Crypto – The Merge

Good for payments. Bad for Nvidia / AMD.

• The Ethereum Blockchain will attempt to change the way it operates this week which will make it much cheaper to operate (good for payments) but at the same time, there will be a load of graphics cards that suddenly are no longer needed by their owners (bad for Nvidia).
• There are two big blockchains in use today, the Bitcoin blockchain which has a market capitalisation of $370bn and the Ethereum blockchain which has a market capitalisation of $201bn.
• However, a key difference is that the Ethereum blockchain does far more than the Ether cryptocurrency that competes with Bitcoin as it also runs most NFT systems as well as decentralised finance.
• Consequently, the Ethereum blockchain is larger than it looks and in my opinion is probably more important to the crypto industry than the Bitcoin blockchain.
• While the blockchain systems are decentralised and independent of any centralised control they also suffer from being expensive to transact, inflexible, slow and unstable in terms of monetary value.
• This makes crypto wholly unsuitable for being used as a payment system or medium of exchange.
• This explodes the myth that crypto is needed to run the economy of the Metaverse as in its current state, systems like AliPay and WeChat Pay are far better suited (see here).
• Furthermore, blockchain transactions are verified by a system called proof-of-work which involves very computationally heavy mathematics which is where the enormous energy consumption of blockchains comes from.
• To put it in context, Ethereum consumes more power than Chile and produces more carbon than Finland while Bitcoin consumes more energy than Belgium.
• These calculations are best executed on Nvidia graphics chips which is one reason why Nvidia did so well when the crypto craze was in full swing.
• Gaming revenues increased by a factor of 3x between FY2017 and FY2022 to top $11bn.
• The Merge involves the Ethereum blockchain merging with another system that does not use proof-of-work but uses a system called proof-of-stake.
• Simply put, proof-of-stake means that those that hold Ether tokens pledge those tokens as collateral in order to validate the transaction.
• There are some potential shortcomings with this system, but because it no longer uses computation for verification, its energy consumption should fall by 99% or more.
• Given how much energy Ethereum consumes and how much carbon it produces, this is a very significant move and while it will not solve any of the problems of blockchain as a payment mechanism, I think it is a big step in the right direction.
• One of the reasons why transactions are expensive on the blockchain is due to the cost of verification which will now fall substantially should the Merge be successful.
• I see this as the first of many steps towards making crypto usable for transactions, but a lot still needs to be done if crypto wants to run the economy of the Metaverse or anything else.
• Furthermore, if the Merge is successful then I suspect that many of the other blockchains will begin to migrate over to proof-of-stake over time.
• This will be nothing short of a complete disaster for the picks and shovels of the crypto industry as demand for their products could effectively go to zero.
• The largest one of these is privately held Chinese mining equipment maker Bitmain but further down the supply chain are Nvidia and to a lesser extent, AMD.
• Nvidia’s unique system of massively parallel computing with its Cuda cores is ideal for executing the computations required for proof-of-work.
• If this moves to proof-of-stake then the bitcoin miners will no longer need the graphics cards they have bought and the increase of graphics cards we have seen coming into the second hand market as a result of the crypto winter could become a flood (see here).
• While this could easily put vendors like Bitmain out of business, the ramifications for Nvidia are far less as Cuda still remains ideal for graphics for gaming, AI in the datacentre, the Metaverse as well as a number of other areas.
• Nvidia has millions of developers across a wide range of applications and so while proof-of-stake may hurt Nvidia in the short term, it is a dent rather than a crater.
• Hence, if the Merge is a success, I can see another couple more bad quarters for Nvidia as gaming performs worse than expected as gamers will be required to absorb graphics cards being sold by obsolete miners into the second-hand market.
• It is important to note that as of FQ2 2023, the data centre is twice the size of the gaming segment meaning that its ability to hurt Nvidia’s overall performance has already been significantly curtailed.
• I still remain extremely cautious on cryptocurrency which shows no real sign of emerging from its current deep freeze and in semiconductors I continue to prefer the cheaper, more defensive end of the sector.
• Qualcomm, MediaTek and TSMC all offer good growth at valuations which are much cheaper than AMD or Nvidia.
• I would like to own Nvidia one day, but I suspect that the crypto crunch may give me the opportunity to buy it at levels below even here.