spaceSpace and Physics

Nuclear Fusion On The Brink Of A Breakthrough Thanks To Google Algorithm


Jonathan O'Callaghan

Senior Staff Writer

The Norman machine. TAE

Google has revealed it is helping make fusion power a reality, by joining forces with a private fusion company to create machine learning algorithms that could lead to a breakthrough.

The company Google has been working with since 2014 is called Tri Alpha Energy (TAE), which has reportedly raised a huge $500 million so far. TAE, based in California, has been investigating how to store plasmas for a prolonged period of time, with first their C2-U machine and now a new machine called Norman.


Google’s involvement was detailed in a new study published in Scientific Reports. The two companies have been working on something called the Optometrist Algorithm, which seeks to help engineers by showing them what works and doesn’t work in fusion reactors.

“Just like getting an eyeglass prescription, the algorithm presents the expert human with machine settings and the associated outcomes,” Ted Baltz, from the Google Accelerated Science Team, said in a blog post. “The key improvement we provided was a technique to search the high-dimensional space of machine parameters efficiently.”

What does this all mean exactly? Well, the idea behind nuclear fusion is to store a hydrogen plasma for long enough to get more energy out than was put in, known as breakeven. Once this happens, fusion will be the real deal and provide a revolutionary zero-carbon form of energy.

This is the previous C2-U machine. TAE

At the moment, a number of experimental reactors around the world have succeeded in storing a plasma, but for barely more than a minute. The problem is that as the plasmas heat up, they get harder to store.


TAE’s machine is a bit different, in that it uses a field-reversed configuration that gets more stable as the energy increases. In the latest breakthrough, they managed to get a 50 percent reduction in the energy loss rate of the total plasma energy, a key step towards keeping a plasma contained and stable.

“Results like this might take years to solve without the power of advanced computation to rapidly scale our understanding of the complex properties of plasma,” said Michl Binderbauer, TAE’s President and Chief Technology Officer, in a statement. “This research, years in the making, has already allowed us to advance our science, and I am grateful for the longstanding collaboration with Google.”

With their new Norman machine, they're going to investigate how plasmas in the machine behave over the next 18 months. And with Google’s help, they’re hoping to find out if the field-reversed configuration technique is the way forward. If it is, they’ll build a machine that can actually generate power. Fingers crossed.


spaceSpace and Physics
  • tag
  • google,

  • nuclear fusion,

  • plasma,

  • breakthrough,

  • TAE