The World Computer – A look at those trying to build it

One of the promises of the decentralised web (web3) is the idea that you will be able to harness all the idle power of computers all over the world and use that power to do amazing things that require lots of computational power. You will in effect be able to create huge supercomputers by aggregating the computing power of thousand or even millions of small ones.

in order to do this, so the theory goes, all you have to do is create a marketplace with the right incentives for people to make their computing power available, and others to buy it and pay for the privilege of running their computations on them.
In practice this is fiendishly complicated, but that hasn’t stopped people from trying to create such a marketplace with the correct incentives.

In this post I will review three outfits trying to create this World Supercomputer marketplace: Golem, SONM and

I do not claim to have anywhere near enough technical knowledge to assess the merits or otherwise of these platforms. I am merely looking at it from a layman’s perspective, and from the perspective of someone who is trying to figure out whether or not to invest money in these tokens.

SO let’s start with the characteristics they have in common:

  1. In all of them there is a marketplace where some users offer up resources and other users bid for and buy these resources. When the work is done (the computational task that was requested) the results are given to the buyer and the seller gets paid.
  2. They all claim to be decentralised, i.e. there is no central authority deciding who can or cannot participate, the price of things or any kind of dispute resolution. It is all done in the code.
  3. They each have created their own currency (token) that will be used to pay and get paid in their marketplace. If their marketplace is the one that succeeds, then their tokens will be in great demand and they will be rich!
  4. All the tokens are tokens on the Ethereum Platform. Therefore transactions of their tokens are transactions on the Ethereum Network and the consensus mechanisms will be those of Ethereum. Ultimately, to a greater or lesser extent, their success is tied to the success of Ethereum.

So what makes them different?


In the Golem marketplace, as well as buying and selling computational resource, you can also buy and sell the software required to run various tasks. So in their marketplace there are “providers” or computational resource, there are “requestors” of resources, but there are also “software developers”, who can upload their software to an “Application Register” and charge people for using it. For example, and this is the first proposed Golem use case, someone who needs to do CGI rendering (special effects for movies, which is often computationally expensive) could upload software that does that and charge others to use it.
The Golem team recognises that the potential for defective or malicious code is high, so it introduces other concepts such as “validators” (who do QA on software) and whitelists and blacklists of applications that the validators (or others) have checked and used.

How much money did they raise?

Their ICO (Initial Coin Offer) happened in November 2016. They issued one billion Golem Network Tokens (GNT) and raised one million ETH (around US$10 million at the time)
According to their White Paper, “The supply of GNT will be limited to the pool of tokens created during crowdfunding period”, so no more will be issued, ever.

What stage are they at?

Their first iteration, Brass Golem, was initially scheduled for May 2017 (six months after crowdfunding). There have been some delays and now they are being more coy about a release date.

Who is the team

The team is based in Poland and composed largely of Polish professionals with backgrounds in Economics, Software Engineering and Mathematics.


SONM (Supercomputer organized by network mining) claims to be not just a marketplace for computational tasks but a whole “World Computer”. Its White Paper criticises Golem and for their lack of grand vision and explains in quite a lot of technical detail how its architecture in effect mimics, or accounts for, all the building blocks of a computer but in a distributed, mega-scalable, way.

Their use cases, interestingly, are not dissimilar to the Golem ones, one of them being “rendering video and computer graphics”. They also include scientific research like drug development and weather prediction, areas of research that traditionally require lots of number-crunching and computational expense.

But because it is more than computational marketplace, SONM also claims that you can run your applications (say, your website) on their World Computer.

The ICO happened in June 2017 and 331.360.000 SONM (SNM) tokens were issued. They say they raised US$ 42 million.
They also say that no more tokens will be issued.

What stage are they at?

They say version 0.1 (which they call an Alpha) is out and version 0.2 is due in June 2017. There appear to be some implementations using the current release, see here for example.

Who is the team?

According to their site the main movers are two Russian software developers, one of whom works for Yandex, the big Russian search engine and technology company. The SONM Platform as a Service (PaaS) relies heavily on the Yandex “Cocaine” framework.

Like SONM, claim that their solution is a distributed Cloud infrastructure that can provide distributed applications equally distributed access to all the resources they need, i.e. computation, data and other services. stems from the field of Desktop Grid Computing, i.e. its founders have been involved in the area of harnessing idle computing power on people’s desktops to carry out computational tasks. The advent of the blockchain, they say, gave them a “mechanism to register and prove the contribution of participants”.

At its simplest it is similar to the other two in envisaging a marketplace where some users will offer various infrastructure services and others will consume and pay for them. In the case of, they say the platform will take care of finding the most optimal resources given the needs of the application that wants them.

Like with the others, the platform will contain an Ethereum-based transaction model, as well as various consensus mechanisms like a “Proof of Contribution” to ensure trust and traceability.

The ICO of their RLC token took place in May 2017 and they raised US$12 million
They say they aim to have a total of 87 million RLC although it is not clear whether their model envisages more tokens being created.

What stage are they at?

Their first release (v1) is scheduled for 6 months after the ICO, i.e. towards the end of 2017. They claim to have demoed some proofs of concept here and here.

Who is the team?

It is a multinational team based in France. The founders are French and Chinese computer science professionals. As mentioned before, they have a background in developing Desktop Grid Computing software like XtremWeb-HEP, an open-source software solution for this.

Further Reading

The Golem whitepaper is here
The SONM whitepaper is here
The whitepaper is here