Mega-IPOs Skew the Tally as Lockups Test Demand and the SEC Rewrites Market Rules

The headline numbers for the 2026 IPO market are deceptive in a way that matters for how investors should read the pipeline. Total US IPO proceeds have reached $145.5 billion year-to-date, a 543% jump from the prior year^[1] — but the number of deals priced (those with a market cap of at least $50 million) is down 25%, at 102^[1]. A single transaction accounts for the distortion. Below the surface, the market is telling two stories: one about supply concentration and one about whether the window is genuinely broadening.

The SK hynix Effect: One Deal, Six Times the Historical Average

South Korean memory chipmaker SK hynix (SKHY) raised $26.5 billion in its July Nasdaq debut through American Depositary Receipts, the largest US equity offering ever from a foreign issuer^[2][3]. The company sold 177.9 million ADRs^[3] and rose 12.8% on its first day, closing at $168.01^[3]. The deal was so large that July’s eight IPOs collectively raised $29.3 billion — more than six times the 10-year July historical average of $4.7 billion^[2].

Strip out SK hynix, and July’s remaining seven deals raised roughly $2.8 billion combined. That is a more modest figure, and it puts the “543% proceeds growth” headline in perspective. The IPO market has not flooded with capital so much as it has been skewed by a single mega-listing tied to the AI memory cycle. SK hynix is the world’s second-largest memory chipmaker, and its offering was fundamentally a play on HBM (high-bandwidth memory) demand from AI accelerators. Chairman Chey Tae-won told CNBC that AI agents and robots require “a lot of memory chips”^[3]. Whether that demand trajectory sustains is the variable that will determine whether SK hynix proves to be the first of several AI-infrastructure mega-IPOs or an anomaly.

SpaceX Lockups: The Real Supply Test

If SK hynix represents the supply coming through the front door, SpaceX (SPCX) represents the supply coming through the side. The company’s June IPO — the largest on record — priced 639 million shares, but the first lockup expiration on August 6 turned over 911 million additional shares, roughly 7% of shares outstanding^[4]. The stock was already under pressure: it had fallen more than 50% from its mid-June peak above $225 to close at $108.27 the day before the lockup lifted^[4]. The decline accelerated after SpaceX’s first earnings report showed capital expenditures more than twice as high as revenue^[4].

More supply is coming. On August 20, another 319 million shares become eligible for sale, followed by roughly 700 million in September and a similar amount in October^[4]. Elon Musk’s holdings — more than 6 billion shares — remain locked until June 2027^[4]. Mizuho analysts noted that shares becoming eligible for sale does not mean the full tranche will hit the market^[4], and that is the right framing. The question is not how many shares unlock but what fraction insiders choose to monetize. Early evidence is mixed: an Atlanta Falcons safety who invested $150,000 in 2022 announced plans to sell his entire stake^[4], while institutional holders have been more circumspect.

The broader lockup calendar extends well beyond SpaceX. According to a Reuters report, a staggered schedule will free additional SpaceX shares through mid-2027^[5]. For the IPO market as a whole, the number of recent listings means lockup expirations will be a recurring source of potential supply pressure through the fall.

The Pipeline: Nscale, Anthropic, and the AI Infrastructure Queue

Behind the priced deals, the filing pipeline remains active. Six companies submitted S-1 filings on August 12 alone, spanning AI infrastructure, biopharma, fintech, and minerals^[6]. The most consequential name in the queue may be Nscale, a London-based AI data-center company targeting a US IPO as early as September. The company has told prospective investors it has approximately $51 billion in total contracted revenue, with Goldman Sachs and JPMorgan advising on the listing^[7]. Quarterly revenue reportedly jumped to over $100 million in Q2 2026, up from about $37 million in Q1^[7].

The contrast with the broader market is worth noting. The Renaissance IPO Index fell 14% in July, underperforming a flat S&P 500^[2], as the rotation out of tech and AI infrastructure names weighed on recent debuts. Nscale’s filing is, in effect, a bet that the AI infrastructure thesis remains intact despite that rotation — that contracted revenue will be valued more generously than the sentiment-driven selloff in recent IPOs suggests. For that bet to pay off, the deal needs to price before any further deterioration in AI-spending confidence.

On the calendar for the current week, Renaissance Capital lists Lyntris (LYNX), with 24 million shares in a $19–$22 range for an estimated deal size of $492 million^[8]. The broader August calendar has been thin: the week of August 11 saw three IPOs and one SPAC price, with copper-foil producer Londian Wason (FOIL) raising $94 million and finishing up 10%, and biotech Vogenx (VOGX) raising $81 million^[9]. Renaissance characterized the period as the market’s “summer slowdown in full swing”^[9].

The Consumer Test: Jersey Mike’s and the Window’s Edge

The most instructive deal of recent weeks may not be a tech name at all. Jersey Mike’s (JMKE), the sandwich chain backed by Blackstone, priced 43.5 million shares at $23 — the midpoint of its $21–$25 range — raising approximately $1 billion in the largest restaurant IPO since 2011^[10]. The initial valuation was roughly $7.3 billion^[10]. Shares opened at $21 and closed at $21.63, down nearly 6% from the offer price^[10].

By historical analogy, consumer IPOs that break issue price on day one are not necessarily failures — but they do signal that institutional demand at the offered valuation was thinner than underwriters assumed. The deal’s structure is relevant context: Blackstone had owned the company for only about 18 months before listing^[10], making this a relatively quick sponsor-to-public exit. When a private-equity sponsor pushes a company public quickly and the stock trades below offer, the market is implicitly questioning whether the private-to-public valuation arbitrage has been stretched too far.

SEC Rewires Market Structure: Reg NMS and Tokenized Securities

While the IPO market processes its supply dynamics, the SEC is simultaneously rewriting the plumbing underneath it. On June 11, 2026, the Commission proposed rescinding Rule 611 of Regulation NMS — the trade-through prohibition that has required trading centers to execute orders at the best available displayed price across all venues since 2005^[11]. The proposal would also eliminate Rule 610(e), which restricts locked and crossed markets^[11].

Commissioner Mark Uyeda called the proposal “an important beginning in the broader, more complex journey of reforming the Commission’s equity market-structure rules”^[11]. Commissioner Hester Peirce supported the move, arguing that technological advances have rendered the rules unnecessary^[11]. The trade-through rule was designed for a fragmented market where best execution was difficult to verify; in an era of sub-millisecond routing and consolidated feeds, the SEC’s view is that the rule may be constraining competition among venues rather than protecting investors.

Separately, the SEC granted immediate effectiveness to a NYSE proposed rule change — filed April 9, 2026 — to enable the trading of securities on the exchange in tokenized form^[12]. The rule, designated Release No. 34-105260, references the DTCC’s December 2025 no-action letter for its tokenization services^[12]. NYSE American filed a parallel rule^[12]. The practical effect: NYSE can now list and trade securities represented as tokens on a blockchain, settled through DTC’s infrastructure. This is not crypto trading — it is traditional securities in a new wrapper — but it opens the door to settlement-cycle compression and potentially new distribution channels for IPO shares.

The Jones Day law firm characterized the Reg NMS proposal as potentially reshaping “listed equities and on-chain markets,” noting that removing the rigid trade-through routing mandate could give exchanges more flexibility in how they match orders^[13]. Both changes are at the proposed or early-effectiveness stage, and the comment period for the Reg NMS rescission is ongoing. But the direction of travel is clear: the SEC is pulling back from prescriptive order-routing rules while enabling exchanges to adopt blockchain-native settlement.

What the Numbers Say — and What They Don’t

The gap between the proceeds figure (+543%) and the deal-count figure (-25%) is the single most important data point in this market. It means new-listing supply is concentrated in a handful of large deals rather than broadly distributed. That has implications for liquidity: fewer new names means less diversification opportunity for IPO-focused strategies, and more sensitivity to how the mega-deals trade in their first 90 days.

What to Watch Next

• SpaceX lockup tranche 2 (August 20): 319 million additional shares become eligible^[4]. Watch the volume-to-price relationship — if selling is absorbed without a break below the $100 level, it signals real institutional demand beneath the headline volatility.
• Nscale IPO pricing (September target): The $51 billion contracted-revenue figure^[7] is the key variable. If the deal prices at a premium to comparable AI-infrastructure names, it validates the contracted-revenue thesis. If it prices at a discount or is delayed, the AI infrastructure rotation is deeper than the pipeline assumes.
• Reg NMS comment period: Watch for filings from large broker-dealers and market makers. The trade-through rule rescission would most directly affect internalizers and dark pools; their comment letters will signal whether the industry views this as deregulation or disruption.
• NYSE tokenized securities first listing: The rule is effective^[12], but no tokenized security has yet traded on NYSE under the new framework. The first listing — and whether it is an IPO or an existing security — will set the precedent.
• Biotech IPO cadence: August saw a flurry of biotech filings and pricings^[9][6]. Biotech has been the most consistently active sector in the 2026 IPO market. If biotech deal flow continues into September while AI-infrastructure deals stall, it would confirm that the IPO window is sector-specific rather than broadly open.
• Jersey Mike’s post-IPO trading: The stock’s ability to recover above its $23 offer price would signal that consumer-IPO demand is deferred rather than absent. A continued discount would suggest the sponsor-to-public valuation gap remains a structural headwind for consumer listings.

The base-rate view is this: IPO markets that appear flush because of one mega-deal tend to revert to a deal-count-driven reality within a quarter. The 2026 market is not an exception to that pattern yet — it is an illustration of it. What would have to be true for the optimistic case is straightforward: Nscale prices well, SpaceX lockup selling is absorbed, and the September calendar fills with more than biotechs and SPACs. What would have to be true for the pessimistic case is equally clear: the AI infrastructure rotation deepens, lockup supply overwhelms demand, and the $145.5 billion proceeds figure turns out to have been a SK hynix artifact rather than a market signal. The evidence available right now does not resolve that question — but the next six weeks will.

FN2 Research provides financial research and education, not personalized investment advice.