Intel – The bandage

Intel puts AMD in its box.

• Intel has finally launched new products that have a shot at putting AMD back in its box but the real competitor for Intel to deal with is no longer AMD but Arm.
• Intel launched its Raptor Lake series of processors the flagship of which is the Core i9-13900K touting 24 cores and clock speeds of up to 5.8Ghz.
• The chips are made on the Intel 7 process which is broadly equivalent to the TSMC 5/4 process and so these combined with AMD’s latest chips on TSMC 5nm should begin to answer the question of whether the x86 processor is obsolete.
• Assuming that Intel’s numbers hold when tested by 3^rd parties, the 13^th generation chips will match or slightly exceed AMD’s latest releases putting the two companies on an even footing for the first time in nearly 5 years.
• This will put a bandage over the wound, but there are much more serious issues to overcome.
• The most pressing of these in my opinion is technology leadership.
• There was a time when Intel was the most advanced and most efficient manufacturer of semiconductors in the world but its slip-ups at 10nm and 7nm cost it leadership to TSMC and Samsung which now needs to be wrestled back.
• Intel is spending hard to achieve this but the current economic downturn which is causing the semiconductor cycle to begin turning downwards is going to make this more difficult.
• This is because Intel is not nearly as strong financially as it was, and it will have to dig deep to get ahead of TSMC.
• The second problem is the x86 architecture itself where the question of its obsolescence remains open.
• For many years the situation in CPUs has been that if one needs performance one went to Intel and if one wanted battery life one went to Arm.
• The launch of the M-series of processors called this doctrine into question as Apple demonstrated that it was possible to have excellent performance and excellent battery life if one used the Arm processor in a certain way.
• This was one of the first products made on TSMC’s 5nm process and so there was always the possibility that this sudden jump in performance and power efficiency was in large part due to TSMC.
• To restore the situation, AMD and Intel’s processors need to substantially outperform the M-series when measured on single core performance.
• The time is drawing close when the silicon will be made available for testing and there is very likely to be a large number of these comparisons conducted.
• My suspicion is that the answer will end up somewhere in the middle with x86 still outperforming Arm, but the margin will be less than it was in the past.
• This will make life more difficult for Intel as its historical dominance has been predicated on the performance advantage that has been conferred by using the x86 design, the sales of which it dominates.
• However, almost all server software is still based on the x86 meaning that it will be a long and slow process to migrate server software to an alternative should their owners choose to do so.
• This will give Intel both time and cash flow to shift its business model, but it does mean that the outlook remains very unclear.
• This is reflected in the shares which have fallen substantially this year.
• My position has long been that when the shares hit a 12-month forward PER of 9x to 10x, the risks of investing in Intel are more than accounted for in the share price.
• The problem is that the earnings have been falling hard this year with 2022 guidance now standing at GAAP EPS of $2.57.
• At those multiples, this implies a share price of $23 – $26 compared to the current share price of $26.8.
• The outlook for the market, in general, remains very difficult and so I am expecting to be forced into a decision on Intel sometime before the end of this year.