The NVIDIA GeForce RTX 5080 launched with a $999 original price. That number is worth writing down, because almost nobody pays it anymore. As of August 2026, the same card routinely sells between $1,390 and $2,100 depending on the model — and the most expensive RTX 5080 variants now cost more than the RTX 5090’s original MSRP did at launch.
This guide breaks down what the RTX 5080 actually cost at launch, what it costs today, why the gap got so wide, and whether buying one right now makes any sense.
RTX 5080 Original Price at a Glance
| Original MSRP (Founders Edition) | $999 |
| Release date | January 30, 2025 |
| Typical new price (Aug 2026) | $1,390 – $1,800 |
| Premium AIB models (Aug 2026) | Up to $2,099 |
| Typical used price (Aug 2026) | ~$1,249 |
| Increase over MSRP | ~40% to 110% |
What the RTX 5080 Cost at Launch
NVIDIA announced the RTX 5080 at CES 2025 and put it on sale on January 30, 2025 at a Founders Edition MSRP of $999. That was a deliberate signal: the previous-generation RTX 4080 had launched at $1,199 and was widely panned for it, so NVIDIA cut $200 off the sticker for the Blackwell replacement.
Board partners — ASUS, MSI, Gigabyte, Zotac, PNY and the rest — were free to price above that. At launch, third-party cards clustered between $999 and roughly $1,300, with the overclocked flagship models (ROG Astral, SUPRIM, AORUS Master) sitting at the top of that band. The $999 figure applied to the Founders Edition and a handful of baseline partner cards.
Availability at that price lasted weeks, not months. By mid-2025 the $999 RTX 5080 was already a stock-alert lottery rather than a shelf price.
RTX 5080 Specs: What $999 Was Supposed to Buy
| Specification | RTX 5080 |
|---|---|
| GPU die | GB203-400-A1 (full die) |
| Architecture | Blackwell |
| Process node | TSMC 4N |
| CUDA cores | 10,752 (84 SMs) |
| Memory | 16 GB GDDR7 |
| Memory bus | 256-bit |
| Memory bandwidth | 960 GB/s |
| Total board power | 360W |
| Power connector | Single 12V-2×6 (up to 600W) |
| Ray tracing / Tensor | 4th gen RT, 5th gen Tensor |
The headline upgrade over the 4080 Super was GDDR7 memory, which pushed bandwidth to 960 GB/s on the same 256-bit bus. The 16 GB capacity was the sticking point even in 2025, and it has aged worse than the raw compute has — a point worth weighing before paying 2026 prices.
What the RTX 5080 Actually Costs in August 2026
Here is where the “original price” and the real price diverge sharply.
| Model | Price (Aug 2026) | vs $999 MSRP |
|---|---|---|
| PNY RTX 5080 | $1,390 | +39% |
| Amazon average, new | $1,478 | +48% |
| MSI SUPRIM SOC | $1,800 | +80% |
| ASUS ROG Astral | $2,099 | +110% |
| Used, eBay average | $1,249 | +25% |
The ASUS ROG Astral figure is the one that has drawn the most attention. At $2,099 it sits above the $1,999 original MSRP of the RTX 5090 — the generation’s flagship. A step-down card now lists for more than the halo card’s launch sticker.
European pricing followed the same curve. RTX 5080 cards that sat at €1,089–€1,190 during the May–June 2026 low have climbed to €1,269–€1,387, an increase of roughly 12% to 16.5% in a single quarter.
Why RTX 5080 Prices Went Up Instead of Down
GPUs normally get cheaper as a generation matures. Blackwell did the opposite, for four compounding reasons.
1. GDDR7 and DRAM costs exploded
GDDR7 modules have been running around $70 each. The RTX 5080 carries 16 GB of it. That is a meaningful chunk of bill-of-materials cost that did not exist at these levels when the $999 price was set, and it flows straight through to the shelf.
2. NVIDIA pulled back pricing subsidies
Retail sources point to the withdrawal of NVIDIA’s pricing-support program for board partners. While that program was active it kept baseline models near MSRP. Without it, partners price to their actual costs. NVIDIA has not publicly confirmed every detail of this, so treat it as well-sourced retailer reporting rather than official policy.
3. AI took the wafers
Data-center Blackwell parts carry far higher margins than gaming cards. When foundry capacity is finite, the allocation decision is not close. Gaming GPU supply gets squeezed, and squeezed supply against steady demand only moves one way.
4. No new competition arrived
NVIDIA has not launched a new gaming card in 2026. With no successor and no serious pressure at this tier, there is nothing forcing prices down. A product drought is a price floor.
US–China tariffs on Asian-assembled components add a further layer on top of all of this for American buyers.
Is the RTX 5080 Worth Buying at 2026 Prices?
The honest answer depends entirely on what you pay.
- At or near $999: If you find a genuine MSRP card in stock, buy it. That is the price the hardware was benchmarked and reviewed against, and it still holds up.
- $1,250–$1,400: Defensible if you need the card now for 1440p or 4K gaming, or for a workload that benefits from the memory bandwidth. You are overpaying, but not absurdly.
- $1,800–$2,100: Hard to justify. At that level you are within reach of used RTX 5090 listings, which bring 32 GB of VRAM — and 16 GB is the RTX 5080’s weakest specification.
- Used at ~$1,249: Currently the most sensible entry point, provided you verify the card was not used for sustained mining or AI training and you inspect the 12V-2×6 connector for heat damage before paying.
If you can wait, waiting is reasonable. Memory pricing is cyclical, and the current elevated GDDR7 costs are not a permanent structural feature of the market.
RTX 5080 vs the Alternatives Right Now
Because the RTX 5080 no longer occupies the price slot it was designed for, the competitive picture has shifted. At $1,400 it is fighting cards it was never meant to fight.
| Option | Approx. Aug 2026 price | Case for it |
|---|---|---|
| RTX 5080 (new, baseline) | $1,390–$1,478 | Best raw performance per dollar in the tier, if you can find a baseline SKU |
| RTX 5080 (used) | ~$1,249 | The value pick, with the usual used-GPU risk |
| RTX 5090 (used) | Well above $2,000 | 32 GB VRAM solves the 5080’s main weakness — but the premium is steep |
| Previous-gen 4080 Super | Varies, thin supply | Cheaper entry, but GDDR6X bandwidth and an aging feature set |
| Wait it out | — | Memory pricing is cyclical; the current spike is not permanent |
The pattern to notice is that every rung of the ladder has moved up together. This is not an RTX 5080 problem — it is a market-wide repricing. The RTX 5060 Ti tells the same story more starkly: the MSI VENTUS 2X OC PLUS went from €282 to €409 in the same window, a 45% jump on a mid-range card. When the budget tier rises 45%, “just buy something cheaper” stops being useful advice.
The practical takeaway: set a price alert rather than refreshing retailer pages. Baseline models do surface at or near MSRP during restocks, and those windows close in minutes. Knowing the $999 anchor is what lets you recognize one when it appears.
Frequently Asked Questions
What was the RTX 5080’s original price?
$999 for the Founders Edition at launch on January 30, 2025. Partner cards ranged from $999 to about $1,300.
Why is the RTX 5080 more expensive now than at launch?
Four factors stacked: GDDR7 memory costs near $70 per module, the end of NVIDIA’s board-partner pricing subsidies, foundry capacity diverted to higher-margin AI products, and no new competing gaming GPU launch in 2026 to apply downward pressure.
Can you still buy an RTX 5080 for $999?
Occasionally, through restock drops at Newegg and similar retailers. It is not a standing shelf price. Treat an MSRP listing as a deal to act on quickly, not as the baseline.
Is the RTX 5080 more expensive than the RTX 5090 was?
Some models, yes. The ASUS ROG Astral RTX 5080 at $2,099 exceeds the RTX 5090’s $1,999 launch MSRP. Current RTX 5090 pricing is far higher still, with some models above $4,300.
How much VRAM does the RTX 5080 have?
16 GB of GDDR7 on a 256-bit bus, delivering 960 GB/s of bandwidth. This is the card’s most-criticized specification and the main reason to consider a 5090 if your budget stretches.
The Bottom Line
The RTX 5080’s original price was $999 — a genuinely competitive number when it was set in January 2025. Eighteen months later, that figure has become a historical footnote rather than a shopping guide. Real-world pricing sits 40% to 110% above it, driven by memory costs, subsidy withdrawal, AI-driven supply constraints, and an absence of competition.
Know the $999 number so you recognize a real deal when one appears. Just don’t shop as though it is still the going rate.