If you're staring at your aging GPU and wondering when does rtx 5070 come out, you aren't alone. It's been a weird year for PC gaming. Honestly, the timeline has been a bit of a moving target. While the massive RTX 5090 and its smaller sibling, the 5080, already hit the shelves back in January 2025, the mid-range has felt a little... neglected.
The short answer? The base RTX 5070 officially launched on March 5, 2025.
But that’s only half the story. If you’re looking for the newer "Super" refresh or just trying to find one at the original $549 MSRP, things get complicated fast. We are currently in early 2026, and the "Super" variants that everyone expected to see at CES 2026 earlier this month were a total no-show.
The RTX 5070 Release Timeline (And Why It’s Shifting)
Nvidia likes to follow a pattern, but they also love to pivot when they’re making billions from AI chips. The initial Blackwell rollout was relatively fast. We saw the big enthusiast cards in January 2025, followed by the RTX 5070 Ti in February 2025, and then finally the standard RTX 5070 in March 2025.
Since then, the market has been a roller coaster. If you’re checking stock today, you’ve probably noticed that prices aren’t exactly sticking to that $549 launch price. Between a global shortage of GDDR7 memory and Nvidia focusing heavily on data center production, the 5070 has become surprisingly hard to find at a fair price.
Why the "Super" Refresh is MIA
Everyone—and I mean everyone—expected Nvidia to announce the RTX 5070 Super at CES 2026. Rumors from reliable leakers like kopite7kimi suggested we’d see an 18GB monster that would basically fix the VRAM complaints people had at launch.
Instead, CES 2026 came and went with a focus on DLSS 4.5 and some wild RTX 5090 Lightning Z models from partners like MSI. No Super cards. Word from the grapevine (and reported by outlets like PCMag) is that Nvidia has "indefinitely delayed" the Super refresh.
Basically, they don’t have to rush. AMD’s RDNA 4 lineup, specifically the Radeon RX 9070 XT, hasn't exerted enough pressure to force Nvidia’s hand. Why launch a better, cheaper card when the current one is still selling out?
Specs That Actually Matter (No Fluff)
If you’re still hunting for the original 5070, here is what you’re actually getting for your money. It’s a 12GB card. Yeah, 12GB in 2026 feels a bit tight for some, but the move to GDDR7 memory makes a massive difference in bandwidth.
- CUDA Cores: 6,144
- Memory: 12GB GDDR7 (28 Gbps)
- Bandwidth: 672 GB/s
- Power Draw: 250W TGP
- Interface: PCIe 5.0 x16
The performance is actually impressive. It’s roughly twice as fast as the old RTX 4070 in path-traced games like Cyberpunk 2077 and Black Myth: Wukong. But that "2x" number is heavily carried by DLSS 4.0 and 4.5. Raw rasterization? It’s a solid jump, but not quite a miracle.
The VRAM Dilemma
The biggest point of contention remains that 12GB buffer. In a world where Resident Evil Requiem is pushing VRAM limits, 12GB feels like it might have a shorter shelf life than we’d like. This is exactly why so many people were holding out for the rumored 18GB Super version.
Current leaks suggest that if the Super does ever come out, it will use 3GB memory modules to hit that 18GB sweet spot. But with memory prices expected to climb another 40% by the middle of 2026, don’t expect it to be cheap.
🔗 Read more: How to Search for a Word on a Mac Without Losing Your Mind
Real-World Pricing in 2026
Forget the $549 MSRP for a second. That was a lovely dream from March 2025. Today, because of the "AI tax" on memory, you’re more likely to see partner cards from ASUS, Gigabyte, and PNY sitting between **$599 and $649**.
If you find one for $549, buy it. Seriously.
The supply chain is currently strangled. Reports from Tom’s Hardware have highlighted that German and Japanese retailers are even rationing 5070 stock because the memory chips are being diverted to AI accelerators. It’s a "gamers vs. data centers" world right now, and we’re kinda losing.
Should You Wait for the RTX 5070 Super?
This is the $600 question. Honestly? I wouldn't wait anymore.
Since the CES 2026 snub, the most optimistic timeline for an RTX 5070 Super launch has slipped to late Q2 or even Q3 of 2026. Some insiders are even whispering that it might be canceled entirely in favor of an early RTX 60-series (Rubin architecture) teaser in 2027.
If you need a card today for 1440p gaming at high refresh rates, the 5070 is still the best balance of power and efficiency on the market. Just be prepared to pay a bit of a premium.
📖 Related: When Did MySpace Come Out? The Real Story of the Site That Changed Everything
Actionable Next Steps
If you’re ready to pull the trigger or just want to stay ahead of the curve, here’s how to play it:
- Monitor GDDR7 Prices: Follow industry news on memory supply. If GDDR7 prices continue to spike, GPU prices will follow.
- Look for the "Ti": The RTX 5070 Ti is already out (launched Feb 2025) and offers 16GB of VRAM. It’s more expensive at $749, but it solves the longevity issue if you plan to keep the card for 4+ years.
- Check Open-Box at Microcenter/Best Buy: Now that the card has been out for nearly a year, "buyer's remorse" returns are becoming more common. You can often snag a 5070 for under $500 this way.
- Skip the Hype: Don't wait for a "Super" that might never come. If you see a card that fits your budget and plays your favorite games at the frames you want, take the win.
The window for "early adoption" has closed, and we’re now in the "stable but expensive" phase of the Blackwell lifecycle. Keep an eye on local stock, avoid the scalpers on eBay, and prioritize cards with decent cooling—the 250W TGP means these things can run a bit hot in smaller cases.