RTX 5090 and M4 MacBook Air: Can It Game?

TL;DR

A user demonstrated that it is technically possible to connect an NVIDIA RTX 5090 GPU to an M4 MacBook Air using Thunderbolt eGPU docks. While the setup shows promise, practical use for gaming or AI inference remains limited due to driver and software support issues.

A user has demonstrated that it is technically possible to connect an NVIDIA RTX 5090 graphics card to an M4 MacBook Air via Thunderbolt eGPU docks, raising questions about gaming performance and practical use on Apple Silicon Macs.

The setup involves connecting a high-end desktop GPU, the RTX 5090, through a Thunderbolt dock that adapts PCIe to Thunderbolt, and then into the MacBook Air. This process relies on PCIe tunneling over Thunderbolt 4, which provides up to 4 lanes at 40Gbps, enabling the Mac to recognize the GPU as a PCIe device. However, macOS does not natively support NVIDIA GPUs on Apple Silicon, and drivers are limited. A recent development includes tinygrad releasing open-source drivers for macOS, but these are primarily suited for AI inference and not gaming. Linux can run on Apple Silicon Macs and support NVIDIA GPUs via PCI pass-through in virtual machines, but current kernel support for Thunderbolt on Apple Silicon is lacking. As a result, practical use for gaming or AI tasks remains constrained, with performance and stability issues reported during testing. The demonstration is mostly at a proof-of-concept stage, highlighting technical feasibility but not yet practical deployment.

Why It Matters

This development is significant because it suggests that high-performance desktop GPUs could potentially be used with MacBook Airs, expanding the hardware capabilities of Apple Silicon Macs beyond official support. For gamers and AI developers, this could open new avenues, but current limitations mean it is not yet a practical solution. The experiment underscores ongoing challenges in driver support and hardware compatibility, which are critical hurdles for wider adoption.

Plugable Thunderbolt 4 Dock for M4/M5 MacBook Air & Pro: 100W Charging Driverless Laptop Docking Station with 4K HDMI Dual Monitor or Single 6K, 6X USB, SD, Ethernet, Thunderbolt Certified (TBT-UDM)

Plugable Thunderbolt 4 Dock for M4/M5 MacBook Air & Pro: 100W Charging Driverless Laptop Docking Station with 4K HDMI Dual Monitor or Single 6K, 6X USB, SD, Ethernet, Thunderbolt Certified (TBT-UDM)

13-in-1 Thunderbolt Performance: Need a high-performance Mac docking station? The Plugable TBT-UDM laptop docking station is a space…

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Background

Apple Silicon Macs lack native support for discrete NVIDIA or AMD GPUs, limiting their gaming and AI capabilities. Enthusiasts have explored eGPU setups via Thunderbolt, which tunneling PCIe over Thunderbolt 4 makes technically feasible. Linux support for NVIDIA GPUs on ARM64 is improving, but driver support remains limited. Recent efforts like tinygrad’s open-source drivers aim to bridge this gap but are not yet suitable for demanding tasks like gaming. The demonstration of connecting an RTX 5090 via PCIe passthrough on a MacBook Air is a recent example of pushing hardware boundaries, though practical use faces significant software and stability challenges.

“Connecting a desktop GPU like the RTX 5090 to a MacBook Air is technically possible through Thunderbolt eGPU docks, but practical use is still limited.”

— tech enthusiast

“Our open-source drivers are not optimized for gaming or AI inference on Mac, but they demonstrate progress in driver support.”

— tinygrad developer

ASUS ROG Astral NVIDIA GeForce RTX 5090 32GB GDDR7 OC Edition Gaming Graphics Card (PCIe 5.0, HDMI/DP 2.1, 3.8-Slot, 4-Fan Design, Axial-tech Fans, Patented Vapor Chamber), 3 Year Warranty

ASUS ROG Astral NVIDIA GeForce RTX 5090 32GB GDDR7 OC Edition Gaming Graphics Card (PCIe 5.0, HDMI/DP 2.1, 3.8-Slot, 4-Fan Design, Axial-tech Fans, Patented Vapor Chamber), 3 Year Warranty

Powered by the NVIDIA Blackwell architecture and DLSS 4

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

What Remains Unclear

It remains unclear whether this setup will become stable enough for regular gaming or AI inference use. Driver support on macOS for NVIDIA GPUs is limited, and current workarounds like tinygrad are experimental. Performance benchmarks and long-term stability are still unknown, and hardware compatibility may vary with different GPU models and Mac configurations.

ANQUORA ANQ-L336 Mini eGPU Enclosure 40Gbps Thunderbolt3/4 USB4 for Laptop Compatible with Win/Linux/MacOS/ROG Ally X/Claw 8, PCIe External GPU Dock with PD 85W, Support NVIDIA/AMD, DC/ATX/SFX/Flex

ANQUORA ANQ-L336 Mini eGPU Enclosure 40Gbps Thunderbolt3/4 USB4 for Laptop Compatible with Win/Linux/MacOS/ROG Ally X/Claw 8, PCIe External GPU Dock with PD 85W, Support NVIDIA/AMD, DC/ATX/SFX/Flex

Excellent Performance: With officially certified JHL7440 controller module, showing high-performance on gaming, AI training/inference, 3D modeling & rendering,…

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

What’s Next

Next steps include refining driver support, improving stability, and conducting performance testing under real-world gaming and AI workloads. Future updates to macOS or third-party driver projects could enhance usability, but widespread adoption is still uncertain. Monitoring progress in driver development and user reports will be key to assessing practical viability.

AOOSTAR AG01 External GPU Docking Station, Supports Mainstream NVIDIA and AMD Graphics Cards, Built-in 800W Power Supply, Oculink Expansion Graphics

【AG01 EGPU Dock Features】 The AOOSTAR AG01 external GPU docking station has a built-in Huntkey 800W power supply….

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Key Questions

Can I use an RTX 5090 with my MacBook Air for gaming?

Currently, while the hardware connection is possible, practical gaming use is limited due to driver and software support issues. Performance and stability are still being tested.

What are the main challenges in using external GPUs on Apple Silicon Macs?

The primary challenges include lack of native driver support for NVIDIA GPUs on macOS, limited Thunderbolt support for PCIe passthrough, and stability issues during hardware configuration.

Will this setup improve AI inference or machine learning tasks on MacBook Air?

While theoretically possible, current software support like tinygrad is too limited for demanding AI inference, and performance is significantly lower than native solutions.

No, connecting high-end desktop GPUs to MacBook Air via eGPU is experimental and not supported officially by Apple.

You May Also Like

Led by industry, Japan and Taiwan plant seeds of drone cooperation

Japan and Taiwan are beginning cooperative efforts in drone technology amid efforts to lessen reliance on China’s supply chain, signaling strategic shifts.

Apple Silicon costs more than OpenRouter

Recent analysis shows Apple Silicon costs more per token for local AI inference than OpenRouter, raising questions about cost-efficiency for on-device models.

AI tool for radiotherapy can support the global effort to eliminate cervical cancer

A new AI tool effectively plans radiotherapy for cervical cancer, potentially expanding access and saving lives in low-income countries, according to recent trial results.

Forza Horizon 6 boots up in just 4 seconds instead of 90 with new Advanced Shader Delivery tech and AMD GPUs — Microsoft claims 95% reduction in gaming load times

Microsoft’s new Advanced Shader Delivery reduces Forza Horizon 6 load times from 90 seconds to just 4 seconds on supported GPUs, marking a major performance breakthrough.