Newsletter · · Ashutosh Agarwal

SpaceX's Hangover and the IPO Traffic Jam - Capital Markets: IPOs, M&A & Exchanges - Week of August 11, 2026

A synthesis of the past week's capital-markets podcasts for the week of August 11, 2026: SpaceX's rocky first earnings and lockup expiration, an IPO traffic jam behind Anthropic and OpenAI, NVIDIA's $500 billion AI-financing consortium, the prediction-markets legal war, and an M&A wave from P&G-Thorne to a potential $400 billion AstraZeneca-Bristol Myers deal.

Capital Markets: IPOs, M&A & Exchanges

Week of August 11, 2026: SpaceX's Hangover and the IPO Traffic Jam

The pipes that carry money into public markets got a real stress test this week. The biggest company ever to go public reported its first earnings, and the results were less a celebration than a warning about what happens when you buy a stock nobody knows how to value. A logjam of would-be blockbuster listings is stacking up behind it. Wall Street's biggest names sat down together to invent a new way to fund the AI build-out. And a scrappy set of "prediction markets," where you bet real money on whether something will happen, spent the week getting sued by state after state.

Below: what happened, why it happened, and what named investors, executives, and hosts said about it, all drawn from the past week's podcasts.

1. SpaceX: the biggest IPO in history, and the biggest hangover

Start with the deal itself, because how it came together explains almost everything about the mess that followed.

SpaceX went public in June 2026 at $135 a share, valuing the company at roughly $1.75 trillion, the largest initial public offering (IPO) ever, meaning the first time a company sells its shares to the general public. On The Wall Street Skinny (Aug 5), the hosts walked through just how unusual the whole thing was. Normally a company does a "roadshow," where management meets big investors like Fidelity and BlackRock, collects orders at different price points, and lets supply and demand set the price. None of that happened here. As they put it:

Elon Musk just came down on high and said, we're going out at $135... there was a take it or leave it price. There was no matching supply and demand. It was just $135.

They noted the eerie coincidence that $135 was also, almost exactly, the price that tipped Musk into being a trillionaire on paper. Demand was enormous anyway, the company sold $75 billion of stock (all of it going to SpaceX's own balance sheet, since Musk and other insiders didn't cash out) against roughly $250 billion of orders, four times oversubscribed. Fidelity alone reportedly put in for about $5 billion, more than the size of some entire past IPOs.

Here's the part the hosts flagged that most people miss: the first-day "pop" (the jump from the offer price once trading opens) wasn't luck; it was engineered. Three things guaranteed the stock could only go up at first:

  • Tiny float. Only about 3–5% of the shares actually traded freely, versus a normal 10–20%. Less supply, same demand, higher price.
  • No shorting allowed. For the first 30 days, the banks that underwrote the deal couldn't lend out shares to investors who wanted to bet against it, so there were almost no sellers.
  • A self-fulfilling target. With no comparable companies and no profits to anchor a valuation, traders in "perpetual futures" markets (a kind of never-expiring bet on the price, common in crypto) had already pegged the stock around $174, exactly a 20% pop on $135. So that's roughly where it opened. As the hosts said, "you don't have fundamentals... a pretty good guess is a 20% pop."

Even NYU valuation professor Aswath Damodaran, whom the hosts called one of the most respected names in finance, ran a full discounted-cash-flow model and landed around $1.2–1.3 trillion, well short of the $1.75 trillion price. Their dry aside: "What's a half a trillion in between friends?"

Then reality arrived. SpaceX reported its first quarterly results as a public company in early August, and the market didn't like the bill for its AI ambitions:

  • Revenue came in around $7.8 billion, beating estimates near $6.8 billion, per The Rundown (Aug 5) and Bloomberg Intelligence (Aug 5). But the company spent a staggering $18.4 billion on capital projects in the quarter, about $15.8 billion of it on artificial intelligence, and posted a loss (The Rundown cited $541 million; Bloomberg's team pegged the AI-business operating loss around $1.26 billion, which they framed as actually better than feared).
  • On CNBC's Fast Money (Aug 4), the math behind the AI spending got concrete: SpaceX plans to build 15 gigawatts of computing capacity, which early investor Gene Munster of Deepwater Asset Management estimated could mean roughly $600 billion of spending at $40–50 billion per gigawatt. Munster remains a believer, describing SpaceX as a vertically integrated AI company with fast payback. Starlink, the satellite-internet business, now has about 12 million subscribers and drives 55–60% of revenue.
  • Bloomberg's analysts pointed to a $47.5 billion backlog of contracted work, including about $1.25 billion a month from AI lab Anthropic and more than $2 billion from Google starting next quarter, as reason to expect revenue to keep climbing.

But the stock fell hard, down somewhere between 6% and 10%+ on the print across the various recaps, and it has roughly halved from its post-IPO peak. Two things are weighing on it:

  • The lockup expiration. A "lockup" is the period after an IPO when insiders are barred from selling. SpaceX's expired this week: The Rundown described roughly $100 billion of stock (around 900 million shares) becoming eligible to trade, which effectively doubles the available supply and removes the artificial scarcity that inflated the price. Risk and Return with Charlie Gasparino and Bob Sloan (Aug 6) framed the whole setup bluntly in its title: "SpaceX Beat the Quarter. Shorts Await the Unlock." (Musk himself is locked up for a full 366 days, and other insiders face price hurdles, they can only sell if the stock trades ~30% above the IPO price for a stretch, a structure designed to prevent a flood of selling.)
  • The valuation is a leap of faith. On Best Stocks Now (Aug 6), Bill Gunderson warned against the stock at a ~$1.45 trillion valuation on maybe $30–50 billion of annual sales, in a world of 4.6% ten-year Treasury yields that make far-off profits less valuable. The Rundown cited Morningstar analyst Nicholas Owens' June fair-value estimate of $63 a share, against a stock trading around $100–130. Not everyone's bearish: on The KE Report (Aug 6), Joel Elconin called it a possible institutional accumulation phase and said he was a buyer near $110.

Why it matters: SpaceX was supposed to be the deal that reopened the IPO market. Instead, its wobble is the cautionary tale hanging over everyone else who wants to go public.

2. The IPO traffic jam, and Wall Street's new machine for funding AI

Behind SpaceX sits a long line of companies waiting to test the public markets, and the mood is nervous.

Who's in line (per Bloomberg Intelligence, Aug 10):

  • Switch, a data-center landlord (a real-estate investment trust), filed confidentially for a November IPO at an enterprise value near $50 billion, after a $2 billion funding round led by venture firm Andreessen Horowitz.
  • Formlabs, a 3D-printer maker backed by SoftBank, is in early talks for a roughly $500 million raise.
  • Anthropic is expected to go public in September–October, with OpenAI possibly following later.

The two AI labs are the real prize, and the podcasts disagreed on timing and wisdom. On The Rundown (Aug 2), The Information's Alex Heath said Anthropic will likely list in October 2026 and OpenAI probably not until early 2027, adding that both may have already missed their peak-hype window. On The Liquidity Event (Aug 6), the hosts reported OpenAI is eyeing a $1 trillion valuation, with CEO Sam Altman pushing for a faster timeline while the finance chief wants to wait given how much cash the company is burning. And on WSJ's Take On the Week (Aug 9), the worry was explicit: SpaceX's cold reception could scare off appetite for the OpenAI and Anthropic deals unless they can show a fast path to actually making money.

The plumbing question underneath all of this: where does the money come from to build AI at this scale? This week produced two answers.

  • Intel launched a $15 billion stock offering to fund its AI push, and Bloomberg reported it was oversubscribed almost immediately (Bloomberg Intelligence, Aug 10). Selling new shares to raise cash is the oldest capital-markets tool there is, and it still works when demand is hot.
  • NVIDIA convened a $500 billion AI-financing consortium, arguably the most striking capital-markets story of the week. On Closing Bell Overtime (Aug 10), CNBC sat down with NVIDIA CEO Jensen Huang alongside the heads of six of the biggest names in finance: Goldman Sachs (David Solomon), BlackRock (Larry Fink), Blackstone (John Gray), KKR (Vladimir Szlezak), Apollo (Jim Zelter), and Brookfield (Bruce Flatt). The idea: NVIDIA plays matchmaker, connecting customers who need to fund AI build-outs with these firms, which have pledged to find over half a trillion dollars. Crucially, NVIDIA puts in none of its own money, this is not a "circular" deal where the chipmaker funds its own sales. Huang described the moment as a "phase shift," saying compute has become an asset class the way water and power did a century ago. The scale behind it: Morgan Stanley estimates the big cloud companies alone will spend about $3.5 trillion between 2026 and 2028, and Fink noted each gigawatt of AI capacity costs $50–60 billion to build, with the U.S. needing more than 70 gigawatts.

Why it matters: The AI boom has outgrown any single company's balance sheet or any single bank's loan book. Wall Street is now building shared machinery to fund it, and that machinery, not the IPO calendar, may be where the real capital-markets action is.

3. Prediction markets go to war with the states

If you want to see a brand-new market being born in real time, lawsuits and all, look at prediction markets: platforms where you buy contracts that pay out based on whether an event happens (an election result, a Fed rate move, whether a politician shows up to a speech).

The fight came to a head when New York Attorney General Letitia James sued Kalshi, the leading platform, calling it an illegal, unlicensed gambling operation. Kalshi's head of enforcement and legal counsel, Robert DeNault, made the company's case forcefully on Bloomberg Businessweek (Aug 5). His core argument: Kalshi is a federally licensed exchange overseen by the Commodity Futures Trading Commission (CFTC), the federal derivatives regulator, not a casino.

Calci is a licensed federally regulated exchange. By her logic, any federally regulated exchange... can suddenly be subject to the whims of state criminal enforcement.

DeNault drew a sharp line between an exchange and a casino: gambling, he argued, is when you play "against the house," a business that profits when you lose and limits winners. On Kalshi, he said, "traders set the price. We match traders in an open marketplace with one another. We are not on the other side of individuals trading." He also warned the same legal logic could be aimed at Robinhood, crypto trading, or any derivatives venue, and noted James had already sued Coinbase and Gemini in April on similar "illegal gambling" grounds.

The state-versus-federal collision is now a genuine legal brawl:

  • TechLinked (Aug 1) reported the CFTC filed an emergency motion to stop New York's enforcement action less than an hour before the lawsuit landed, the federal regulator actively defending the platform against the state.
  • Reuters World News (Aug 1) noted at least four states, Massachusetts, Michigan, Nevada, and Washington, have already won court orders restricting Kalshi.
  • On NerdWallet's Smart Money (Aug 6), Brookings scholar Aaron Klein explained why this is so unsettled: the cases are splitting across courts (Minnesota went against the platforms, Nevada went for them), and he expects it ultimately reaches the Supreme Court. On InvestTalk (Aug 5), attorney Justin Klein argued New York has a strong case because these contracts fit the legal definition of gambling and let 18-year-olds bet, where sports betting requires being 21.
  • The stakes are eye-watering: The Smashi Business Show (Aug 4) reported New York is seeking up to $36 billion in penalties, at $100,000 per alleged illegal wager.

There was also a vivid enforcement story. Former Congressman George Santos agreed to pay more than $35,000 to settle allegations he manipulated a bet on his own platform activity, specifically a wager on whether he'd attend the 2026 State of the Union. DeNault confirmed on Bloomberg that Kalshi's own surveillance detected the trading, interviewed Santos, and referred the case to the CFTC, his proof that the platform polices itself like a real exchange.

Meanwhile the industry keeps landing mainstream deals even as it fights in court. Front Office Sports Today (Aug 7) reported the New York Yankees signed a deal with Polymarket, the fourth individual pro-team partnership with a prediction market (after the Blackhawks, Rangers, and Mets), on top of Major League Baseball's $300 million league-level Polymarket deal.

Why it matters: Prediction markets are trying to graduate from novelty to regulated financial infrastructure. Whether they're "exchanges" or "gambling" is now a multi-billion-dollar legal question, and the answer will decide whether a whole new asset class survives.

4. The deal wave: from vitamins to a $400 billion drug megamerger

M&A, mergers and acquisitions, where one company buys another, had a busy week across sizes.

The headline deals:

  • Procter & Gamble bought Thorne, a premium supplements brand, for $3.8 billion. On the Joshua Schall Audio Experience (Aug 10), industry strategist Joshua Schall explained why the price shocked people: P&G paid 5.85 times Thorne's projected 2026 revenue of $650 million, versus the roughly 3x that supplement companies usually fetch. His argument is that P&G wasn't really buying pills, it was buying "a data rich AI powered predictive health ecosystem." Thorne's at-home testing kits feed into a machine-learning platform (Wongevity) that builds a personalized supplement plan, which locks customers into subscriptions: "Once a consumer hands over their biological health data, switching to a generic retail vitamin brand becomes nearly impossible." Schall also flagged the private-equity windfall: L Catterton had taken Thorne private for just $680 million less than three years ago, so the sale locked in a fast, massive gain.
  • Visa is acquiring fraud-prevention firm BioCatch for $2.4 billion in cash (Bloomberg Intelligence, Aug 3), to bolster a risk-and-security business already generating about $2 billion in revenue. BioCatch's technology reads behavioral signals during logins and transactions to catch scams.
  • The whale: AstraZeneca and Bristol Myers Squibb. The Financial Times reported the two drugmakers are having "discussions about discussions" over what would be a roughly $400 billion combination. On Investors' Chronicle (Aug 7), Julian Hoffman noted AstraZeneca's market value is roughly double Bristol Myers' (~$264 billion) and framed the deal as an "insurance policy" to hit AstraZeneca's sales targets. But The Rundown (Aug 3) cautioned that antitrust regulators could tie it up for years, since both companies own competing blockbuster lung-cancer drugs.

The broader trend looked genuinely healthier beneath the megadeals. On Business Valuation... Real Talk (Aug 1), Stephen White of Onyx Partners Group reported U.S. M&A announcements rose 5.8% in June to 1,202 deals, with private-equity activity jumping 30% to 149 deals, and middle-market deal counts growing after 13 months of stagnation.

But the antitrust wall is the story. The clearest example is the Paramount–Warner Bros. Discovery merger (roughly $110 billion), which has become a landmark test of who gets to block a deal:

  • Federal regulators cleared it: the FTC, the EU, and the UK's Competition and Markets Authority all approved it (the UK with conditions, like keeping Channel 5 independent for five years), per The Information's TITV (Aug 7) and The Powers That Be: Daily (Aug 7).
  • But 12 Democratic state attorneys general sued to block it anyway, led by Washington's Nick Brown and California's Rob Bonta, after the Trump administration's DOJ declined to challenge it. On Pitchfork Economics (Aug 4), Brown laid out the argument that the merger illegally reduces competition in movie production and cable distribution; the states won a temporary restraining order freezing the deal.
  • A federal judge has now set a trial for March 2–19, 2027 (Strictly Business, Aug 5). That timing is the whole ballgame: The Powers That Be noted the real pressure is a June 4 "outside date," after which Warner's David Zaslav could walk away. The Information's Martin Peers predicted the deal only closes if Paramount agrees to sell CNN.
  • On Our Curious Amalgam (Aug 3), an antitrust lawyer drew the lesson for every boardroom: getting federal clearance is "no longer the finish line" for a merger, directors now have to plan for state challenges too.

Why it matters: Deal-making is picking up, but the rules of who can stop a merger just got a lot murkier. A green light from Washington no longer guarantees a deal closes.

Also worth hearing

  • Retail traders are now a real revenue engine. On The Compound and Friends (Aug 4), the hosts dug into Robinhood's quarter: options-contract volume up 50% year-over-year, $215 million of its $389 million in net interest revenue coming from margin lending, and $156 million from event contracts (its prediction-market business). This Week in Startups (Aug 5) put Robinhood's quarterly options revenue at $342 million.
  • The 0DTE phenomenon. On Deep Values (Aug 9), the host noted that "zero-days-to-expiration" options, contracts that expire the same day they're traded, now make up over 61% of all S&P 500 option volume. When enough people pile in, the dealers who take the other side are forced to buy and sell the underlying market to hedge, which can mechanically accelerate rallies (and drops) regardless of the fundamentals.
  • A rare open window: biotech IPOs. BioCentury This Week (Aug 11) reported four biotech companies went public on Nasdaq last week, Latigo, Blossom Hill, Braveheart, and Atovia, raising more than $1.2 billion combined. Per Lindsay Martin, that brings 2026's Nasdaq biotech IPO count to 20 companies raising $6.9 billion, with 14 posting first-day gains. Braveheart jumped 66% on day one for a $2.2 billion market value.
  • A price war in index funds. On Trillions (Aug 6), Bloomberg's Eric Balchunas described new competitors (iShares and State Street) launching Nasdaq-100 ETFs at 10–12 basis points to undercut Invesco's giant QQQ (15 bps) and QQQM (18 bps), a reminder that the economics of tracking an index are being squeezed thinner every year.
  • AI is now reshaping the bond market. On Exchanges (Aug 5), Goldman Sachs credit strategists Amanda Lynam and Zach Ablon laid out how much of the AI build-out is being funded with debt: the big cloud companies issued $108 billion in bonds in 2025 (27% of their capital spending) and $194 billion so far in 2026 (33%), with a forecast of $250 billion for the full year. AI-related debt now makes up 18% of investment-grade bond issuance, up from 1% in 2024.
  • The private-equity exit backlog. On The Dividend Cafe (Aug 10), David Bahnsen cited PitchBook data showing about 33,600 companies globally are still sitting unsold in private-equity portfolios (roughly 13,500 in the U.S.), 1,100 more bought than sold this year. It's far easier to buy a company than to sell one right now, which is exactly why so many firms are itching for the IPO window to reopen. He also noted credit spreads tightened from 287 to 270 basis points in under two weeks, a sign of returning risk appetite.
  • A cynical take on the IPO rush. On The Julia La Roche Show (Aug 8), analyst Chris Whalen offered the week's most quotable warning about why markets are being held together:

The big investment banks are going to hold this market together until they get a few more IPOs done and put those big fees in their pocket. And then they're going to stand back and let it go.

  • The macro backdrop. The overlay for all of this is a jittery bond market. July payrolls showed 23,000 jobs lost against an expected 80,000 gain (The Dividend Cafe), cutting September rate-hike odds to a coin flip. On Forward Guidance (Aug 3), the discussion centered on new Fed Chair Kevin Warsh's gamble to tighten financial conditions by pulling support from long-term bonds rather than hiking rates outright, pushing 30-year yields up and widening credit spreads. Higher long-term rates make every far-off-profit story (SpaceX, the AI labs) harder to justify, and make the debt funding the AI boom more expensive.
  • The tokenization angle. On Unchained (Aug 4), Securitize CEO Carlos Domingo discussed his firm's own IPO and argued tokenized stock, shares recreated on a blockchain, should come only from the companies that issue them, not third parties. And on Capitalisn't (Aug 6), Renée Jones explained the deeper backdrop to the whole IPO drought: decades of deregulation (the 1996 securities law and the 2012 JOBS Act) made it so easy to raise money privately that companies increasingly just… don't go public.

The through-line: capital markets are wide open for the right story and slammed shut for the wrong one. SpaceX proved a company can raise a fortune on almost pure faith, and then proved how fast that faith gets tested once the lockup lifts and the earnings are real. Behind it, a queue of AI names is waiting to find out which kind of week they'll have.