Nvidia's RTX 3080 Ti Lands as Computex Turns Into a Chip-Shortage Confessional

Computex is usually where the GPU hype trains leave the station, and today Nvidia delivered right on schedule: the GeForce RTX 3080 Ti is official, priced at $1,199, slotting in just under the 3090 as the new “I have a problem” tier for anyone still trying to build a gaming rig in 2021. The RTX 3070 Ti follows next week on June 10 at $599, rounding out Nvidia’s mid-refresh of the Ampere lineup.

Here’s the thing nobody can quite get past, though: none of this matters much if you can’t actually buy one. Availability has been the real story of this GPU generation since launch, and a new SKU at a higher price point doesn’t fix that on its own. I’d love to be proven wrong here, but “paper launch” is going to be the phrase hovering over this release until stock proves otherwise.

AMD’s quieter, more interesting reveal

While Nvidia grabbed the gaming headlines, AMD used its own Computex keynote to show something that feels more consequential long-term: 3D V-Cache. The pitch is stacking extra cache directly on top of future CPU dies using through-silicon-via bonding, which in theory lets AMD dramatically boost cache capacity without blowing up the physical footprint of the chip. No shipping product yet, just a technology demo, but if it pans out this could be a meaningful lever for gaming and compute performance on future Ryzen and EPYC parts alike.

AMD also announced new EPYC “Milan” server chips at the same event, continuing the steady cadence of Zen 3-based server silicon that’s been chipping away at Intel’s data center dominance for the past couple of years. Milan isn’t a surprise architecturally, it’s the expected iteration, but the fact that AMD keeps hitting its server roadmap on schedule is itself notable given how much of the industry is currently struggling with supply.

Intel plays a different game

Intel’s presence at Computex this year felt less about new silicon and more about messaging. The company leaned hard into its plans to expand domestic chip manufacturing, which reads as much like a policy and investor pitch as a product announcement. Given how central “who actually fabs the chips” has become to the current conversation, it’s a smart place for Intel to plant a flag, especially while it’s not the one setting the pace on raw GPU or CPU performance this cycle.

And that brings us to the elephant in every one of these keynotes: the chip shortage. Nvidia, AMD, and Intel all acknowledged, in one form or another, that the constrained supply situation isn’t wrapping up anytime soon and will likely persist through the rest of the year. That’s not exactly a shock to anyone who’s tried to buy a GPU, a console, or even a car in the last several months, but it’s a rare moment of alignment between three companies that usually spend their keynotes trying to one-up each other.

So where does that leave a buyer? Honestly, in the same spot as before: excited about the announcements, skeptical about the shelves. The 3080 Ti is a genuinely strong card on paper. Whether “on paper” is as close as most of us get to it remains the open question.