SK Hynix Q2 26 – Greed and Fear

A question of who you believe.

  • Missing earnings is not what one should do after a blistering rally, but with the shares on less than 5x PER and demand apparently strong as ever, perhaps this is a chance for those that missed it to get in.
  • In the natural resources and commodity businesses, pullbacks of 50% are not uncommon during a secular bull market, but they do not usually occur in the space of a month as they have with SK Hynix.
  • This means that if one believes what the AI industry, the hyperscalers and the model companies are saying, SK Hynix, Samsung and Micron will probably recover their losses and go on to make new all-time highs.
  • However, SK Hynix’s results have seriously spooked the market, sending the shares down almost 20% before recovering to a 10% loss for the day.
  • Q2 26 revenues / operating profit were KRW79.3tn / KRW60.5tn, which missed forecasts of KRW84.0tn / KRW64.0tn and which raises fears that the medium estimates are too high.
  • SK Hynix did not miss estimates because demand was weak but because the street got too excited about how much product the company had available, which led to forecasts being too high.
  • This, combined with the IPO of CXMT and stories about Chinese companies being able to manufacture restricted semiconductor equipment to get around import restrictions, only added to the panic.
  • The reality remains that SK Hynix has more demand than it can handle at the moment and, like Micron, it is using its current market power to reduce its cyclicality.
  • This involves entering into long-term contracts with customers that guarantee volume and pricing for SK Hynix but also put a cap on how much more the price of memory will rise.
  • In effect, SK Hynix is selling further upside in terms of pricing in return for longer-term volume and a minimum price level, which will go a long way to ensure that the notorious volatility in the memory market is mitigated.
  • Despite what the memory makers say, this will not stop the market from being cyclical, but it will mean that its effect on the memory makers’ financial performance will be smoothed over a number of years rather than occurring in 6 months.
  • This means that the quality of earnings for the memory makers will improve substantially, which in turn makes the shares that are supported by those earnings less cyclical and therefore they should be able to support a higher earnings multiple than they have historically.
  • So, it comes down to a question of who you believe.
  • If you think that this is the end of the AI roll-out, that demand is not materialising, and that all of the capex budgets are going to be cut, one should sell the memory companies and sell them immediately.
  • However, so far this earnings season, capex is going up, not down, while the model makers are seeing parabolic revenue growth and the hyperscaler revenues are accelerating to new levels of growth.
  • Hence, there is no sign of a slackening of demand for memory, leading me to agree with Micron that the first move back towards a balance between supply and demand is going to come in 2028.
  • Hence SK Hynix is likely to make at least KRW300,000 per share in 2026, with KRW450,000 likely in 2027.
  • This puts SK Hynix on 4.7x 2026 PER and 3.0x 2027, which is a bargain if one thinks that AI rollout is going to continue for a while.
  • There are obvious parallels with the internet bubble here, but in 2000, there was no demand whereas this time there is more demand than anyone can supply.
  • Hence, for as long as demand holds, I suspect that the AI trade will recover once Korea gets its wild leveraged ETF trading under control (Friday) and when market participants return from their summer holidays (September).
  • I still hold a position in Samsung Electronics, which has lost 37% of its value in a single month, but this has been cushioned by my position in ServiceNow and Adobe, which are trading with a strong negative correlation to the AI trade.
  • I am tempted to increase my position in Samsung or take a position in SK Hynix or Micron here, given that the immediate panic looks overdone.

RICHARD WINDSOR

Richard is founder, owner of research company, Radio Free Mobile. He has 16 years of experience working in sell side equity research. During his 11 year tenure at Nomura Securities, he focused on the equity coverage of the Global Technology sector.

Leave a Comment