Connect with us


Helium: A Valuable Gas Not To Be Taken Lightly



Helium makes up 25% of the atoms in the known universe, so one would guess that the inert gas would be quite plentiful on Earth.

Unfortunately, a familiar property of helium prevents this from happening. Helium gas is lighter than air and literally rises into space, depleting the Earth of almost all valuable helium resources over time.

Where do we get helium?

So how do we actually obtain new helium gas, which is necessary for important technological applications such as MRI machines, superconductors, and even the Large Hadron Collider?

Today’s infographic from Helium One shows everything you need to know on helium, including where we can find it on Earth, as well as the most important uses of the gas.

Helium: A Valuable Gas Not To Be Taken Lightly

Although helium is plentiful in the universe, on Earth it is quite rare and difficult to obtain.

Why Do We Need Helium?

Helium has several properties that make it invaluable to modern humans, particularly for technological uses:

Helium PropertyBenefits
InertDoesn’t react with other elements, and doesn’t explode like hydrogen
Non-toxicCan be used by humans in a variety of applications
Lighter than airAbility to lift and/or float
Melting point -272˚CLiquid at ultra-cool temps enables superconductivity
Small molecular sizeCan be used to find the smallest of leaks

Helium Demand

Helium demand has risen consistently since 2009, and the market has been increasing at a CAGR of 10.1% since 2010. With that in mind, here are the specific constituents of helium demand today:

Helium UseGlobal ShareDescription
Cryogenics23%Superconductors use ultracooled helium liquid.
Lifting15%Used in airships and balloons
Electronics14%Used to manufacture silicon wafers
Optical Fiber11%Necessary to make optical fiber cables
Welding9%Used as a shielding gas for welding
Leak Detection6%Helium particles are small, and can find the tiniest leaks
Analytics6%Used in chromatography and other applications
Pressure & Purging5%Used in rocket systems
Diving3%Mixed into commercial diving tanks for various reasons
Other8%Helium's diverse properties give it many other minor uses

Helium’s melting point, which is the lowest found in nature, allows it to remain as a liquid at the coolest possible temperature. This makes helium ideal for uses in superconductors, including MRI machines – one of the fastest growing components of helium demand.

Helium Supply

But where do we obtain this elusive gas?

It turns out that new helium is actually created every day in very tiny amounts within the Earth’s crust as a by-product of radioactive decay. And like other gases below the Earth’s surface (i.e. natural gas), helium gets trapped in geological formations in economical amounts.

Today, much of helium is either produced as a by-product of natural gas deposits, or from helium-primary gas deposits with concentrations up to 7% He.

Here’s helium production by country:

Country2016 production (in billion cubic feet)Share
USA (from Cliffside Field)0.814%

USA (from Cliffside Field)
The USA government has a helium stockpile at the Cliffside Field in Texas, developed as part of a WWI initiative. It is in the process of being phased out, and by as late as 2021 it will no longer contribute to supply.

In December 2013, the Qatar Helium 2 project was opened. This new facility combined with the first helium project makes the country the 2nd largest source of helium globally.

Russia is looking to become a player in helium as well. Gazprom’s Amur LNG project will be one of the biggest gas facilities in the world, and it will include a helium processing plant. This won’t be online until 2024, though.

Though not a helium player yet, scientists have recently uncovered a major helium find in the Rift Valley of Tanzania which contains an estimated 99 billion cubic feet of gas.

The Future of the Helium Market?

Because of inflated demand, especially for cryogenics in MRI machines, helium prices have risen significantly over the years.

And with these market dynamics in mind, it’s clear that the future of helium is not full of hot air.

Click for Comments


Charted: The Exponential Growth in AI Computation

In eight decades, artificial intelligence has moved from purview of science fiction to reality. Here’s a quick history of AI computation.



A cropped version of the time series chart showing the creation of machine learning systems on the x-axis and the amount of AI computation they used on the y-axis measured in FLOPs.

Charted: The Exponential Growth in AI Computation

Electronic computers had barely been around for a decade in the 1940s, before experiments with AI began. Now we have AI models that can write poetry and generate images from textual prompts. But what’s led to such exponential growth in such a short time?

This chart from Our World in Data tracks the history of AI through the amount of computation power used to train an AI model, using data from Epoch AI.

The Three Eras of AI Computation

In the 1950s, American mathematician Claude Shannon trained a robotic mouse called Theseus to navigate a maze and remember its course—the first apparent artificial learning of any kind.

Theseus was built on 40 floating point operations (FLOPs), a unit of measurement used to count the number of basic arithmetic operations (addition, subtraction, multiplication, or division) that a computer or processor can perform in one second.

ℹ️ FLOPs are often used as a metric to measure the computational performance of computer hardware. The higher the FLOP count, the higher computation, the more powerful the system.

Computation power, availability of training data, and algorithms are the three main ingredients to AI progress. And for the first few decades of AI advances, compute, which is the computational power needed to train an AI model, grew according to Moore’s Law.

PeriodEraCompute Doubling
1950–2010Pre-Deep Learning18–24 months
2010–2016Deep Learning5–7 months
2016–2022Large-scale models11 months

Source: “Compute Trends Across Three Eras of Machine Learning” by Sevilla et. al, 2022.

However, at the start of the Deep Learning Era, heralded by AlexNet (an image recognition AI) in 2012, that doubling timeframe shortened considerably to six months, as researchers invested more in computation and processors.

With the emergence of AlphaGo in 2015—a computer program that beat a human professional Go player—researchers have identified a third era: that of the large-scale AI models whose computation needs dwarf all previous AI systems.

Predicting AI Computation Progress

Looking back at the only the last decade itself, compute has grown so tremendously it’s difficult to comprehend.

For example, the compute used to train Minerva, an AI which can solve complex math problems, is nearly 6 million times that which was used to train AlexNet 10 years ago.

Here’s a list of important AI models through history and the amount of compute used to train them.

Perceptron Mark I1957–58695,000
Neocognitron1980228 million
NetTalk198781 billion
TD-Gammon199218 trillion
NPLM20031.1 petaFLOPs
AlexNet2012470 petaFLOPs
AlphaGo20161.9 million petaFLOPs
GPT-32020314 million petaFLOPs
Minerva20222.7 billion petaFLOPs

Note: One petaFLOP = one quadrillion FLOPs. Source: “Compute Trends Across Three Eras of Machine Learning” by Sevilla et. al, 2022.

The result of this growth in computation, along with the availability of massive data sets and better algorithms, has yielded a lot of AI progress in seemingly very little time. Now AI doesn’t just match, but also beats human performance in many areas.

It’s difficult to say if the same pace of computation growth will be maintained. Large-scale models require increasingly more compute power to train, and if computation doesn’t continue to ramp up it could slow down progress. Exhausting all the data currently available for training AI models could also impede the development and implementation of new models.

However with all the funding poured into AI recently, perhaps more breakthroughs are around the corner—like matching the computation power of the human brain.

Where Does This Data Come From?

Source: “Compute Trends Across Three Eras of Machine Learning” by Sevilla et. al, 2022.

Note: The time estimated to for computation to double can vary depending on different research attempts, including Amodei and Hernandez (2018) and Lyzhov (2021). This article is based on our source’s findings. Please see their full paper for further details. Furthermore, the authors are cognizant of the framing concerns with deeming an AI model “regular-sized” or “large-sized” and said further research is needed in the area.

Methodology: The authors of the paper used two methods to determine the amount of compute used to train AI Models: counting the number of operations and tracking GPU time. Both approaches have drawbacks, namely: a lack of transparency with training processes and severe complexity as ML models grow.

Continue Reading