Why I skipped the RTX 5090 and kept my RTX 4090

At a glance:

• Nvidia's RTX 5090 launched at $2,000, but street prices climbed above $4,000 within months.
• The RTX 4090 still receives meaningful driver updates, including DLSS 4 and DLSS 4.5 improvements.
• Multi-Frame Generation is the main RTX 50-series feature being missed, but DLSS Enabler can provide a workaround.

The upgrade that suddenly stopped making sense

When Nvidia launched the RTX 5090 last year, the RTX 4090 looked like a sensible trade-in candidate. At a $2,000 launch price, the RTX 5090 was already expensive, but that was the same amount paid for a liquid-cooled RTX 4090 in 2022. From that perspective, the upgrade did not immediately look absurd, even if the generational gains were not as dramatic as previous jumps from Nvidia's flagship cards.

The RTX 4090 remains a very capable GPU, but it is not invincible. Some newer AAA titles can push it hard once path tracing is enabled and settings are cranked to the maximum at 4K. Early benchmarks showed around a 30% FPS uplift for the RTX 5090, which was not transformative, but it was enough to make the upgrade feel worth planning for.

That calculation changed once real-world pricing became clear. By the time the money was saved, the RTX 5090 was going for over $4,000 in the market. At that point, the upgrade stopped looking like a premium GPU purchase and started looking like a second high-end PC, especially if the RTX 4090 was sold to offset part of the cost.

Street prices changed the value proposition

The RTX 5090's $2,000 MSRP was already $400 higher than the RTX 4090's launch price. For a card that did not deliver the same generational leap as its predecessor, that starting point was already difficult to justify. The situation became more complicated because AMD did not offer a true high-end competitor, leaving Nvidia with unusually strong leverage at the top of the market.

For someone who genuinely needed more 4K performance, the RTX 5090 was effectively the only option. The RTX 4090 was still the second-fastest GPU, meaning the RTX 5090 sat above it as the clear flagship. That lack of competition made the pricing frustration sharper: the upgrade was not merely expensive, it was expensive without a realistic alternative.

When vendors began charging over $4,000 for the same GPU within months, the MSRP almost looked reasonable by comparison. Paying twice the MSRP for a 30% frame-rate increase was not an enthusiast bargain; it was a premium that exceeded the value of the performance being gained. Even after selling the RTX 4090, the net cost would still be far higher than the original RTX 4090 purchase, which made the upgrade feel hard to defend.

DLSS made the old card feel newer

The decision to keep the RTX 4090 was not only about avoiding a bad price. It was also about the software support Nvidia has continued to deliver. Driver updates have brought major improvements to DLSS upscaling and frame generation, making the older GPU feel less obsolete than it might have after a new generation launch.

The author was not always a DLSS fan. After buying the RTX 3090, the first experience with DLSS in Cyberpunk 2077 was negative, with the image looking softer and less detailed than native rendering. After upgrading to the RTX 4090, DLSS looked better, but native 4K still seemed preferable because it preserved more detail.

DLSS 4, which launched alongside the RTX 50 series, changed that view. Upscaled 4K began to look almost as good as native 4K, and in some cases it could even look better than native resolution with TAA enabled. That made the RTX 4090 feel more future-proof, because the card was gaining image-quality improvements without a hardware purchase.

Nvidia's DLSS 4.5 update pushed that further earlier this year. The newer model presets improved upscaling quality enough that DLSS Balanced with Preset L became appealing for gaining extra FPS without a noticeable image-quality compromise. Frame generation also became more viable when base performance was already decent, especially in slower single-player titles where input feel matters less than in competitive games.

The RTX 5090's biggest missing feature

The RTX 4090's biggest hardware disadvantage compared with the RTX 5090 is Multi-Frame Generation. Nvidia's Dynamic Multi-Frame Generation can automatically adjust the number of generated frames based on the target frame rate, which makes the feature more useful in demanding AAA single-player games. The tradeoff is latency, and inputs can feel off when too many frames are generated, but the feature can still be worthwhile when raw smoothness matters more than perfect responsiveness.

That said, Multi-Frame Generation was not enough to justify a $4,000-plus upgrade on its own. The author experimented with Multi Frame Generation using DLSS Enabler, and while the FPS gains were impressive, the feature did not completely change the gaming experience. For the kind of single-player gaming described here, 60 to 100 FPS was already more than enough.

Cyberpunk 2077 and Assassin's Creed: Shadows were cited as examples of games where pushing beyond 200 FPS was not the goal. A 30% raw performance jump would have made higher frame rates easier to reach, but the price made that advantage feel less compelling. If a workaround exists for the one major missing feature, the incentive to upgrade becomes weaker.

Why patience may be the smarter upgrade path

The RTX 4090 was bought with the expectation that it would not last this long, but the four-year-old GPU has aged well. That is largely because Nvidia has continued to improve the software stack around it. In many GPU generations, older owners feel left behind quickly, but the RTX 4090 has remained close enough to the top through a mix of raw performance, DLSS improvements, and frame generation support.

Skipping the RTX 5090 does not mean the card is unimpressive. It means the combination of modest generational gains, high MSRP, and inflated street prices made the upgrade hard to recommend from a personal value perspective. If the RTX 5090 were closer to $2,000, the decision might have been different, especially for users who want the fastest possible 4K setup.

For now, the smarter move is to rely on AI upscaling and frame generation to extend the life of a GPU that remains the second-fastest on the planet. The next upgrade will need to deliver a bigger performance leap, a better price, or both. Until then, patience is not just a compromise; it is the most rational response to the current GPU market.

Related