Micron FQ4 26 – Bottleneck Continues

Micron promises more of same.

• Good results from Micron and the promise of more to come, confirm that everyone is still sold out at least 12 months in advance, meaning that the transfer of value to the semiconductor sector is a theme that remains very much ongoing.
• Micron FQ4 27 revenues / Adj-EPS were $54.2bn / $33.42, around 5% ahead of expectations of $51.3bn / $31.77, which were powered by historically high gross margins of 86.8%.
• Micron also raised expectations for FQ1 27 with revenues / Adj-EPS of $61.5bn / $38.15, again comfortably ahead of expectations of $60.9bn / $37.68.
• Micron is in a position to take a leaf out of Nvidia’s book, meaning that it is able to guide just below what it thinks it will report next quarter, as it has at least 12 months of visibility.
• Hence, I think that all four quarters of fiscal 2027 are pretty much in the bag already, as when any customer can’t take its allocation for some reason, there will be plenty of others waiting in line.
• Hence, I see very little chance of an earnings miss in the next 12 months, which does reduce some of the risk on the shares.
• The company is also leveraging its current market power to try and reduce the extreme cyclicality of its business by offering somewhat better pricing to customers that are willing to commit to volume and price for 4-5 years.
• This won’t stop Micron from being cyclical, but it will have a smoothing effect where a collapse in the spot price takes 4 or more years to make its full impact felt on Micron.
• All of its competitors are engaged in exactly the same activity, which is proving to be fairly effective, as Micron stated on its conference call that it had 26 contracts covering over 36% of its revenues.
• This assumes that when the shortage is alleviated (which it will be), that everyone remains disciplined and no one breaks ranks and cuts price in an attempt to gain market share.
• Given the history in this industry, breaking ranks is a virtual certainty, and so there is a chance that customers fail to adhere to their contracts and switch suppliers when the price starts to fall.
• This will put memory makers in a difficult position because in a commoditised market it can be extremely foolish to sue one’s customers even when they are in breach of contract.  
• Hence, I think that there is a good chance that the cyclicality of memory remains exactly the same as it always has been, just with the cycle taking much longer given the unprecedented imbalance between demand and supply.
• The net result is that I think memory remains in a great place for the next 12 months, but it looks to me that I should be starting to look for the exit from this trade rather than allocating more capital to it.
• I continue to consider selling the position in Samsung Electronics and putting the initial capital and all of the profit into Nvidia for 2027.