrea Richetta. It is not a DigiKey event.
  • The new DigiKey VENTUNO Q demo is the one named in the PR Newswire headline.

What is rumored

The Moorcheh team says in a LinkedIn post that it is putting its search engine and its Memanto agentic memory on VENTUNO Q. The post claims sub-27 ms on-device retrieval and says joint webinars with Arduino and Qualcomm are planned. This is a single social post by the vendor about its own product. We rate it grade D. The performance figures are self-reported, and we found no independent benchmark.

There is a useful reference point. The Memanto paper (arXiv 2604.22085) describes Moorcheh’s information-theoretic search engine as the base for agent memory and reports sub-90 ms latency. That is a different number from the LinkedIn claim. Neither source tells us whether the two figures come from the same hardware, dataset size or query type, so they should not be compared directly.

On price, an update attached to CNX Software’s article says Arduino has ended the $299 introductory offer and VENTUNO Q now sells for $399. CNX Software is normally a reliable outlet, but we could not verify this change against Arduino’s store or a second source. We rate it grade C until Arduino’s own listing confirms it.

What we don’t know yet

The DigiKey webinar’s date, time and presenters. We could not retrieve the PR Newswire release text, so only the headline is confirmed. DigiKey’s webinar page or the full release should settle it.

Whether the DigiKey demo will include the Moorcheh search and memory work, or focus on the board and App Lab in general.

Whether the joint Arduino, Qualcomm and Moorcheh webinars mentioned on LinkedIn will happen. Arduino and Qualcomm have not confirmed them in anything we reviewed.

The current price. It is either the confirmed $299 introductory figure or the reported $399, and Arduino’s store is the place to check.

Independent retrieval numbers. Nobody outside Moorcheh has published a VENTUNO Q latency figure for this workload that we could find.

What it means

For builders, VENTUNO Q is a different class of board from a typical maker SBC. If the reported specs hold, a 40 TOPS NPU, 16 GB of LPDDR5 and NVMe storage on one board, with a Cortex-M33 handling real-time pins, is a serious platform for robots and vision projects that need local inference and deterministic motor or sensor control at the same time.

The Moorcheh showcase on the Arduino Blog is the most concrete sign so far of what Arduino expects people to run on it. Multimodal search over millions of records, with an agent memory layer on top, is a workload you would normally send to a server. Keeping it on the device helps with privacy, offline robots and latency. Treat the 27 ms figure as a vendor claim, though. The vendor’s own paper reports sub-90 ms, so measure on your own data before you design around either number.

If you are deciding whether to buy, a live demo is the cheapest way to answer practical questions. How does App Lab bridge the Linux side and the STM32H5? What does the NPU toolchain look like? How much of that 16 GB is left once a model and a vector store are loaded? Bring those questions to the DigiKey session once its details are published, or to element14’s separate session.

Budget for the board at $299, but plan for $399. The CNX Software update may turn out to be right, and Arduino has already shown with the Q2 slip that its timelines can move.

Follow every development in one place on our board release tracker.

Notes on this article

Written by

Wren covers board news and our Build of the Week, rounding up new dev boards, price changes and the most interesting community projects. She is drawn to clever builds made from cheap, easy-to-find parts.

Covers: Board News, Build of the Week

  • Fact-checked by Hugo Lindgren before publication.
  • Published

Fact-checking

  • Every claim is checked by Hugo Lindgren; anything we cannot confirm is corrected, labelled or removed.