More bells at the top?

Chip mania mascot SK Hynix — talk of the town in South Korea, where regulators are discovering that genies are loath to return to lamps once loosed — is coming to America.

The planned US listing could see the ADRs trading on the Nasdaq within two weeks, according to the F-1.

This is a big deal. Figuratively and literally. SK Hynix says it’ll raise $29.4 billion in the share sale. That’d be on par with the Aramco IPO (which was local, obviously).

The figure above, which’ll be updated at least twice over the next 12 months when Sam Altman and Dario Amodei give the investing public an opportunity to buy into the bleeding edge, gives you some context.

There’s a lot going on here in terms of narrative overlap. From a 30,000-foot view, the SK Hynix listing is more equity supply atop the SpaceX deal, Google’s follow-on, possible equity raises from other hyper-scalers and public debuts from OpenAI and Anthropic.

Equity prices, you’re gently reminded, are like all prices: A function of supply and demand. For nearly two decades, stocks benefited from buyback-assisted float shrink, which is to say net supply was negative. Perennially. That may change going forward, and if it does… well, what happens to prices when supply outstrips demand?

Zooming in, no one’s going to blame SK Hynix for jumping at what it’s fair to call an unprecedented opportunity. In addition to (and because of) its status as the dominant member of the HBM oligopoly, SK Hynix is among the most talked about companies on the planet.

The parabolic surge in its locally-traded shares, shown above in all its log scale glory, makes more headlines every week, including and especially this one.

You gotta strike while the iron’s hot, and to mix proverbs, there’s no time like the present if you’re SK Hynix.

The paradox — and if you know the space, you know this is always the quandary — is that despite rallying by more than 1,100% in 12 months, the stock’s cheap. As discussed at some length here a month ago, it trades at just 7x on a forward multiple, 2.5 turns cheaper than Micron, 16 turns cheaper than Nvidia and 22 turns cheaper than the Mag7.

How’s that possible? Well, again, if you know the space, you know the story. This industry, perhaps more than any other, is notorious for boom-bust cycles.

The bull case says this cycle’s different, though. That AI’s changed the game. That, at the very least, the boom phase can last much longer this time. We’ll see.

Here’s the language the company employed in the F-1 “risk factors” section:

Recent growth in our revenue and profitability has been driven in significant part by strong demand for our memory products, including HBM and server DRAM products, from the expansion of AI infrastructure, such as AI accelerators and data centers. Demand for AI infrastructure has been driven in large part by significant capital expenditures by hyperscale cloud service providers and other large technology companies. If such customers reduce, delay or reprioritize their capital expenditures, including as a result of macroeconomic conditions, changes in business priorities, concerns regarding returns on investment or a sudden correction following a period of elevated spending, demand for AI infrastructure and related components could slow down materially, which in turn would decrease the demand for our products.

As boilerplate corporatespeak goes, that’s actually pretty useful. I’m not sure you even need sell-side research when SK Hynix can summarize it that concisely.

For now, SK Hynix’s position looks enviable, and that’s an understatement. The company “sit[s] at the chokepoint of the global AI buildout, with their products forming a critical bottleneck for data-center expansion,” to quote from Bloomberg’s coverage of the imminent US listing.

To reiterate: SK Hynix’s HBM market share is something like 60%. As one analyst put it, in remarks to CNBC last week, “It has the best product [at the] lowest cost. What else do you need?”

When you put it that way…