Newsletter · · Ashutosh Agarwal

The IPO Window Is Wide Open and the Skeptics Are Getting Loud - Capital Markets Weekly - Week of July 14, 2026

Capital Markets Weekly for the week of July 14, 2026. US listings just printed their best half-year on record, led by SK Hynix's $26.5 billion Nasdaq debut, even as Gary Gensler, Michael Green and George Noble warned that sky-high prices, SpaceX's fast-tracked index inclusion and lockups, prediction-market froth, and a roughly $500 billion private-credit maturity wall make the boom feel late.

Capital Markets Weekly

Week of July 14, 2026: The IPO Window Is Wide Open and the Skeptics Are Getting Loud

IPOs, M&A and the plumbing of markets, this is what the podcasts were saying this week (July 7–14, 2026).

The headline of the year in capital markets is a happy one: this has been the best half-year for US stock listings that anyone has on record. Companies are racing to sell shares, foreign giants are choosing New York, and a South Korean chipmaker just pulled off the largest first-time US listing by a foreign company ever. Bankers are busy, the fee pool is fat, and the pipeline is full.

But listen a little longer to the people who do this for a living, and a second story runs underneath the first, a quieter, more nervous one. The same conditions that make it a great time to sell stock (very high prices) make a lot of veterans uneasy about buying it. And away from the glamour of the IPO stage, in the unglamorous machinery of private loans and buyout funds, the first cracks of a slow-moving hangover are starting to show.

The IPO Machine Is Running Hot

Start with the fact that set the tone. On July 10, SK Hynix, the South Korean memory-chip maker that supplies the high-bandwidth memory chips inside Nvidia's AI systems, began trading on the Nasdaq. It was, as Bloomberg Tech put it on the day of the debut, "the biggest ever U.S. listing by a foreign company."

The details are worth pausing on, because they tell you how hungry investors are right now. SK Hynix raised $26.5 billion, priced at $149 per share (technically an American Depositary Receipt, a way for a foreign company's shares to trade in the US). Bloomberg's equity reporter Bailey Lipschultz said the deal drew $200 billion of demand and was more than seven times oversubscribed, meaning buyers asked for seven times as many shares as were available. The stock opened around $171 and traded near $172 to $175, up roughly 15 to 17% on the first day. On Squawk on the Street (July 10), analysts noted Barclays estimated about $14 billion of automatic "passive" buying still to come as index funds are forced to add the stock over the coming months. JPMorgan ran the deal as lead stabilization agent.

Why does a profitable Korean company bother? Bloomberg Intelligence's Mandeep Singh explained on Bloomberg Tech that SK Hynix wanted the branding and the access, to become "a household name" among long-term US investors, and the chairman hinted at selling more ADRs and tapping the US debt market later to fund a planned $35 billion of capital spending. There is also a valuation gap the company would love to close: SK Hynix commands about 57% of the high-bandwidth-memory market but has traded at a discount to US rival Micron (roughly 6x forward earnings versus Micron's ~7x), partly because Americans simply could not buy the stock before.

SK Hynix was not alone. The week's pipeline was crowded:

  • Hub International, a Chicago insurance broker owned by private-equity firm Hellman & Friedman, is seeking to raise about $3 billion in an IPO, according to Crain's Daily Gist (July 7). It filed confidentially in late June and is working with Goldman Sachs and Morgan Stanley, plus Bank of America, Barclays, BMO and JPMorgan. It was valued at a $29 billion enterprise value in a minority investment round in May 2025. Crain's tied it directly to the boom: this "has resulted in the best half-year ever for U.S. listings, according to data compiled by Bloomberg."
  • Jersey Mike's, the sandwich chain, filed its S-1 (the document a company files before going public) targeting roughly a $12 billion valuation "later this month," per The Best One Yet (July 7) and a deep Run the Numbers breakdown (July 9). Blackstone bought it for about $8 billion in January 2025; a new store costs around $515,000 to open and throws off roughly 42% cash-on-cash returns, the kind of economics that make a franchise attractive to public buyers.
  • Bending Spoons, the Italian app company, went public at a $22 billion valuation with a $1.68 billion raise (and about $4.1 billion of debt), discussed on Slate Money (July 11).
  • And it is not just IPOs. On Bloomberg Talks (July 10), former SEC Chair Gary Gensler rattled off the capital-raising spree, "Why did Google raise $85 billion? Why did SK Hynix raise $26 billion...and SpaceX go public?", and gave the honest answer:

"These are all tells that people are saying, this is a high valuation market. I want to raise capital while I can."

That is the core insight of the week. A flood of new stock and bonds is not automatically a sign of health. It can also be a sign that management teams think their shares are as expensive as they will ever be, and want to cash in before the mood changes.

SpaceX, the Index, and the "Manipulation" Argument

No listing has been more divisive than SpaceX, which went public a few weeks ago in a roughly $75 to $85 billion raise at a valuation near $1.75 trillion, the largest ever. This week the debate shifted from the price to the plumbing: how it got into the indexes, and who benefits.

On Prof G Markets (July 7), host Ed Elson walked through the mechanics with Michael Green of Simplify Asset Management. As of that morning, SpaceX was officially part of the Nasdaq 100, meaning its ups and downs now sway the more than $1.4 trillion benchmarked to that index, and tens of millions of Americans own it indirectly through their retirement funds. The catch: SpaceX got in under new fast-track rules that cut the required trading history to just 15 days and eliminated the usual "float" requirement (the rule about how many shares actually trade freely). The S&P 500 looked at SpaceX and said no. Nasdaq said yes.

Green did not mince words about why:

"It is an opportunity to effectively exploit the index position to bring the company public, get it listed, drive the capacity for insiders to sell shares, and ultimately create liquidity... I think it is candidly quite manipulative."

His argument, in plain terms: index funds are supposed to be neutral, passive vehicles that simply own "the whole haystack." When an exchange bends the rules to usher one specific company in, one whose value, Green argues, is mostly tied to a "money-losing, nonprofitable, and lagging AI business," the losers are ordinary passive investors who do not know the rules were changed for them, and the winners are the insiders who now have a deep pool of forced buyers to sell into.

The unease showed up again on Bloomberg Surveillance (July 10), where veteran investor George Noble (once Peter Lynch's auto analyst at Fidelity) zeroed in on SpaceX's staggered lockup, the schedule for when insiders are allowed to sell. Starting next month, 20% of shares unlock, then roughly 7% more every 20 to 30 days, until 100% is freely trading by December. His warning:

"Grandma's 401(k) is the exit liquidity for this manipulation. And I think the regulators are asleep at the wheel."

Noble called SpaceX priced at 120 times revenues (Scott McNealy's famous "10 times revenues is crazy" line, times twelve), and flagged that the bonds of SpaceX already "can't find a bid," a red flag when a company's debt is struggling even as its stock gets celebrated. The stock itself, after opening near $150 and peaking around $217, has slid back to roughly $145 to $150.

Even Gensler, who is broadly pro-innovation, landed in the same neighborhood. He reminded listeners that markets tend to over-build during every technology boom, canals, railroads, electricity, the internet, and warned that today's valuations are "the highest that we've been in a very long time" on essentially every measure. His question about the AI trade driving all this issuance: "Just what happens if the data center space just doesn't grow next year, just stays flat?"

Why it matters: the IPO boom and the "we're near a top" chorus are two readings of the same fact, sky-high prices. For a portfolio manager, the useful takeaway is not "sell everything." It is that the supply of new stock is being timed to the mood, the insider selling on names like SpaceX is only just beginning, and the market's ability to keep absorbing mega-deals is now the thing to watch.

Prediction Markets Grow Up, and Show Their Teeth

The fastest-moving corner of "capital markets" this week was not stocks at all. It was prediction markets, regulated exchanges where you buy and sell contracts that pay out based on whether some future event happens (an election, a game, the weather). They are booming, they are fighting regulators, and this week they got both a founder's victory lap and a brutal reality check.

The victory lap came from Kalshi co-founder Tarek Mansour, interviewed on Long Strange Trip (July 9). Kalshi now claims 95% U.S. market share, having passed its offshore rival Polymarket. Mansour's whole strategy was to do the slow, painful thing, stay regulated onshore, while competitors operated offshore without a license. It took Kalshi four years to get its CFTC license (the CFTC is the federal commodities regulator), and years of what he called regulatory "warfare": enforcement actions, audits that normally take 10 days stretched to nine months, the government "yanking the clearinghouse from under us." He described the moment his litigator called to say Kalshi had won its make-or-break court case, about three and a half weeks before the 2024 election:

"He's like, we won. And after that, I don't remember, like we were throwing chairs in the office... we like basically destroyed the office."

His framing of the whole bet is a good one for anyone trying to understand these companies: "building the next generation New York Stock Exchange from within." That is the ambition, not a betting app, but a new kind of exchange.

The scale is real. On Squawk Pod (July 9), Kalshi advisor and former Nevada Senator Dean Heller sparred with a clip of former New Jersey Governor Chris Christie (now advising the casino industry), who called prediction markets unregulated sports-betting in disguise, claiming "about 90% of their business" is sports. Heller pushed back hard: Kalshi is regulated, with "a watchdog watching over what these prediction markets are doing" daily, and sports is closer to two-thirds of volume, not 90%. He cited $22 billion in prediction-market bets placed on the World Cup. His key distinction, and it is a fair one to understand, is that prediction markets are peer-to-peer: they match a buyer and a seller and take a fee, rather than betting against you the way a casino does. Elsewhere, on Around the Coin (July 8), a fintech executive cited a Bernstein Research projection that prediction markets could reach $1 trillion in annual volume by 2030. Polymarket, per Tech Brew (July 10), was already doing about $4 billion in weekly notional volume in June.

And Robinhood is charging in. On The Milk Road Show (July 7), a guest described Robinhood's prediction-market volumes swelling to $15 to $20 billion daily around the World Cup, with 1.5 million Robinhood users having touched the contracts.

Now the reality check, and it is an important counterweight, because the boom has a dark underbelly. On The Personal Finance Club Show (July 9), the host laid out the research plainly. According to data Kalshi shared with CNBC, over 70% of its traders were unprofitable over six months. The winners are extraordinarily concentrated: one study found the top 1% of users capture 76% of profits, another that the top 0.1% capture 67%. As he put it:

"If there is 1,000 people in the room, one guy is getting 67% of the profits... And then the 700 people are losing money. And would you want to walk into that room and see if you're the one?"

He also flagged the genuinely disturbing edge cases: a market on whether Venezuela's Maduro would remain in power, where a bettor put in $100,000 on "no" and won about $500,000 two days before a US operation captured him, the trader turned out to be an American soldier on the mission, now facing wire-fraud charges. And the "moral hazard" problem: a market on whether the LA Palisades fire would be contained creates, in theory, a financial incentive for someone to start a fire. Insider trading and manipulation are not hypothetical risks here; they are structural.

The legal fight, meanwhile, is tilting Kalshi's way for now, a judge denied a preliminary injunction against it (The Paul Barron Crypto Show, July 8), but the jurisdiction battle between the states and the CFTC is headed for a much bigger showdown.

Why it matters: prediction markets are becoming a real, monetizable business line for Robinhood, Kalshi and eventually the incumbent exchanges, a genuine new revenue stream. But the same episodes that celebrate the growth also make clear the product is, for most users, closer to gambling than investing, with real regulatory and reputational tail-risk. Both things are true at once.

The Stock Market's Edges Are Going 24/7

A related theme kept surfacing: the slow blurring of the line between the traditional stock market and the always-on crypto world.

Tokenized stocks, blockchain-based versions of shares that can trade around the clock, had a breakout month. Per The Wolf of All Streets (July 9) and Thinking Crypto (July 8), tokenized equities hit roughly $3.4 to $3.9 billion of volume in June, up on the order of 145% from May and more than 1,000% year over year, driven largely by pre-IPO SpaceX shares trading on crypto platforms. Backpack's SpaceX token alone did about $1.08 billion, and a rival version about $852 million. It is still tiny next to Bitcoin's daily turnover, but the growth rate is the story. The SEC is reportedly working on a "tokenization innovation exemption," and both NYSE and Nasdaq are running blockchain pilots (per crypto lawyer Ashley Ebersole on Thinking Crypto, July 9).

And Robinhood is trying to own the rails outright. On Empire (July 10), hosts detailed the launch of "Robinhood Chain," the company's own blockchain, which quickly logged over $500 million in trading volume on Uniswap (more than any chain except Ethereum itself) and about 150,000 active wallets. The strategic logic is verticalization: control the trading flow, monetize crypto users, and sidestep the regulatory friction of listing assets. Whether crypto's core problem is "value capture," actually keeping the profits rather than giving them away to users, was the honest debate the hosts kept circling.

Why it matters: for now this is a rounding error next to the New York Stock Exchange and Nasdaq. But it is the clearest sign of where the competitive threat to the exchange franchises eventually comes from, not another exchange, but a 24-hour, blockchain-native venue that a broker like Robinhood builds itself.

Underneath It All: the Private-Credit and Private-Equity Hangover

Here is the part of the week that the champagne-popping IPO coverage mostly skipped. While public markets celebrated, several thoughtful investors described the beginnings of a slow-motion reckoning in private equity (buyout funds) and private credit (loans made outside the banks), the corner of finance that ballooned during the cheap-money years.

The bluntest take came from Dan Rasmussen on Hedgeye's "The Great Private Equity Hangover" (July 11). His thesis: risk piled up, unnoticed, for years. Buyout firms bought software companies at "25 times EBITDA with 12 turns of leverage" during the 2021 to 2022 boom and assumed it would work. Now the bill is coming. The tell is showing up first in the publicly traded versions of these vehicles:

"If this publicly traded BDC is down 30% and you're telling me that my privately traded BDC is not down at all... seems like something's wrong."

(A "BDC" is a business development company, essentially a listed fund that makes private loans.) His deeper point is a matter of arithmetic: private equity sits below private credit in the capital structure, so if the loans are impaired, the equity has to be worth even less, yet private-equity marks have barely budged because "there's nothing to ever force anyone to take the marks." He expects a wave of "private credit on private equity, sponsor-on-sponsor violence," where lenders end up owning the companies. And he was witheringly funny about the roll-up craze, buying dentists, lawn-care outfits, pest control, even (he swears) public libraries, noting "there are more private equity funds in this country than there are McDonald's," and that "50 dry cleaners is not just as good as Google."

Torsten Slok, chief economist at Apollo, put hard numbers on the danger zone on The Real Eisman Playbook (July 13). Private credit is roughly a $2 trillion market, and about $500 billion of it is loans to software companies made in the last six or seven years. Those loans typically have seven-year maturities, so a wall of them comes due in 2028 and 2029, right when, if inflation stays sticky and new Fed Chair Kevin Warsh keeps rates "higher for longer," refinancing gets painful. You can already buy some of these software loans in the open market at a discount that implies a 12%-plus yield, the market's way of saying it is worried:

"It's a double whammy to software coming from the terminal value being questioned, and at the same time rates higher for longer, meaning that they're not able to service their debt."

Slok's reassurance, though, is worth holding onto: $500 billion is real pain for the lenders involved, but it is not systemic in a $33 trillion economy, largely because it sits in BDCs (limited by law to 2-to-1 leverage), pension funds and insurers, not in banks levered 30-to-40x the way they were before 2008. This is a slow leak, not a 2008-style detonation.

The credit backdrop was filled in nicely by CRE Exchange (July 9), which is worth reading even if you do not care about real estate, because it is the clearest single summary of first-half financing conditions:

  • Investment-grade bond spreads compressed to 75 basis points, near the tightest since before the 2008 crisis, with high-yield around 280 basis points.
  • Investment-grade issuance is on pace for about $1 trillion, roughly 25% ahead of last year. Of that, hyperscalers alone issued over $100 billion to fund the data-center build-out, "very deep, very cheap capital."
  • But the leveraged-loan market told the opposite story: private-equity-backed issuance fell nearly 30% quarter over quarter, sponsors froze, and "amend-and-extend" deals (kicking the can on debt maturities) hit record highs as everyone pushes their walls out to 2028.
  • Tellingly, software's share of leveraged-loan issuance collapsed to about 8.6%, the lowest since 2013, because credit markets are already pricing in AI's threat to software revenue.

RenMac (July 10) framed the mood as a "rolling train wreck": the very riskiest (CCC-rated) spreads are still fine, but the higher-quality BB and BBB tier is "a watch point," and the stress at the market's fringes, private credit, gold, a strong dollar, has not yet filtered through to mainstream corporate borrowing. On The Julia La Roche Show (July 11), Chris Whalen made a similar observation about spreads between government and corporate bonds beginning to widen, a classic medium-term slowdown signal.

Why it matters: this is the counter-narrative to the IPO party, and it is the more important one for risk managers. The public equity and investment-grade bond markets are wide open and cheap. The leveraged-buyout and private-credit machine, especially anything touching software, is quietly seizing up, and the 2028 to 2029 maturity wall is the date circled on everyone's calendar.

Also Worth Hearing

A few shorter items that rounded out the week:

  • M&A is oddly quiet where you'd expect it to be loud. On Oil Markets (July 9), analysts noted that despite high prices and a friendly regulatory regime, US oil-and-gas dealmaking has stayed subdued in 2026. Devon's roughly $58 billion acquisition of Cotera stands out precisely because so little else is happening, most tier-one crude assets have already changed hands, and buyers are waiting for the bid and ask to line up while eyeing gas and LNG instead. It is a useful reminder that a "favorable regulatory climate" does not automatically produce deals.
  • Antitrust is still biting, just mostly abroad and in media. On Pivot (July 10), the hosts detailed the Paramount and Warner Bros. merger drawing challenges from multiple US states (led by California's Rob Bonta) and British regulators, with possible concessions like divestitures. And on Grumpy Old Geeks (July 9), the Getty Images and Shutterstock merger was reported terminated over UK regulatory objections, a deal break driven by London, not Washington.
  • Bank earnings are the next test. WSJ's Take On the Week (July 12) previewed the big banks, JPMorgan, Bank of America, Citigroup, reporting Q2 results, with investment banking, trading revenue and loan-loss provisions as the swing factors. The SpaceX IPO and the wider issuance boom should show up as a tailwind to the advisory and underwriting lines; whether the credit worries above start denting provisions is the thing to watch.

The Bottom Line

This was a week of split-screen. On one side: the busiest, most confident new-issue market in living memory, a record foreign listing, a fat pipeline, and a fee bonanza for Wall Street. On the other: a growing list of serious people, Gensler, George Noble, Michael Green, Dan Rasmussen, Torsten Slok, pointing out that the same high prices fueling the boom are also what make it feel late, that insider selling on the marquee names is only starting, and that in private credit and buyouts a slow, quiet reckoning has already begun.

Neither story cancels the other. The window is genuinely open, and the smart move is to notice who is climbing through it, and in which direction.