Graphics card shoppers searching for “RTX 5070 Ti vs RX 9070 XT” in October 2026 are not chasing old news by accident. With flagship cards scarce and expensive (our earlier report found the GeForce RTX 5090 essentially gone from US retail shelves at prices north of $9,500), buyers are circling back to the mid-range tier that actually ships at something resembling a normal price. Nvidia’s GeForce RTX 5070 Ti and AMD’s Radeon RX 9070 XT, both launched in early 2025, have become the default “sane person’s upgrade” question of late 2026, and outlets including TechRadar and KitGuru have been refreshing their head-to-head coverage as a result.
The renewed interest isn’t happening in a vacuum. Average European GPU prices jumped 41.3% for Nvidia cards and 28.3% for AMD cards between June and September 2026, a surge we covered in detail here, driven by a DRAM and HBM shortage that has nothing to do with either chip’s underlying design. That same memory crunch is why the RTX 5070 Ti and RX 9070 XT, cards that debuted 19 months ago, are still the two most-discussed GPUs in their price class rather than having been cleanly replaced by a next generation. This piece breaks down the actual spec sheet, the pricing reality in 2026, what reviewers measured, and what it means for buyers trying to decide between the two before the holiday shopping season.
The 2026 context: why this matchup is trending again
Both cards are, strictly speaking, last year’s news. As TechRadar’s editorial team put it, “the Nvidia RTX 5070 Ti was released on February 20, 2025, with a starting price of $749”, built on the Blackwell architecture. AMD’s Radeon RX 9070 XT followed on March 6, 2025, at $599, running RDNA 4. Neither company has shipped a direct successor in this exact tier as of this writing, and that gap is precisely why the comparison keeps resurfacing. Desktop GPU shipment data cited in our coverage of Nvidia’s roughly 90% share of a 12.5 million-unit desktop GPU market shows just how dominant this generation remains in actual sell-through, more than a year after launch.
Valve’s own hardware telemetry backs that up from the player side. The most recent Steam Hardware Survey data, which we broke down in our report on 32GB RAM overtaking 16GB as the most common configuration, showed the RTX 5070 family leading the GPU rankings among Steam users, ahead of both higher-end and older cards. That’s a strong signal that this exact price bracket, not the $2,000-plus flagship tier, is where most real purchase decisions are actually happening in 2026. With memory costs squeezing every card’s margins and flagship inventory drying up, the RTX 5070 Ti and RX 9070 XT have effectively become the ceiling of what a typical PC gamer will spend, rather than a mid-tier stepping stone.
Specs at a glance: Blackwell vs. RDNA 4
On paper, the two cards are closely matched in memory capacity but diverge sharply in architecture and memory technology. Nvidia’s GeForce RTX 5070 Ti packs 8,960 CUDA cores and 70 fourth-generation RT cores onto its Blackwell die, paired with 16GB of GDDR7 across a 256-bit bus delivering 896 GB/s of bandwidth. AMD’s Radeon RX 9070 XT counters with 4,096 stream processors and 64 ray accelerators on RDNA 4, also with 16GB of memory on a 256-bit bus, but using the more mature GDDR6 standard for 640 GB/s of bandwidth.
| Spec | Nvidia GeForce RTX 5070 Ti | AMD Radeon RX 9070 XT |
|---|---|---|
| Architecture | Blackwell | RDNA 4 |
| Shader units | 8,960 CUDA cores | 4,096 stream processors |
| Ray-tracing hardware | 70 fourth-gen RT cores | 64 ray accelerators |
| Memory | 16GB GDDR7 | 16GB GDDR6 |
| Memory bus | 256-bit | 256-bit |
| Memory bandwidth | 896 GB/s | 640 GB/s |
| Board power | 300W | 304W (total board power) |
| Recommended PSU | 750W | 750W |
| Upscaling tech | DLSS 4 with Multi Frame Generation | FSR 4 |
| Launch date | February 20, 2025 | March 6, 2025 |
| Launch price | $749 | $599 |
That 256 GB/s bandwidth gap in Nvidia’s favor comes down to GDDR7 adoption, which is also one reason the RTX 5070 Ti carries a 25% higher launch price. AMD’s choice to stick with GDDR6 kept the Radeon RX 9070 XT’s bill of materials lower, and as the ongoing memory shortage has shown, that decision also made the card less exposed to the GDDR7 and HBM pricing spikes rattling the broader market this year.
Price then and now: from MSRP to 2026 street prices
The launch-day numbers are clean: $749 for the GeForce RTX 5070 Ti against $599 for the Radeon RX 9070 XT, a $150 gap. What’s murkier is what each card actually costs today. Some 2026 retail-tracking reports have pegged average street prices closer to $979 for the RTX 5070 Ti and $659 for the RX 9070 XT, but those figures are specific reported averages rather than confirmed, universal current prices, and they move by region, retailer, and week given the memory shortage distorting the entire GPU market.
| Metric | GeForce RTX 5070 Ti | Radeon RX 9070 XT |
|---|---|---|
| Launch price (2025) | $749 | $599 |
| Reported 2026 street average* | ~$979 | ~$659 |
| Implied change | ~+30.7% | ~+10.0% |
| Price gap vs. rival at launch | +$150 | Baseline |
| Price gap vs. rival, reported 2026 | +$320 (reported) | Baseline |
*Reported street averages are not official MSRPs and should be treated as directional, not exact. If accurate, the gap between the two cards has roughly doubled in dollar terms since launch, which tracks with the pattern we found when covering the 41% Nvidia and 28% AMD price surge in the European market: Nvidia’s GDDR7-equipped cards are more exposed to the current memory shortage than AMD’s GDDR6-based lineup, and that exposure shows up directly in how much each card’s street price has drifted from its original MSRP.
Ray tracing, DLSS 4, and FSR 4: the upscaling fight
Hardware specs only tell part of the story, because both companies lean heavily on software to extend performance. Nvidia’s GeForce RTX 5070 Ti supports DLSS 4 with Multi Frame Generation, which can generate multiple interpolated frames per rendered frame to boost displayed frame rates well beyond native rendering output. AMD’s Radeon RX 9070 XT runs FSR 4, AMD’s current-generation upscaler, which closed much of the image-quality gap with DLSS that plagued earlier FSR versions.
Tom’s Hardware’s 2026 test suite found the GeForce RTX 5070 Ti had superior ray-tracing performance overall, which lines up with Nvidia’s two-generation head start in dedicated RT hardware and the maturity of its denoising and upscaling stack. The same outlet described the Radeon RX 9070 XT as “way cheaper” than the RTX 5070 Ti at the time of its report, framing the choice as a tradeoff between raw ray-tracing headroom and dollars saved rather than a clean win for either card. That framing still holds in late 2026: nobody buying a Radeon RX 9070 XT today is doing it because they expect it to beat Nvidia at path tracing, they’re doing it to save money while still getting competitive performance in the far more common rasterized workloads that make up most of what people actually play.
Real-world benchmarks: what reviewers found
Independent reviews back up that rasterization-versus-ray-tracing split. KitGuru’s Dominic Moass tested the Radeon RX 9070 XT directly against the GeForce RTX 5070 Ti and reported that “at 1440p it averages 101fps, putting it bang in line with the 4080 Super, while also making it 6% faster than the RTX 5070 Ti”. That 6% edge was measured in rasterized gaming performance, the workload that still dominates most people’s actual play time, and it’s a meaningful data point given the RX 9070 XT costs $150 less at launch price.
Moass’s broader verdict on the Radeon RX 9070 XT was unambiguous on that front: “Overall then, there’s no doubt the RX 9070 XT delivers impressive rasterisation performance”. On the Nvidia side, IGN’s Matt Smith positioned the RTX 5070 Ti as the stronger all-around 4K option, writing that “at $749, the Nvidia GeForce RTX 5070 Ti is the best 4K graphics card for most people, delivering much better value than either the RTX 5080 or 5090”. IGN’s Adam Patrick Murray summed up the rivalry from AMD’s side just as directly: “The AMD Radeon RX 9070 XT is a $599 graphics card that trades blows with the $749 GeForce RTX 5070 Ti”. Taken together, the reviewer consensus is less a knockout and more a genuine split: AMD wins on rasterized frame rates and price, Nvidia wins on ray tracing, upscaling maturity, and 4K headroom.
Power draw, PSU headroom, and thermals
Power requirements are nearly a wash. Nvidia rates the GeForce RTX 5070 Ti at 300W board power, while AMD rates the Radeon RX 9070 XT at 304W total board power, a 4W difference that won’t move anyone’s electricity bill. Both companies recommend the same 750W power supply, which means PSU shopping isn’t a deciding factor between the two cards the way it sometimes is between a mid-range and flagship part.
That near-identical power envelope matters more than it looks, because it means the performance and efficiency differences between the two cards come almost entirely from architecture and memory choice, not from one vendor simply feeding its GPU more electricity. Buyers upgrading from an older, higher-wattage card in either camp should see case thermals improve rather than worsen, since both the RTX 5070 Ti and RX 9070 XT sit in a power band that’s comfortable for most mid-tower cases with decent airflow.
Market share and shipment data: who’s actually buying
Specs and reviews are only half the picture; the market has already rendered its verdict in unit terms. The desktop GPU shipment figures we reported, showing Nvidia holding roughly 90% of a 12.5 million-unit desktop GPU market, make clear that AMD is fighting from a structural minority position even when it wins individual benchmark categories. That imbalance makes AMD’s pricing strategy on the Radeon RX 9070 XT less about maximizing margin and more about defending whatever share it can hold in the one tier where it has a genuinely competitive product.
Steam’s hardware telemetry tells a complementary story about where demand is actually concentrated. Our analysis of the latest Steam Hardware Survey data showing the RTX 5070 family leading GPU adoption suggests that most PC gamers are making exactly the kind of mid-range purchase decision this comparison addresses, rather than chasing a flagship. That’s a function of both price sensitivity and the practical reality that flagship cards are either unavailable or absurdly priced right now, pushing demand down into the tier where the RTX 5070 Ti and RX 9070 XT compete most directly.
Historical context: a decade of Nvidia-AMD mid-range battles
This isn’t the first time Nvidia and AMD have fought it out in the roughly $600-to-$750 bracket, and it won’t be the last. Across multiple GPU generations, the pattern has repeated: Nvidia ships a card with a meaningful ray-tracing and upscaling advantage at a premium price, while AMD answers with a cheaper card that wins on raw rasterized frame rates and value, betting that most buyers care more about frames-per-dollar in the games they already own than about path-traced lighting in the handful of titles that showcase it best.
What’s different in 2026 is the backdrop. Previous generations of this rivalry played out against a relatively stable memory market, where price cuts and refreshes arrived on a predictable cadence. This time, the fight is happening inside a supply shock. The same DRAM and HBM shortage squeezing AI data center buyers has reached straight into consumer GPU pricing, and it’s hitting Nvidia’s GDDR7-based cards harder than AMD’s GDDR6-based ones simply because of which memory standard each company picked for this generation, not because of anything either company did wrong on the silicon itself.
Competitive landscape: where else the budget could go
Buyers weighing the RTX 5070 Ti against the RX 9070 XT also have to consider what else their money could buy in a distorted 2026 market. Stepping up to a GeForce RTX 5090 is largely off the table at anything resembling a sane price, given the card’s near-total disappearance from US retail at prices above $9,500. Grey-market workarounds exist at the extreme end too: we reported on a China-based mod shop offering a 96GB VRAM version of the RTX 5090 for $3,888, but that’s a niche, unofficial path aimed at AI workloads rather than a realistic option for a mainstream gamer comparing two $600-$750 cards.
Waiting for next-generation hardware isn’t a clean alternative either. Leaked roadmap chatter around Nvidia’s next flagship tier has been inconsistent, and we covered one prominent leaker reversing an earlier 2027 timeline call to 2028 for next-gen high-end parts. With that kind of uncertainty hanging over the top of the stack, and no confirmed direct successor to either the RTX 5070 Ti or RX 9070 XT, buying now in this exact tier looks like the more rational move for anyone who needs a GPU in the next few months rather than a hypothetical one over a year out.
Market impact: what this means for retailers and AIB partners
For add-in-board partners and retailers, the renewed relevance of this specific comparison is a mixed blessing. On one hand, sustained demand for 19-month-old cards means existing inventory and manufacturing lines stay useful longer, reducing pressure to rush a refresh before the memory market stabilizes. On the other hand, rising input costs on the memory side squeeze margins on cards that were priced under a very different cost structure back in early 2025, forcing partners to either eat the difference or quietly raise prices on cards whose MSRP was set over a year ago.
That tension is visible in how differently the two companies’ ecosystems have responded. AMD’s partners have been relatively transparent about staged price increases tied to memory costs, while Nvidia’s pricing moves have been less openly communicated, according to the same tracking data behind our report on Europe’s 41% Nvidia and 28% AMD price surge. For shoppers, the practical upshot is that the RX 9070 XT’s price has stayed comparatively more stable than the RTX 5070 Ti’s, widening the value gap between the two cards well beyond their $150 launch-day difference.
Predictions: where this rivalry goes next
Based on current pricing trends, shipment data, and the state of the memory market, a few outcomes look likely over the next two to three quarters. First, expect the dollar gap between the RTX 5070 Ti and RX 9070 XT to keep widening rather than closing, since Nvidia’s GDDR7 exposure to the memory shortage shows no sign of easing before 2028 on most industry timelines. Second, AMD likely gains incremental share in the sub-$700 bracket specifically, not because RDNA 4 suddenly pulls ahead technically, but because its GDDR6-based bill of materials is simply less exposed to the current supply shock.
Third, FSR 4 adoption across more game titles should keep narrowing the upscaling gap with DLSS 4, even if Nvidia retains its ray-tracing hardware lead. Fourth, with next-generation flagship timelines reportedly sliding toward 2028 rather than 2027, don’t expect either company to retire the RTX 5070 Ti or RX 9070 XT from active sale anytime soon. Fifth, retail availability for both cards will likely stay choppier than usual through the holiday quarter, as the same scarcity dynamics pushing flagship buyers down-market also thin out mid-range inventory at a faster clip than retailers are used to planning for.
Frequently asked questions
Is the RTX 5070 Ti or RX 9070 XT better for gaming in 2026?
It depends on the workload. KitGuru’s testing found the Radeon RX 9070 XT about 6% faster in 1440p rasterized gaming, while the GeForce RTX 5070 Ti holds a clear edge in ray-traced titles and 4K upscaling thanks to DLSS 4 with Multi Frame Generation.
How much did each card cost at launch?
The GeForce RTX 5070 Ti launched at $749 on February 20, 2025. The Radeon RX 9070 XT launched at $599 on March 6, 2025, a $150 difference.
Why are GPU prices higher in 2026 than at launch?
A global DRAM, NAND, and HBM memory shortage has pushed component costs up across the board. Nvidia’s GDDR7-based cards appear more exposed to that shortage than AMD’s GDDR6-based cards, which is reflected in reported 2026 street-price averages running higher for the RTX 5070 Ti than for the RX 9070 XT relative to their launch prices.
What power supply do I need for either card?
Both Nvidia and AMD recommend a 750W power supply for the GeForce RTX 5070 Ti and Radeon RX 9070 XT, respectively. Board power draw is nearly identical: 300W for the RTX 5070 Ti and 304W total board power for the RX 9070 XT.
How much memory does each card have?
Both cards ship with 16GB of memory on a 256-bit bus. The RTX 5070 Ti uses GDDR7 for 896 GB/s of bandwidth, while the RX 9070 XT uses GDDR6 for 640 GB/s.
Should I wait for the next generation instead of buying now?
That’s a personal call, but leaked timelines for next-generation flagship GPUs have reportedly slipped toward 2028, and neither company has announced a direct successor to this exact tier. For anyone who needs a GPU within the next few months, waiting carries real opportunity cost given how uncertain those next-gen timelines currently look.
Does the RX 9070 XT support DLSS?
No. DLSS is Nvidia-exclusive technology. The Radeon RX 9070 XT uses AMD’s own FSR 4 upscaler, which narrowed much of the image-quality gap with DLSS compared to earlier FSR generations, though Nvidia’s hardware ray-tracing lead remains the bigger differentiator.
Which card has better long-term value?
Reported 2026 pricing data suggests the Radeon RX 9070 XT has held its value closer to launch price than the GeForce RTX 5070 Ti, widening its price advantage beyond the original $150 gap, though exact street prices vary significantly by region and retailer.
