SK hynix Tests $28 Billion of Market Capacity as Reg NMS Rewrite Looms

The week of July 7, 2026, compresses three structural stories into a single frame. SK hynix is pricing what would be the second-largest share sale in history. The SEC has proposed gutting the order protection rule that has governed equity trading for two decades. And the exchanges are filing for 23-hour sessions while the clearinghouse quietly extends its capacity to support them. None of these developments is happening in isolation – and the trajectory each one implies points toward a market that will look fundamentally different by this time next year.

The $28 Billion Test

SK hynix kicked off formal marketing on Monday for a US listing of American Depositary Shares on Nasdaq, with pricing expected Thursday and trading to begin Friday morning, July 10{{cite:1a34cfc02a46}}. The South Korean memory-chip maker is offering 17.79 million shares at a reference price of approximately $158.14, targeting roughly $28 billion in proceeds{{cite:c91e730cfd05}}. Each ADR represents one-tenth of a common share{{cite:1a34cfc02a46}}.

If it reaches that target, the offering would surpass Saudi Aramco’s $25.6 billion IPO in 2019 to become the second-largest share sale on record, behind only SpaceX’s $85.7 billion debut last month{{cite:1a34cfc02a46}}. It would also set the record for the largest overseas company listing in the United States{{cite:f4f44011ce4a}}.

The deal is an AI-memory story. SK hynix’s first-quarter sales nearly tripled year over year, and the stock has gained more than 260% since the start of 2026{{cite:1a34cfc02a46}}. The company dominates high-bandwidth memory (HBM), the category most directly exposed to data-center AI accelerators, and it is raising capital to fund manufacturing expansion alongside Samsung, with the two firms pledging more than $550 million in new facilities{{cite:1a34cfc02a46}}.

But the offering also tests something broader: the market’s capacity to absorb new AI-adjacent supply after a year of extraordinary gains. Micron, the only major US-listed pure-play memory maker, has surged more than 700% over the past year and crossed $1 trillion in market capitalization{{cite:1a34cfc02a46}}. The Roundhill Memory ETF, which holds Korean and American memory stocks, became the fastest ETF in history to surpass $20 billion in assets after launching in April{{cite:1a34cfc02a46}}. The question is whether that appetite extends to a secondary listing that dilutes existing shareholders – SK hynix’s Seoul-listed shares edged lower on Monday on dilution concerns{{cite:1a34cfc02a46}} – and whether the bookbuilding process this week confirms or cools the narrative.

My read: demand is likely robust but not unbounded. Petra Capital Management analyst Albert Yong told Reuters he expected “relatively robust” demand despite recent volatility{{cite:1a34cfc02a46}}, which is a calibrated endorsement, not a victory lap. I would put the probability of the deal pricing within its reference range at roughly 75% – the 25% tail involves a softer book that forces a discount, not a withdrawal. The bigger tell will be first-week trading volume: a premium to the reference price on heavy volume signals the market wants more AI-memory supply; a drift below it signals saturation.

The Reg NMS Rescission: Rewriting the Floor

While SK hynix tests demand, the SEC is proposing to rewrite the rules that govern how every order in the market finds its price. On June 11, 2026, the Commission proposed amendments to Regulation NMS that would rescind Rule 611 – the Order Protection Rule, also known as the trade-through rule – and Rule 610(e), the prohibition on locked and crossed markets{{cite:26b9f29c8583}}. Comments are due by August 17, 2026{{cite:26b9f29c8583}}.

If adopted, this would be the most significant change to US equity market structure in two decades{{cite:26b9f29c8583}}.

Rule 611 currently requires every trading center – exchanges, alternative trading systems, and broker-dealers that internalize orders – to establish policies preventing “trade-throughs” of protected quotations. In plain terms: you cannot execute a trade at a price inferior to the best displayed quote on another venue. The rule was designed as a regulatory backstop to the duty of best execution, ensuring that investors who post aggressive limit orders are not bypassed{{cite:26b9f29c8583}}.

The SEC’s rationale for rescission rests on four arguments: markets are now highly automated and interconnected, making mandatory linking obsolete; Rule 611 has increased costs and complexity, including exchange proliferation and convoluted routing logic; best execution obligations alone are a sufficient safeguard; and rescission may facilitate competition and innovation, including tokenized securities trading{{cite:26b9f29c8583}}.

The counterarguments are not trivial. Without Rule 611, trades may lawfully occur at prices that would previously have constituted a trade-through, placing heightened pressure on broker-dealers’ best execution analyses and documentation{{cite:26b9f29c8583}}. The rescission of Rule 610(e) means locked and crossed markets could be displayed to investors making routing decisions, potentially causing confusion and undermining confidence in displayed market data{{cite:26b9f29c8583}}. And the connectivity that the SEC cites as evidence that mandatory order protection is no longer needed developed in part because Rules 611 and 610(e) required or incentivized access to protected quotations in the first place{{cite:26b9f29c8583}}.

A parallel filing on June 12 proposed amending the CT Plan’s revenue allocation formula to impose a 5:1 cap on each member’s quote-to-trade revenue ratio, targeting low-volume exchanges that earn outsized market data revenue by posting quotations with minimal trading activity{{cite:26b9f29c8583}}. In combination, these two proposals could remove the economic incentive for displayed liquidity on marginal venues and create pressure on low-volume exchanges to consolidate or close.

The trajectory here is clear: the SEC is moving toward a market structure that relies less on regulatory mandates and more on competitive incentives and broker obligations. Whether that produces a more efficient market or a more fragmented one is genuinely uncertain. I would estimate a 60% probability that adoption within the next 12 months produces a net reduction in venue count and routing complexity, and a 40% probability that the loss of the trade-through prohibition leads to wider effective spreads in less liquid names as internalizers gain discretion over where trades execute. The comment period will be the first real signal of which way the industry leans.

23-Hour Sessions and the Overnight Plumbing

The third structural shift is the extension of trading hours toward a 24-hour cycle. NYSE Arca filed a proposed rule change on May 12, 2026, to establish an Overnight Trading Session and amend its Early and Late Trading Session hours, with the goal of operating 23 hours a day – from 9:00 PM to 8:00 PM, five days a week, with a one-hour break for clearance and maintenance{{cite:5866ef7461e2}}.

The clearing infrastructure is being built in parallel. On May 27, 2026, the SEC approved NSCC’s proposed rule change to support extended trading hours for US equity markets, providing clarity around processing times for equity clearing services{{cite:24cc9972ad2a}}. And on May 27, Nasdaq filed the Twenty-Seventh Amendment to the National Market System Plan to Address Extraordinary Market Volatility, proposing temporary price band protections for overnight trading – a recognition that circuit-breaker frameworks designed for the 9:30-to-4:00 session need recalibration for a market that may never truly close{{cite:24cc9972ad2a}}.

The convergence of these filings tells you the exchanges and the SEC are coordinating, not competing, on the transition. NYSE Arca’s FAQ, updated May 2026, notes the extended-hours launch is “pending SEC approval” with a “targeted launch in 2026”{{cite:5866ef7461e2}}. The NSCC approval has already landed. The price-band proposal for overnight sessions is the kind of detail that gets filed when launch is quarters away, not years.

Citadel Securities’ Scott Rubner framed the stakes in his 1H 2026 Market Structure & Flows review published June 30: “Markets entering the second half of 2026 bear little resemblance to the markets investors navigated for most of the past two decades. The defining story of 2026 has not been a single macro event, it has been the structural transformation of equity markets”{{cite:4fbd3f744706}}. That is a market-maker telling you that the infrastructure layer, not the macro layer, is where the action is.

The $1.2 Trillion Buyback Floor

Underneath the IPO calendar and the rule changes, corporate repurchases are running at a pace that constitutes a structural demand source in their own right. US corporations are on track for approximately $1.2 trillion in share repurchases in 2026{{cite:8ad2abef2ec8}}. Daily active repurchase programs have surged from roughly 10 to between 50 and 60, with the buyback ETF PKW up 17% over the past year as mid-caps and industrials join the wave alongside the usual mega-cap tech suspects{{cite:8ad2abef2ec8}}.

Deutsche Bank argued in a June note that record corporate earnings are keeping buybacks on track even as AI capital spending climbs sharply, with capex across major hyperscalers accelerating without cannibalizing repurchase budgets{{cite:8ad2abef2ec8}}. Manufacturing profits jumped 31% year over year, giving non-tech CFOs fresh ammunition for buybacks{{cite:8ad2abef2ec8}}.

This matters for the market-structure story because buybacks are a predictable, price-insensitive flow that liquidity providers can model. If the Reg NMS rescission widens effective spreads, buyback execution desks will be among the first to feel the cost. If 23-hour sessions launch, buyback algorithms will need to decide whether to participate overnight, when liquidity is thinner and the new price-band protections are untested. The $1.2 trillion floor is both a stabilizer and a stress test for the plumbing changes above it.

The Convergence

Stack the three stories and a single thesis emerges: the market is being asked to absorb a record-breaking IPO, a rewrite of its core execution rules, and a near-doubling of its trading hours – all within a window of months, and all while corporate demand for its own shares runs at an all-time high.

Each development is defensible in isolation. The SK hynix IPO reflects genuine AI-memory demand. The Reg NMS rescission reflects a market that is more automated and interconnected than the one the rule was written for. Extended hours reflect global demand for US equity access. The buyback wave reflects strong earnings.

What is less certain is whether the system can handle all four simultaneously without friction. The most likely scenario – I would put it at 65% – is that each change lands with moderate disruption: the IPO prices successfully but with some book softness, Reg NMS comments reveal a divided industry, extended-hours launch slips to Q4, and buybacks continue undisturbed. The tail risk – roughly 20% – is that one of these transitions produces a visible market-quality incident: a poorly priced overnight session, a trade-through controversy in the first weeks without Rule 611, or an IPO that prices below range and signals supply saturation.

The remaining 15% is the upside case: the IPO prices above range, the comment period on Reg NMS is constructive, extended hours launch on schedule, and the market absorbs the structural shift with minimal disruption – confirming Rubner’s thesis that the transformation is already largely priced in.

What to Watch Next

The market’s structural transformation is not a single event to watch for – it is a process to track quarter by quarter. The SK hynix IPO is the first real test of whether the demand side can keep pace with the supply and infrastructure changes that are already in motion.