Folding@home in 2026: What It Is, How to Install v8, and Does It Still Help?

Folding@home is a free volunteer computing project that uses the spare power of home computers to simulate how proteins move and fold, research that helps scientists understand diseases such as Alzheimer's, cancer, Parkinson's and COVID-19. It is still active in 2026: the current software is the v8 client (version 8.5 at the time of writing), for Windows, macOS and Linux. To join, download the client from foldingathome.org, install it, open the web control at app.foldingathome.org, choose which CPU and GPU to use, and select Fold. It does not pay you, it does use electricity, and it is safe when downloaded from the official site. Your graphics card does most of the useful work, and you can stop or uninstall it any time.

In April 2020, when this page was first written, Folding@home was in the news every day. Millions of people joined to help with COVID-19 research, and the network briefly became the most powerful computer system on Earth. Ethan was one of them: he installed it on his gaming PC that spring, left it running for months, then forgot about it. This autumn his PC fans started roaring at night, and he asked Jake whether the old program was still doing anything, whether it was still worth it, and whether it was safe to leave. Jake found the old v7 client still installed, replaced it with the current v8 client, set it to fold only while Ethan was away from the PC, and explained what his GPU had actually been contributing. This page covers the same ground: what Folding@home is, what it has achieved, how to install and set up v8, what it costs you, and every common problem.

What is Folding@home, and what does it do?

Proteins are the small machines inside every living cell. They are long chains that fold into precise shapes, and the shape decides what the protein does. When proteins fold wrongly, or when a virus's proteins change shape to attack a cell, disease can follow. Alzheimer's, Parkinson's, Huntington's, many cancers and infectious diseases all involve proteins behaving in ways scientists want to understand.

The trouble is that proteins move. They wiggle, open, close and change shape constantly, and the movements that matter for disease can be rare and fleeting. Simulating that motion atom by atom takes enormous computing power, far more than most labs can afford. Folding@home's answer, since it started at Stanford University in 2000, is to split the work into small pieces called work units and send them to volunteers' computers around the world. Each computer simulates a short stretch of a protein's motion, sends the result back, and gets another piece. Put together, millions of short simulations reveal how a protein behaves over much longer times.

So "what does Folding@home do" has a simple answer: it turns your idle computer into a small part of a very large scientific instrument. Researchers use the results to find hidden pockets in proteins where drugs might attach, understand how mutations change a protein, and test ideas before running expensive lab experiments. The project is run today by a consortium of academic laboratories, and its results are published in scientific journals and shared openly.

How does Folding@home work on your side? You install a small program called the client. It downloads a work unit, runs a scientific program called a core on your CPU or GPU, uploads the result when it is done, and repeats. You see progress, estimated time and points in a web page. Everything runs in the background, at low priority, so other programs come first.

The research causes, in plain English

The client lets you choose a cause. Here is what the main research areas are about, so you can choose with some idea of what your PC is doing.

• Alzheimer's disease: proteins such as amyloid-beta and tau clump together in the brain. Simulations explore how those clumps start, which could point to ways to prevent them.
• Parkinson's disease: a protein called alpha-synuclein misfolds and builds up in nerve cells. Understanding its shapes helps the search for treatments.
• Huntington's disease: an inherited change in the huntingtin protein makes it fold badly. Simulations study why.
• Cancer: many cancers involve proteins that send growth signals getting stuck in the "on" position. Simulations look for places where drugs could switch them off.
• Infectious disease: viral proteins, including those of SARS-CoV-2, and bacterial enzymes that make bacteria resistant to antibiotics. Hidden pockets in these proteins are possible drug targets.

Choosing "any" sends you whatever work is most needed, which keeps your hardware busy most reliably. Choosing a cause is a fine way to make the contribution personal; Ethan picked Alzheimer's, for a family reason.

Whichever you pick, the work units your PC receives are real simulations from active research projects, and each one is checked when it comes back before it is used.

Does Folding@home actually help? What it has accomplished

"Does Folding@home actually help?" and "does Folding@home do anything?" are fair questions for a project that asks for your electricity. The honest answer is yes, with a realistic sense of how science works. Folding@home does not cure diseases by itself. It produces data that researchers use, and its results feed into a long chain of lab work, drug design and clinical trials that takes years.

• Hundreds of scientific papers. Over more than two decades, Folding@home data has supported a large body of published research on protein folding, misfolding and drug binding.
• The first exascale computer. In March and April 2020, the flood of volunteers pushed Folding@home past one exaflop, a billion billion calculations per second, before any single supercomputer reached that level. It was the first computing system of any kind to do so.
• COVID-19 research. Simulations of the SARS-CoV-2 virus's proteins revealed hidden "cryptic pockets" that could be targeted by drugs, and Folding@home computing supported the open-science COVID Moonshot effort to design antiviral drug candidates.
• Methods other scientists reuse. Techniques developed to analyze Folding@home's huge datasets are now used in other areas of computational biology.

One thing worth knowing: your individual contribution is small, and that is the design. A single PC's work unit is one piece of a puzzle with millions of pieces. The value comes from many people contributing steadily. Ethan's GPU completed several hundred work units over the 2020 lockdown, each a real simulation used in real research, and that is about as much as anyone's home PC can do.

The 2020 surge: when gamers built the world's fastest computer

Before 2020, Folding@home was a respected but niche project, popular with PC enthusiasts and science fans. Then, in early 2020, it announced that it would simulate the proteins of the new coronavirus. Graphics card makers, game companies, tech sites and Microsoft all encouraged people to donate their idle computers, and gamers with powerful GPUs signed up in huge numbers. This page was first written in that moment, as a short news post about Microsoft's call to contribute.

The result was extraordinary. Within weeks the network passed one exaflop, then grew to more than twice that, faster than the top supercomputers of the time combined. For a while the project's servers could not hand out work units fast enough, which is why so many people remember "no work available" messages from that spring. The team added servers, partner companies donated cloud capacity, and the work flowed again.

As lockdowns ended, many volunteers left, and the network shrank back toward its earlier size, still large, and still doing research. The lasting effects were real: the COVID-19 datasets produced that year are among the largest protein simulations ever run, and the project's software was modernized, leading to the v8 client used today.

Is Folding@home still active in 2026?

Yes. Folding@home is still running, and the software is actively developed. The current client is version 8, which replaced the long-lived v7 client. Version 8.5 is the current public release as of 2026, with updates through the year, and the project's forum is busy with users and developers. If you are asking "what happened to Folding@home?" because the news coverage stopped after 2020, nothing happened to it; it simply went back to steady, quieter work after the pandemic peak.

Is Folding@home still relevant now that AI predicts protein structures? Yes, because the two do different jobs, explained in the AlphaFold section below. Is it still useful? The project continues to send out work for active research, and participation is lower than in 2020, so each volunteer's share of the work is larger than it was then.

To see which version you have, open the web control; the client version is shown with each machine. If you still have the old v7 client, with its separate FAHControl window, the v7 vs v8 section below explains how to move across without losing your name or points. New releases of the v8 client appear on the download page; check it every few months, because newer clients fix bugs and support newer GPUs.

How to install Folding@home on Windows (v8)

Installing takes about five minutes. These steps are for Windows 11 and 10; macOS and Linux are similar, and notes for them follow.

1. Download the client from the official site, foldingathome.org, using its Start Folding or download page. Choose the Windows installer. Download it only from the official site; copies elsewhere may be out of date or tampered with.
2. Run the installer. Accept the license and the default options. The client installs and starts running in the background, with an icon in the notification area near the clock.
3. Open the web control. The installer opens it for you, or go to app.foldingathome.org in your browser. The page connects to the client running on your own computer.
4. Choose what to fold with. In the settings, pick which resources to use: how many CPU threads, and which GPUs. In v8 there are no more "slots" as in v7; you just enable the CPUs and GPUs you want, and the client shares work between them.
5. Optionally pick a cause such as cancer or Alzheimer's, or leave it on any disease so you get work whenever it is available.
6. Optionally set a team by entering a team number, so your points count toward a group. Team 0 is the default for people without one.
7. Select Fold. The client downloads a work unit and starts. You see progress, the estimated time to finish, and your estimated points per day.

To check what it is using, open Task Manager and watch the CPU and GPU graphs on the Performance page; our guide to opening Task Manager covers the quick ways in. The client runs at low priority, so games and other apps take what they need first, but a GPU busy with folding will still run warmer and louder.

Mac and Linux

On macOS, download the Mac installer from the same page; the client supports Apple Silicon and Intel Macs. On Linux, the site offers packages for common distributions, such as a .deb file for Ubuntu and Debian, and the same web control at app.foldingathome.org works in your browser. For servers and home labs, community-maintained Docker images exist, so you can run the client in a container; use a well-known image and keep it updated.

Folding@home settings worth changing

The defaults work, but a few settings make Folding@home fit your PC and your life better. Here is what each one does and a sensible starting choice.

Outside the client, two more settings matter: your GPU's power limit in its driver software, which is the most effective way to cut electricity, and Windows' sleep settings, which decide whether folding continues when you leave the PC. Change them once and the client looks after the rest.

Folding@home account, login, passkey and teams

You can fold without an account. An account is useful if you run Folding@home on more than one computer, because it lets you see and control all of them from one web page, from any browser, including your phone.

• Folding@home login: sign in on app.foldingathome.org to create or use an account, then link each machine to it. On computers without a screen, the account token can be added to the client's configuration file, which is how people set up headless servers and many machines at once.
• Name: the name your points are recorded under in the public statistics. Pick one and keep it.
• Passkey: a private code tied to your name. It proves that work under that name is yours and is needed for the bonus points paid for returning work quickly. Keep it private.
• Team: a number for the group you fold with. Many tech sites, companies, universities and online communities have teams. Your points count toward both you and the team.

Ethan joined a team run by a gaming community he belongs to, which turned out to be the most motivating part. Watching the team climb the leaderboard did more for his folding than any setting.

Folding on several machines, headless servers and Docker

Many people fold on more than one computer: an old desktop in a closet, a home server, or a few spare PCs. The v8 client makes this easier. Install the client on each machine and link it to your account; on a machine without a screen, add the account token and a machine name to the client's configuration file before the first start, which links it automatically. You then manage every machine from one browser tab.

For home labs, community Docker images package the client in a container. CPU folding works in a plain container; GPU folding in a container needs the GPU passed through, which for NVIDIA cards means installing NVIDIA's container toolkit on the host. Keep containers updated like any other software, and use images from well-known maintainers.

Points, PPD and Folding@home stats explained

Folding@home awards points for completed work units. Points do not have money value; they are a way to measure and compare contribution. A few terms you will see:

Your stats, and your team's, appear on the project's public statistics pages and leaderboards after work units are counted, which can take a while. If your stats are not showing, check that the name in your client matches exactly, including capitals, and that the passkey is entered correctly.

Why do GPUs earn far more points than CPUs? Because the simulations Folding@home runs are well suited to the thousands of small cores in a graphics card. A modern mid-range GPU can earn many times the points of an entire desktop CPU, which is also a reasonable guide to how much science each contributes.

CPU vs GPU, and AMD vs NVIDIA

Does Folding@home use CPU or GPU? Both. You choose which. Here is how to decide.

• Use the GPU if you have a dedicated graphics card. NVIDIA and AMD graphics cards are supported. A GPU contributes far more per watt of electricity than a CPU for this kind of work.
• NVIDIA vs AMD: both work. NVIDIA cards have generally been the most efficient for Folding@home's GPU cores, but recent AMD cards also do well. The best card is the one you already own; buying a GPU just to fold rarely makes sense.
• Use the CPU sparingly. CPU folding helps, especially for projects that need it, but it heats a laptop quickly and contributes less per watt. On a desktop with a GPU, leaving a few CPU threads free keeps the PC responsive.
• Integrated graphics in laptops and small PCs are usually not worth using; they contribute little and run the machine hot.
• Keep drivers updated. GPU folding depends on the graphics driver. If folding fails or your screen flickers and recovers, update the driver from NVIDIA or AMD; our guide to the display driver stopped responding error covers that symptom.

Folding@home and the PlayStation 3: from 2007 to 2012, PS3 owners could fold through a built-in app, and PS3s made up a large share of the network for a while. That app ended in 2012, and consoles are no longer part of the project.

How to tell if your folding is working well

A few quick checks show whether your hardware is folding efficiently or wasting electricity:

• GPU use near full while folding. In Task Manager's Performance page, the GPU should show high load. If it is low, the GPU may be waiting for the CPU.
• One CPU thread free for each GPU. GPU folding needs a CPU thread to feed the graphics card with data. If every CPU thread is also folding, the GPU slows down. Leaving one thread per GPU free usually raises total output.
• Temperatures in a safe range. Use your GPU maker's software or a hardware monitor. If temperatures sit near the card's limit, clean the dust, improve airflow, or lower the power limit.
• Steady PPD. Points per day varies with each work unit, but a sudden large drop on the same hardware can mean throttling from heat or a driver problem.
• Few failed work units. The client log shows results. Repeated failures usually point to an unstable overclock or an old driver.

What Folding@home costs you: electricity, heat and wear

Folding@home is free to download, but not free to run. It uses your electricity, and it makes your computer work harder. Here is a realistic way to estimate the cost, using round numbers you can swap for your own.

1. Find the power draw while folding. A mid-range desktop GPU folding at full speed might draw around 150 to 250 watts, plus the rest of the PC. A plug-in power meter gives you the real number.
2. Multiply by the hours. A PC drawing 250 watts for 24 hours uses 6 kWh a day.
3. Multiply by your electricity price. At $0.17 per kWh, 6 kWh is about $1.02 a day, or about $31 a month. Running only overnight, 8 hours, cuts that to about $10 a month.

Other costs to think about: heat in the room, which is welcome in winter and not in summer; fan noise; and some extra wear on fans. Modern components are built to run under load, so folding does not meaningfully shorten a well-cooled PC's life, but a dusty case or a laptop on a bed will run hot. On a laptop, fold only while plugged in, on a hard surface, and consider using the GPU or CPU, not both.

Ways to keep the cost sensible: fold only while the PC is idle or overnight, lower the power level in the client, or cap your GPU's power limit in its driver software, which cuts electricity much more than it cuts work done. Graphics cards run least efficiently at the very top of their power range, so trimming the last 20 to 30 percent of power usually costs only a small share of the points, while the card runs cooler and quieter. If you want folding to continue overnight, make sure Windows does not put the PC to sleep; our guide to sleep settings in Windows 11 and 10 shows where to set that, and a balanced power plan keeps the rest of the PC efficient.

Ethan's setup now folds only when he is away from the PC and overnight, with the GPU power limit lowered by about a quarter. It uses noticeably less electricity than his 2020 setup, and still completes several work units a day.

Folding@home on a laptop: what to know

You can fold on a laptop, and many people do, but laptops need more care than desktops. Their cooling is designed for short bursts of heavy work, not hours of full load, and their batteries do not like heat.

1. Fold only while plugged in. Folding on battery drains it quickly and adds charge cycles.
2. Use a hard, flat surface. A bed, sofa or lap blocks the air vents underneath and traps heat. A laptop stand helps more than any software setting.
3. Use one resource, not both. On a gaming laptop, fold with the GPU and leave the CPU alone, or the other way round on a laptop without a dedicated GPU, and use fewer CPU threads.
4. Fold only when idle. Set the client so it works only while you are away, so the laptop is cool and quiet while you use it.
5. Watch the first session. Check temperatures and fan noise during the first hour. If the laptop gets uncomfortably hot to touch, reduce the load or stop.

Thin laptops with integrated graphics contribute very little and run hot doing it; for them, Folding@home is rarely worth the wear. A laptop that runs hot even when you are not folding may have a different problem; our guide to laptops running hot in the bag covers Modern Standby and other causes.

Is Folding@home safe and legit?

Yes. Folding@home is a long-running, legitimate academic project, open about its research and its software. The client downloads work units and scientific cores from the project's own servers, runs them, and sends back results. It does not mine cryptocurrency, it does not ask for personal information to fold, and you can fold anonymously.

The usual safety rules apply. Download the client only from foldingathome.org. Be wary of "Folding@home" downloads on other sites, browser extensions or apps that claim to pay you for folding. And remember the hardware side: a PC that overheats or a laptop folding on a blanket is a real risk, so keep vents clear and watch temperatures the first time you run it. Our safer internet guide covers safe download habits in general.

Does Folding@home pay?

No. Folding@home is volunteer computing; you donate computer time, and you earn points and a place on leaderboards, not money. A few unrelated third-party projects have offered cryptocurrency tokens based on Folding@home points, but they are not part of Folding@home and carry their own risks. If anything promises to pay you to fold, treat it with suspicion.

Folding@home not working: common problems and fixes

• "No work" or "no appropriate assignment". The servers have no work units for your hardware right now. It is common, especially for CPU-only machines and older GPUs, and for specific causes. Leave the client running; it retries automatically. Choosing "any disease" gives you the best chance of getting work.
• HTTP_SERVICE_UNAVAILABLE or connection errors. A work or collection server is busy or down. This is on the project's side and usually clears on its own. Check your firewall or security software is not blocking the client if it never connects.
• GPU not detected. Update the graphics driver from NVIDIA or AMD, restart, and check that the GPU is enabled in the client's settings.
• Work units fail repeatedly on the GPU. Usually an unstable overclock or an old driver. Return the GPU to default clocks and update the driver.
• Web control shows no machines. The web page cannot find the client. Make sure the client is running (check the notification area), and that the web control version matches; current web control works only with v8.3 or newer clients, so update an old client.
• Stats not showing. Name or passkey mismatch, or the stats have not updated yet. Give it a few hours, then check spelling.
• PC slow while folding. Reduce CPU threads, fold only when idle, or pause folding while you play or work.

If none of these fit, the official Folding@home forum is the best place to ask; volunteers and developers answer questions there, and searching it for the exact error text usually finds someone who solved it already. The Folding@home community on Reddit is friendly for general questions, team recruiting and hardware advice. When you ask, include your client version, operating system, CPU and GPU models and driver version, and the lines from the client log around the error; that turns a long thread into a quick answer.

Folding@home v7 vs v8: should you upgrade?

If you installed Folding@home in 2020, like Ethan, you probably have the v7 client, called FAHClient with the FAHControl window. Version 8 is the current client, and moving to it is worth doing.

To upgrade, uninstall v7 first from Settings, Apps, then install v8 and enter the same name, team and passkey so your points continue under your name. Let any work unit in progress finish first if you can, so it is not wasted.

Folding@home vs AlphaFold, BOINC and other projects

Folding@home vs AlphaFold: AlphaFold, from Google DeepMind, is an AI system that predicts the folded shape of a protein from its sequence, and it transformed structural biology from 2020 onward. It mostly gives a still picture of a protein's shape. Folding@home simulates how proteins move, including rare shapes and hidden pockets that a single still picture can miss. The two complement each other: AI predictions often give researchers a starting structure, and simulations then explore how that structure behaves. So "Folding@home vs AI" is not a contest; AI did not make simulation obsolete.

BOINC vs Folding@home: BOINC is a platform from the University of California, Berkeley, that runs many separate volunteer projects, including Rosetta@home, which designs and predicts proteins, and World Community Grid, which supports health and climate research. Folding@home uses its own client and is not part of BOINC. You can run both, but share your hardware sensibly between them.

Other ways to use a GPU: if you would rather use your graphics card for your own projects, such as running AI models locally, our guide to hardware for local AI covers what a GPU can do there. Folding@home and local AI can share a PC; just pause folding while you use the GPU yourself.

Is Folding@home worth it for you?

Whether Folding@home is worth it is a personal choice, but a few questions settle it for most people.

• Do you have a dedicated GPU that sits idle? If yes, it is worth trying. That is where most of the useful work comes from.
• Do you pay for electricity, and is it expensive where you live? If yes, fold overnight or idle-only with a lower power limit, and work out the monthly cost with the method above.
• Is your PC in a bedroom or small office? Heat and fan noise matter. Idle-only folding during the day, when you are out, may suit you better than overnight.
• Do you care about a particular disease? Choosing that cause can make the contribution feel more personal.
• Is it a work computer? Ask first. Many employers do not allow volunteer computing on their hardware or electricity.

If the answers point to yes, install it, join a team, and check in now and then. If they point to no, that is fine too; donating money to medical research or simply keeping your electricity bill down are perfectly good choices. Ethan's answer was yes, with limits: idle-only, a lower power limit, and his gaming community's team number in the settings.

How to pause, stop or uninstall Folding@home

To pause, open the web control and select Pause, or use the icon in the notification area. Paused work resumes where it stopped. To finish the current work unit and then stop, use the finish option so the work is not wasted.

To uninstall Folding@home on Windows, open Settings, Apps, Installed apps, find Folding@home, and select Uninstall. If an old v7 copy is also listed, remove it too. Our guide to uninstalling programs in Windows 11 and 10 covers programs that leave pieces behind. On a Mac, use the uninstaller included with the client; on Linux, remove the package with your package manager.

Ethan's roaring fans turned out to be the forgotten v7 client grabbing every CPU thread whenever the PC was on. With v8, idle-only folding and a lower GPU power limit, his PC is quiet when he uses it and still folding when he does not.

Folding@home is quieter than it was in 2020, but it never stopped, and the science it supports goes on. If you have a graphics card that sits idle while you sleep, a few minutes of setup turns that time into research data, at a cost you can control. Ethan's PC is back to being quiet when he uses it, and his gaming team's folding total keeps climbing.

Revision note. Written October 3, 2026, turning our April 2020 news post, written when Microsoft and many others encouraged people to donate computing power for COVID-19 research, into a full guide to Folding@home today. It now covers what the project does and has achieved, the v8 client, installation, accounts and teams, points, hardware, electricity cost, safety, common problems, v7 vs v8, and how Folding@home compares with AlphaFold and BOINC. Electricity figures are examples; use your own wattage and rate. Happy folding.