On July 15, FinalWire released AIDA64 beta version 8.30.8337, adding GPU hot-spot temperature measurement for NVIDIA Blackwell—the architecture powering GeForce RTX 50-series cards. The update gives Windows users another officially maintained utility to see a sensor that NVIDIA has not documented through its public NVAPI. PC enthusiasts who’ve been hunting for peak GPU die temperatures on their RTX 50 now have a straightforward, supported path, joining recent additions in HWiNFO and HWMonitor.

What Actually Changed

The AIDA64 beta changelog lists “GPU Hotspot temperature measurement on NVIDIA Blackwell” as the key new capability. In practice, that means the application can now read the highest temperature from one of several internal sensors sprinkled across the GPU die. Previously, users could only pull the edge temperature—the chip’s overall thermal reading—which often masks localised hotspots that can trigger throttling long before the headline number suggests trouble.

VideoCardz confirmed that AIDA64 follows CPUID’s HWMonitor and HWiNFO, both of which rolled out Blackwell hot-spot support in recent builds. The approach isn’t a flip of a switch. Because NVIDIA’s approved NVAPI interface omits hot-spot and VRAM temperature queries for RTX 50, these monitoring developers have had to poke at private interfaces or read registers directly—a technique that carries risk of breakage with future driver updates. Still, for anyone running the beta, the new reading shows up automatically; no configuration is required beyond launching the program.

Rival tool MSI Afterburner remains without native support. Developer Alexey “Unwinder” Nikolaychuk has long said that Afterburner relies on the supported NVAPI path, so until NVIDIA changes its stance, the utility cannot officially expose Blackwell hot-spot or VRAM temperatures. The workaround? A community developer going by Talon2016 has built an Afterburner plug-in, BlackwellHotspot.dll, which VideoCardz reports exposes six discrete temperature values plus a “Hotspot Delta” calculation. Igor’sLAB separately released its own diagnostic, igor’sLAB Blackwell Hotspot Estimation (IBHE), which blends accessible telemetry with model-based estimates.

What It Means for You

For most home users and enthusiasts, the arrival of hot-spot monitoring delivers a more accurate picture of GPU thermal behaviour. A typical scenario: you run a benchmark and see GPU core at 72°C, well within limits, yet the card momentarily drops clock speeds. A hot-spot reading of 95°C or higher explains the throttling instantly, pointing you toward fan-curve adjustments, paste reapplication, or case-flow improvements.

Power users and overclockers gain a datapoint they’ve had in previous GPU generations. RTX 30 and 40 cards exposed hot-spot temperatures through NVAPI, but the RTX 50 launch broke that tradition. With AIDA64, HWiNFO, or HWMonitor, under-volting and overclocking stability testing can once again watch for temperature spikes that the core sensor alone would hide. The IBHE tool even offers a delta value—the spread between the lowest and highest reported sensor—which can highlight uneven paste application or mounting-pressure issues more clearly than a single number.

System administrators and IT professionals managing small render farms or CAD workstations should take note of the unofficial nature of the data. These readings come from reverse-engineering, not a ratified NVIDIA specification. Different tools may report slightly different numbers. If you’re using hardware monitoring for warranty claims or fleet-wide thermal audits, treat hot-spot figures as investigative clues, not certified sensor values. Always cross-reference with the GPU’s own throttling flags and performance logs.

Developers and hobbyist tool authors face a different landscape. The plug-in for MSI Afterburner highlights that the telemetry is accessible—if you’re willing to download unsigned DLLs from a forum. That is a significant caveat: a third-party DLL that reads GPU registers could, in theory, be malicious or clash with anti-cheat engines. VideoCardz notes that the mod’s author warns it may stop working after any driver or Afterburner update. Anyone considering this route should scan the file, keep a system restore point, and be prepared for the plug-in to become obsolete overnight.

How We Got Here

NVIDIA has been tightening telemetry access for several releases. With the GeForce RTX 40 series, the company removed voltage-reading capabilities that third-party tools had long relied on, pushing a “Power Management” narrative around user safety. The RTX 50 series continued that trend: the standard NVAPI does not expose memory junction temperatures or hot-spot readings, despite the sensors being physically present on the die.

This shifted the burden to the monitoring ecosystem. HWiNFO author Martin Malik was among the first to crack the barrier, adding Blackwell hot-spot support in late June 2026 through direct register access. HWMonitor followed shortly after. FinalWire’s AIDA64—a commercial suite with a strong reputation in enterprise diagnostics—waited until its beta channel could extensively test the feature, landing on July 15, 2026.

What’s notable is the speed at which community developers filled the Afterburner void. Within weeks of the first hot-spot reveals, Talon2016’s plug-in appeared on forums, suggesting that the sensor addresses are being shared rapidly among the hardware tinkering crowd. Igor Wallossek’s IBHE tool took a different approach: it reads what it can with administrator privileges and then applies a model to estimate the six hot-spot values. As Wallossek has cautioned, those numbers still require validation, and the industry has not yet settled on a canonical hot-spot reading for Blackwell.

NVIDIA has not responded to VideoCardz’s request for comment on whether future drivers will block the currently used access methods or provide an official interface. The silence means uncertainty: a driver update tomorrow could wipe out all unofficial hot-spot monitoring, or the feature could stay indefinitely at the cat-and-mouse stage.

What to Do Now

If you own an RTX 50 card and want to see your GPU’s hot-spot temperature today, you have several layered options:

Use a stable, well-known monitoring tool. Install the latest beta of AIDA64 (v8.30.8337), or update HWiNFO or HWMonitor to their most recent versions. All three will display the hot-spot reading alongside the standard GPU temperature. This is the safest, lowest-friction path. AIDA64 is commercial but offers a trial; HWiNFO and HWMonitor are free for personal use.

Verify the reading. Hot-spot values fluctuate quickly. Load the GPU with a consistent workload—a 10-minute run of FurMark or a demanding game loop—and watch both the edge and hot-spot temperatures. A normal delta (hot-spot minus edge) is typically 10–20°C under load. If you see a delta exceeding 25°C or a hot-spot that hits the thermal limit (commonly around 105°C for Blackwell) while the edge temp remains in the 70s, begin investigating cooling.

Diagnose with IBHE for deeper insight. Igor’sLAB’s free tool attempts to resolve six individual hot-spot sensors. Run it as administrator and compare its report against the single hot-spot value from AIDA64 or HWiNFO. If one specific sensor repeatedly spikes while the others stay cool, you may have a paste void or uneven mount. However, remember that the numbers are partly model-based and not a ground-truth readout.

Avoid the Afterburner plug-in unless you have a specific need. The community DLL is tempting, especially if Afterburner is already your overlay of choice. But the risk-to-utility ratio is poor for most users. If you do test it, download from the original forum post, check the file hash against what the author publishes, and run a malware scan. Do not use it in environments where system integrity is critical, such as esports tournaments with invasive anti-cheat or on computers handling sensitive data.

Watch for driver updates. Since the hot-spot sensors are accessed through non-public channels, a new GeForce driver could neutralise them. Set a reminder to re-check your monitoring tool after any driver installation. If the reading disappears, visit the tool’s forum or support page; developers will likely scramble to restore access quickly.

Don’t panic over a single high reading. Hot-spot sensors are sensitive to momentary workload spikes. A brief jump to 100°C that instantly drops back to 80°C does not indicate a failed cooler. Look for sustained high deltas or throttling events recorded in GPU-Z or the NVIDIA driver’s performance counters.

Outlook

NVIDIA’s move to hide hot-spot data from NVAPI frustrates a vocal segment of the PC community, but it aligns with broader trends toward deeper driver integration and less user-level telemetry exposure. History suggests that when community tools find a crack, NVIDIA either patches it silently or eventually relents and provides a sanctioned readout—as happened with certain power sensors in the past. For now, Windows users have the information they need, albeit coming through side doors that require a bit of diligence. Keep AIDA64, HWiNFO, or HWMonitor updated, treat the numbers as investigative clues rather than gospel, and stay alert for changes in the next driver release. The hot-spot is back, but it’s still on thin ice.