Is an RTX 5060 Ti 16GB worth $780 for machine learning, and are unfamiliar discount retailers safe?

Established retailers are generally the safer choice for expensive components. Amazon can still have problems with third-party sellers or returned parts being resold, so check who actually ships and sells the item and inspect it immediately. If Best Buy is available, it may be a safer option and can sometimes match a listing sold and shipped directly by Amazon.

Before buying, compare the 5060 Ti against used cards such as the RTX 3080 Ti and newer cards in the same budget. A used 3080 Ti can deliver much stronger raw gaming performance, but it has different power, heat, memory-capacity, and warranty considerations. For ML, make sure the specific software supports the card and that 16GB is enough for the models you plan to run.

$780 is an extremely high price for a 5060 Ti 16GB. Its extra VRAM can make it useful for AI and machine-learning workloads, but the card is not particularly compelling at that price. For gaming, an older high-end NVIDIA card or a comparable AMD card may offer substantially more performance for the money, though CUDA and software compatibility can make NVIDIA preferable for ML.

I’d be very cautious about an unfamiliar store offering a graphics card for $200 less than the normal market price. A deal that far below typical pricing can mean a counterfeit listing, an empty shipment, a used or damaged replacement, or a difficult refund process. Unless the seller has a strong independent reputation and clear buyer protections, the savings probably aren’t worth the risk.