Nvidia shipped a “fix” that breaks more PCs — avoid driver 595.71 and roll back now

Stop updating: the driver that was supposed to fix a disaster looks like a second one

Nvidia rolled out GeForce driver 595.71 on March 2 as a hotfix for the botched 595.59 build – but early reports show it’s creating fresh problems. Users are reporting system crashes, Blue Screens of Death, HDR/refresh monitor issues and aggressive artificial voltage/clock limits that shave performance, particularly on RTX 50/40-series cards. Nvidia has not issued a public retraction or guidance; the safest play right now is to avoid 595.71 entirely and roll back if you’ve already installed it.

Key takeaways

• 595.71 was released March 2 to address issues from the pulled 595.59 driver but is producing new stability and performance problems.
• Reports include BSODs, crashes, monitor refresh/HDR failures and an enforced ~3,000MHz/1.0V cap that cuts OC’d RTX 5090 performance by as much as ~16% in some benchmarks.
• Nvidia remains silent while the driver remains available – unlike the quick pull of 595.59 on Feb 27 – so users should avoid installing and roll back via Windows Device Manager if affected.
• Scope seems variable: stock cards sometimes okay; manual overclocks and RTX 50-series users hit hardest. Independent benchmarks and deeper tests are still pending.

What actually changed — and why it matters

595.59 was already toxic: Nvidia pulled it on Feb 27 after reports of fan control failures and aggressive voltage capping on RTX 50/40 cards. 595.71 was supposed to be the cleanup. Instead, multiple tech outlets and community threads say 595.71 now enforces artificial voltage/clock limits (reports point to roughly a 1.0V/~3,000MHz cap on some RTX 5090 samples, down from >1.05V and >3,100MHz), which translates into measurable FPS losses in OC scenarios. Add to that fresh postings of BSODs and monitor refresh/HDR problems, and you no longer have a narrow OC issue — you have a general stability problem for many users.

Put plainly: a driver update that should restore normal operation is instead adding new failure modes. For gamers, that matters for two reasons. First, the immediate pain: crashes and lost frames. Second, the longer-term signal: repeated post-release driver breakage suggests QA is fraying at a company that’s simultaneously prioritizing AI/data-center growth — something Nvidia’s CEO has openly discussed in investor forums.

The uncomfortable observation Nvidia doesn’t want shouted in a press release

Nvidia is the market leader for gaming GPUs, but it’s also the engine behind the AI boom. That twin focus matters. Public comments from Nvidia’s leadership in recent conferences make clear the company is channeling resources and preferential supply toward data-center customers. The uncomfortable truth: consumer-driver QA looks like an afterthought while the business chases massive, time-sensitive AI deployments. The result is a pattern — driver hotfixes shipped too fast, pulled too late, and users left to clean up the mess.

The question Nvidia hasn’t been asked (but should)

If I had ten seconds with a PR rep I’d ask: why is a fix that was supposed to restore fan and voltage behavior now enforcing lower voltage/clock limits and introducing hard stability problems for a wide range of cards? Gamers deserve to know whether this is a rushed band-aid, a telemetry-driven decision to protect silicon, or a deeper regression in driver testing.

How to protect your PC right now

• Do not install 595.71 if you haven’t already. Wait for independent lab results (Gamers Nexus, Digital Foundry) or an official Nvidia hotfix.
• If you installed it and see crashes or lost performance: roll back the driver via Windows Device Manager — Display adapters → right-click your NVIDIA device → Properties → Driver tab → Roll Back Driver. That reliably restores pre-595.x behavior for many users.
• Disable automatic driver installs in GeForce Experience or Windows Update until this is resolved.

What to watch next

• An official Nvidia statement or a pulled 595.71 build — that would be the clearest sign they acknowledge the regressions.
• Independent benchmark and stability analyses from Gamers Nexus and Digital Foundry; look for quantified FPS and crash-rate comparisons to 591.xx baselines.
• Forum/Reddit threads tracking BSOD frequency across GPU generations — if reports expand beyond OC/R50-series cards, this moves from niche to systemic.

My betting — based on the pattern of quick patches and pulls we’ve seen over the last two years — is Nvidia will issue a hotfix once independent labs confirm the regressions. But until then, the best consumer strategy is simple: don’t be the early adopter of this one.

TL;DR

595.71 was meant to fix the disaster of 595.59 but is introducing new crashes, BSODs and artificial voltage/clock limits that cut performance. Nvidia is still silent and the driver remains available — avoid installing and roll back via Device Manager if you have it. Watch for an Nvidia retraction or independent lab tests; those will tell you whether this is a bad hotfix or a symptom of a bigger QA problem.