[URGENT] Hardware Regression: NVIDIA Ada/Blackwell Support on ChromeOS Flex (Dell Precision 3590 & similar)

[John Cooper]

Hi everyone,

I am writing to advocate for a Priority 1 (S1)escalation regarding the lack of support for modern NVIDIA architectures—specifically Ada Lovelace and the new Blackwell generation—in the current ChromeOS Flex (reven) builds.

The Issue

On certified workstation hardware like the Dell Precision 3590, attempting to use the discrete GPU (dGPU) by disabling "Switchable Graphics" in BIOS results in a total boot failure or a persistent black screen. This occurs because the current Flex kernel lacks the NVIDIA Open GPU Kernel Modules required to initialize these newer chips.

Why this is an S1 Case

1. Hardware Mandate: Starting with Blackwell, NVIDIA has transitioned to requiring open-source kernel modules for GSP-based (GPU System Processor) architectures. Without these, the hardware is functionally a "brick" on ChromeOS Flex.
2. Certified Workstation Regression: Professional laptops (Precision, ZBook, ThinkPad P-series) are being shipped with RTX 500/1000 Ada and RTX 50-series Blackwell chips. Flex cannot currently support the primary selling point of these workstations.
3. Community Impact: This is not an isolated bug; it affects the entire 2024–2026 hardware roadmap.

Call to Action: Seeking Logs

If you are running ChromeOS Flex on a machine with an RTX 40-series/Ada or RTX 50-series/Blackwell GPU, please contribute to the investigation:

1. Expose the dGPU: In your BIOS, Disable Switchable Graphics (or enable Discrete Graphics Controller Direct Output Mode).
2. Capture logs: If you can reach a shell (Ctrl+Alt+F2), run:
   • lspci -nnk | grep -A 3 VGA
   • dmesg | grep -iE "nouveau|nvidia|nvrm|drm"
3. Submit: Please attach these logs to Chromium Issue 483088589.

The goal is to provide the engineering team with enough unique PCI IDs and crash signatures to justify the integration of the nvidia-open modules into the standard Flex kernel image.

Best regards,

John Cooper

https://issues.chromium.org/issues/483088589

[Mike Frysinger]

filing issues is the right first step, but requesting prioritized support is done through existing support agreements & internal channels, not via the public mailing lists.  so please reach out via that communication channel.

https://chromeos.google/products/chromeos-flex/

-mike

--
--
ChromiumOS Discussion mailing list: chromium-...@chromium.org
View archives, change email options, or unsubscribe:
https://groups.google.com/a/chromium.org/group/chromium-os-discuss

[John Cooper]

Hi Mike and Chromium OS Discuss Members,

Thanks for the clarification. I have filed the official feedback report via the device and linked the technical details in the Chromium Issue Tracker (Issue #483088589). To support the case for prioritizing this hardware: the Dell Precision 3590 is the workstation 'twin' to the Latitude 5550, which is a staple in the high-volume enterprise fleets Google targets for Chrome OS Flex.

Currently, the kernel identifies the hardware (platform: NVDA), but as a 'Muxless' design, it requires the NVIDIA Open GPU Kernel Modules to handle the RTX 500 Ada (10de:28b8) correctly. Since this AD107-based architecture is new 2024/25 standard for professional laptop workstations (Dell Precision, HP ZBook, Lenovo P-series), addressing this gap now would provide a clear path for enterprise 'Power Users' looking to migrate their entire fleet to Flex.

I've documented the specific PCI requirements from the NVIDIA-Open GitHub in the bug report to assist the dev team with the initial handshake.

Regards,

John

On Fri, 13 Feb 2026 at 17:47, John Cooper <johnpco...@gmail.com> wrote:

Hi Mike,

Thanks for the clarification. I have filed the official feedback report via the device and linked the technical details in the Chromium Issue Tracker (Issue #483088589). To support the case for prioritizing this hardware: the Dell Precision 3590 is the workstation 'twin' to the Latitude 5550, which is a staple in the high-volume enterprise fleets Google targets for Chrome OS Flex.

Currently, the kernel identifies the hardware (platform: NVDA), but as a 'Muxless' design, it requires the NVIDIA Open GPU Kernel Modules to handle the RTX 500 Ada (10de:28b8) correctly. Since this AD107-based architecture is new 2024/25 standard for professional laptop workstations (Dell Precision, HP ZBook, Lenovo P-series), addressing this gap now would provide a clear path for enterprise 'Power Users' looking to migrate their entire fleet to Flex.

I've documented the specific PCI requirements from the NVIDIA-Open GitHub in the bug report to assist the dev team with the initial handshake.

Regards,

John

[John Cooper]

Additionally enabling the RTX 500 Ada would provide access to its Tensor Cores, which would go together with the NPU cores which are already available with the CPU (Intel Core 5 Ultra).

[Mike Frysinger]

ChromiumOS is an OSS project.  you filed a public bug for the OSS project.  OSS support is best effort, not guaranteed.

for Flex, if you want to prioritize something, you need to do it via the Google partner support channels.  that is not issuetracker.

i.e. requesting urgent responses to issues requires a paid support commitment.  if you aren't paying for support, then on the OSS side, it's mostly "patches welcome".

-mike