Newsletter · · Ashutosh Agarwal

Anthropic Targets a 2 Trillion Dollar IPO as the Deal Wave Returns - Capital Markets: IPOs, M&A & Exchanges - Week of August 18, 2026

A synthesis of the past week's capital-markets podcasts for the week of August 18, 2026: Anthropic reportedly targeting a $2 trillion October IPO, IPO volume roughly doubling to $130 billion, Kalshi raising at $40 billion as prediction markets crash Wall Street's party, exchanges rebuilding their plumbing, and 12 state attorneys general emerging as the real antitrust risk to the deal wave.

Capital Markets: IPOs, M&A & Exchanges

Week of August 18, 2026: The Week the Deal Machine Roared Back to Life

For two years the story of capital markets was a story of waiting. Waiting for the IPO window to open. Waiting for boards to stop hoarding cash. Waiting for the regulators to get out of the way. This week the waiting stopped. The biggest company in the world tried to file for the biggest IPO in history, a $40 billion prediction-market start-up muscled its way onto Wall Street's turf, and an investment bank quietly told everyone that dealmaking has doubled. Underneath it all, a quieter and more important fight is playing out: who actually gets to say no to a merger anymore. Here is everything worth your time from this week's podcasts.

1. The AI IPO Supercycle Is Real, and the Bankers Are Already Playing Games

The headline came from the Financial Times on Wednesday night and by Thursday it was the only thing anyone on the finance podcasts wanted to talk about: Anthropic is targeting a $2 trillion IPO in October. If it prices there, it breaks the record that SpaceX set only weeks ago.

The numbers, as laid out on All-In (Aug 14) and Morning Brew Daily (Aug 14), are genuinely hard to believe:

  • Anthropic will end 2026 at a $100–120 billion annualized revenue run rate, up roughly 10x year over year. Morning Brew's framing: $1 billion in 2024, $9 billion in 2025, and now $100 billion-plus. One investor called it "the fastest revenue growth in technology history."
  • At a $2 trillion valuation, that is only 16 to 20 times sales, which, as the All-In hosts noted, is actually cheaper than where SpaceX and Palantir trade.
  • Alongside the IPO chatter, Anthropic is buying an AI start-up (reported as Descartes/Decart) for $6 billion to make its chips run more efficiently.

Here is the part professional investors should not skip. On All-In, veteran technology investor Gavin Baker poured cold water on the $2 trillion number, and explained a piece of banker behavior that is worth understanding:

"These leaks are always coming from the bankers who probably lost lead left and they want to make lead left look bad. So if it comes out at $2 trillion, it probably means that the testing the waters and the roadshow just went incredibly well."

In plain English: "lead left" is the lead underwriter, the bank whose name sits on the top-left of the deal document. Baker's point is that the losing banks leak a lowball valuation to the press to embarrass the winner. He thinks Anthropic could actually trade to $3 trillion, but that a smart company deliberately prices low so the stock can absorb the "lockup" (the period after an IPO when insiders still can't sell) without a nasty crash.

David Sacks did the scarier math. If Anthropic keeps growing 10x, it hits $1 trillion of annual revenue by the end of next year. Both men ultimately settled on a more sober number, $400 to $500 billion of exit run-rate revenue in 2027, which Sacks noted "would make them the biggest software company." The constraint isn't demand, it's physics: enough chips, enough electricity, enough data centers.

The one genuinely cautionary note came when Baker relayed that Anthropic CEO Dario Amodei has reportedly said the company "might be the only private company in the world at some point." Sacks's reaction:

"I might interpret that as a negative signal because it's so hubristic… This is getting into SBF land a little bit."

Why it matters: this is the deal that reopens the entire IPO market. Every banker, every late-stage fund, every "waiting" company is watching how Anthropic prices and how it trades on day one.

The Cautionary Tale Sitting Right Next to It

Before anyone gets carried away, listen to Scott Galloway on The Prof G Pod (Aug 17) on what actually happened with the SpaceX IPO. It's the best real-world warning about these mega-listings:

  • Priced at $135 a share, roughly a $1.8 trillion valuation, the largest IPO in history, bigger than Saudi Aramco. That put it at 94 times revenue on $18.7 billion of sales. The next most expensive large stock, Palantir, trades at about 67x.
  • Only 4–5% of the company was actually made available to trade (the "float"). Everything else is locked up.
  • It spiked to $225, then fell about 45% to $123.

Galloway's verdict is that Musk is "the greatest engineer of our time, but as a financial engineer." His mechanism: get included in the Nasdaq-100 at IPO, which forces every index fund to buy, huge mandated demand, while releasing only a sliver of stock.

"Smaller supply, greater demand, you know, disco… I think this is a $10 to $30 stock."

On the sell-side research with $800 price targets: "put down the mushroom chocolates… If you believe that investment banking and wealth management fees don't influence analyst reports, this makes Henry Blodgett in 1999 look responsible."

The read-through for the Anthropic deal: a tiny float plus index inclusion can manufacture a huge first-day pop that has little to do with fundamentals, and the real test comes when the lockups expire and insiders start selling.

And Don't Forget OpenAI

OpenAI is running right behind. On Bloomberg Intelligence (Aug 14), analyst Matthew Bloxham noted the company is now over a $40 billion annualized revenue run rate, double where it was at the end of 2025. Per The Rundown (Aug 14), OpenAI confidentially filed for an IPO in June at a valuation above $850 billion, but is widely expected to wait until next year. The same episode flagged the other big index event of the week: Reddit is joining the S&P 500.

2. M&A Has Quietly Doubled, and the Mid-Caps Are the Targets

If the AI listings are the fireworks, the real structural story came from Evercore Edge (Aug 12), where Bill Anderson, who runs Evercore's strategic M&A and shareholder-activism practice, gave a midyear read that every long-short PM should internalize.

The state of play, in his words:

  • The IPO market is running about double last year, from $67 billion to $130 billion so far this year, "wide open across a lot of sectors."
  • M&A is "mainly strategic parties… consolidation focused, a lot of mega caps." Financing for leveraged buyouts is available, cross-border deals are back, and, critically, "antitrust is okay generally."

His single most useful observation is a spread trade hiding in plain sight: the valuation gap between the S&P 500 and the mid-cap S&P 400.

"The disparity between companies trading on the S&P 500 versus 400… the stretch and the value of scale has become more prevalent. And I think some companies are taking that as a call to action."

The mid-caps, frustrated that they "can't attract capital" and stuck with "persistent undervaluation," are becoming willing sellers, and the big companies are paying premiums for scale. Anderson's warning to any under-scaled public company: "you can't wait to get hit because the market can tell when you're not ready."

A few other Anderson nuggets worth filing away:

  • 32% of spun-off companies get sold within three years of the spin. Spinoffs are now a pipeline of future targets, not exits.
  • Private equity firms are now teaming up with activist investors on joint hostile bids: the activist builds a stake, then bids alongside the PE fund.
  • Watch out for the mechanical drop in an acquirer's stock on deal news. Over the last decade, quant and retail ownership has risen "three or fourfold" and short interest has "almost doubled." When a company announces a deal and shuts off its buyback, the quants who were positioned for that buyback sell, so the buyer's stock falls even on a great deal.

The Rebound Is Broad, and Chemicals Is the Tell

On Chemical Week (Aug 13), Federico Minella of DC Advisory (the banking arm of Japan's Daiwa) confirmed the same pattern in his sector, with an important nuance: the big deals are back, the small ones aren't.

  • Announced transaction value has "rebounded significantly," led by mega-deals like Borealis-Bruges-Nova, while the total number of deals is flat and mid-market volume is soft.
  • The cascade effect is visible: once AkzoNobel merged with Axalta, rivals started reassessing, and Axalta itself "may now be sold" (it has added directors and hired bankers).
  • AI-exposed chemical assets (electronic materials, semiconductors, advanced packaging, thermal management, water treatment) are fetching over 20 times EBITDA. Minella's caution: not every company "with an AI-related narrative will command a similar valuation."
  • Private equity is the pressure valve: firms that bought in 2017–2021 have held "five, six, seven, eight years" and now face pressure "to return the capital and move on to the next fund cycle." He notes "more than half of the chemical business in the US" is family- or individually-owned, a huge pool of future sellers.

The Specific Deals That Crossed the Wire This Week

The podcasts surfaced a steady drumbeat of individual transactions across sectors:

  • Cannabis, Curaleaf goes hostile for Aurora. On The Dales Report (Aug 11), Curaleaf CEO Boris Jordan confirmed he took his bid public after Aurora rejected an initial offer "via a one-line email." The logic: cut Aurora's bloated 50% SG&A down to Curaleaf's 29%, harvesting $40–50 million of cost synergies, and push Aurora's high-margin product through Curaleaf's global distribution. Jordan, who owns 25%+ of Curaleaf, was blunt about why Aurora's board could stonewall: "This company has not one shareholder over 2%… It's a bunch of algorithms and retail shareholders." His pitch: "It's a match made in heaven. It just takes adults to sit down at the table."
  • Language services, RWS buys Acolad. Per SlatorPod (Aug 14), RWS is acquiring Acolad for roughly £20–22 million (about 2x adjusted EBITDA), a low price explained by Acolad's heavy debt load, adding 1,200 employees across 22 countries and pulling RWS level with TransPerfect at the top of the sector.
  • Hospitals, a JV wave. On Becker's Healthcare (Aug 12), Alan Condon described a surge in health-system consolidation, headlined by the Intermountain Health and AdventHealth Colorado joint venture combining eight hospitals and dozens of physician practices, modeled on the 2024 Henry Ford and Ascension deal and expected to close in early 2027.
  • Asset management, Goldman buys income. Per Brew Markets (Aug 12), Goldman Sachs (GS) is acquiring NEOS Investments for about $2.25 billion, buying its way deeper into the fast-growing options-income ETF business.
  • Utilities, a $67 billion megadeal meets a governor. More on this in the antitrust section, but note the deal itself: NextEra's proposed $67 billion purchase of Dominion Energy (D), flagged on Squawk Pod (Aug 11).
  • Quantum, IBM keeps buying. On The New Quantum Era (Aug 17), HRL chief scientist Thaddeus Ladd confirmed IBM's (IBM) definitive agreement to acquire HRL Laboratories, folding its spin-qubit technology into IBM's quantum foundry.

3. Prediction Markets Crash the Party, Kalshi at $40 Billion

The most fascinating capital-markets story of the week isn't a bank or an exchange you've heard of. It's Kalshi, the CFTC-regulated prediction market, which is now big enough that Wall Street is genuinely rattled.

The numbers, from The Information's TITV (Aug 12):

  • Annualized revenue has surpassed $4 billion, roughly doubling from $2 billion in June, driven by a frenzy of betting on the World Cup.
  • Kalshi is in advanced talks to raise money at a $40 billion valuation, up from $22 billion in May, when it raised about $1 billion. Its valuation nearly doubled in three months.
  • The relationship with Robinhood (HOOD) is fraying in a way that matters: a year ago Robinhood routed roughly 50% of Kalshi's volume; that is now under 20%, because Robinhood has redirected orders to a rival exchange it backs alongside trading firm SIG.
  • The regulatory sword: multiple states allege Kalshi is running illegal gambling. North Carolina already passed a 6% tax on prediction-market revenue from its residents, a preview of a future where Kalshi keeps operating but hands a cut to every state.

The Two Competing Visions

To understand the fight, listen to two lawyers on opposite sides of the microphone.

On The Future of Money (Aug 14), Neal Kumar, Polymarket's chief legal officer and a former CFTC staffer, made the case that these are derivatives, not gambling, and belong with the federal commodities regulator:

"There's an enormous difference between a platform that sells a product and is the counterparty to its customer versus an exchange that allows people to trade and buy and sell against each other… the CFTC has been regulating exchanges since 1974."

Two facts from that conversation surprised even seasoned listeners. First, sports are not the whole story: on Polymarket's global platform, sports are "below 50%," hovering around 30–40%. Second, and most relevant to this newsletter: Polymarket ran a market on the closing price of the SpaceX IPO, and "our traders got it right… a week or two before the IPO." Kumar sees this as the future, prediction markets as a price-discovery tool for securities themselves.

On Bloomberg Surveillance (Aug 14), Robert Dinalt, Kalshi's head of enforcement, addressed the trust problem head-on. Kalshi has partnered with NASDAQ to run its surveillance tools on Kalshi's data, and it caught a real insider:

"Our surveillance system… identified anomalous trading behavior… the president's prompter who made over $100,000 betting on specific words the president was going to say."

That case was referred to the CFTC, which is now pursuing it. Dinalt argues the exchange model polices itself faster than sportsbooks or even stock exchanges, "where it sometimes takes years." His most telling line about why Wall Street resists:

"I think people are scared of competition, to be honest, at least on Wall Street."

He also offered a great example of self-correcting markets: in the LA mayor's race, someone dropped a $1 million-plus bet on long-shot Spencer Pratt to move the odds, "and the truth is the market corrected that price point within nine seconds."

The Reality Check

Then came Breaking Points (Aug 13), and a much-needed dose of skepticism. Kalshi priced the Wisconsin Democratic primary at 95–5, and the 5% candidate won. The hosts made the fair statistical point (a 5% event should happen one in 20 times) but landed a sharper one: prediction markets are only "as good as the polls," and their track record is short.

The episode also exposed the money machine and the politics behind it:

  • Kalshi's defense was tweeted by an account whose bio read "powered by Kalshi" with "paid partnership" underneath, the influencer-sponsorship model that turned FanDuel and DraftKings into inescapable ad presences, now aimed at prediction markets ahead of football season.
  • The regulatory brawl: NY Attorney General Letitia James sought a temporary restraining order and over $36 billion in damages. In response, on August 11, CFTC chair Mike Selig, a Trump-appointed libertarian, invoked "emergency authority" to order Kalshi to keep operating. Kalshi had offered New York a 6% cut; New York said sportsbooks pay 20%.
  • The conflict nobody should ignore: Don Jr. holds advisory roles at both Kalshi and Polymarket, a paid strategic advisor with an equity stake at Kalshi, and an advisory-board member and investor at Polymarket.

Nearly everyone agrees this ends at the Supreme Court, which will decide whether states or the CFTC govern event contracts. And per Bitcoin And (Aug 14), even JPMorgan (JPM), which cut its banking ties to Polymarket in October 2025, is reportedly still interested in underwriting a Polymarket IPO. That is where this is heading: the disruptors become listings themselves.

4. Exchanges and Market Structure, the Plumbing Is Being Rebuilt

Away from the headlines, the machinery of trading itself changed this week.

CME goes 24/7 on metals. On Crain's Daily Gist (Aug 11): CME Group (CME) will trade its 100-ounce silver futures around the clock, seven days a week, starting September 11 (pending approval), following its one-ounce gold futures launched in January 2025. The context is a genuine retail boom: CME's metals business traded 1.3 million contracts a day in the first half of 2026, with precious metals up 55% year over year. Institutions are still 94% of volume, but retail is the fastest-growing segment, now 650,000+ traders. The strategy is explicit: mimic crypto's always-on markets so investors can react to weekend news.

Single-stock futures return, with a whimper. On TraderMerlin (Aug 11): CME reintroduced single-stock futures on July 27, a product banned since the 1982 Shad-Johnson Accord and briefly alive from 2002 to 2020. There are now 55 full-size and 22 micro contracts. But the host was withering on the launch: the most active contract, NVIDIA, traded just 714 contracts, about 71,000 shares against NVDA's 100 million share daily volume. "That's basically saying it's 0.1% of the total daily volume… I think it speaks volumes for itself."

The appeal is leverage (roughly 10x margin versus 4x in equities) and near-24-hour access. The risk he flagged: thin overnight liquidity, and the potential for institutions to use futures-plus-underlying to manipulate prices. One to watch, not yet one to trade.

Tokenization stops being a slide and starts being a date. On The Pomp Podcast (Aug 17), eToro CEO Yoni Assia dropped the most concrete timeline yet: the DTCC, the clearinghouse that settles US stock trades, is expected by year-end to let self-clearing brokers tokenize US equities on-chain, a shift he framed as potentially unlocking "$100 trillion in capital migration." On The Rollup (Aug 11), the zerohash CEO put it more bluntly: "Wall Street's biggest are all in on tokenization." The plumbing of who owns a share is genuinely being redrawn.

The data and index franchises quietly compound. Not from the exchanges themselves this week, but from Dividend Investing with Longacres Finance (Aug 12), a reminder of why these are among the best businesses in finance: customers who build their workflow around a data platform almost never leave. The show's estimated forward returns: S&P Global (SPGI) about 14.5% (almost all earnings growth), FactSet (FDS) about 15% (trading at a notable discount to its historical valuation), and MSCI (MSCI) about 19%, with a five-year dividend growth rate near 20% and a ten-year rate near 25%. Treat the specific "discount to fair value" figures as one retail investor's yield-based model, not gospel, but the underlying point stands: the toll-collectors of the index and data world keep raising the toll.

5. The Real Story, States Are the New Antitrust Cops

Here is the theme that ties the whole M&A boom together, and the biggest risk to it. The federal government has largely stepped back from blocking deals. So state attorneys general have picked up the hammer. Three fights this week show how it works.

Paramount and Warner Bros. Discovery. This is the marquee case. On The Capitol Forum (Aug 14) and Bloomberg Intelligence (Aug 12): the DOJ approved the roughly $110 billion merger, but 12 state attorneys general, led by California's Rob Bonta and New York's Letitia James, sued to block it, alleging harm to movie-theater distribution and cable licensing. A judge issued a temporary restraining order with what one analyst called "a fairly pointed opinion" suggesting the states' case has merit. Trial is set for March 2026.

The tactics have turned theatrical. On The Powers That Be (Aug 17) and Risk and Return (Aug 13): Paramount's David Ellison is threatening to relocate the company out of California by October 1 unless Bonta drops the case, abandoning the $4 billion Paramount lot and thousands of jobs, while the deal racks up $7 million a day in "ticking fees" after September 1. On Bloomberg Law, an antitrust professor was unimpressed with Ellison's offer to guarantee 30 theatrical releases a year:

"Anytime you say it's a three-year contract, an antitrust enforcer says, okay, what about the fourth year? You want to merge forever and you're giving me something for three years?… If you don't get it, then it's not gettable."

The NFL. On Bloomberg Law (Aug 12), NYU antitrust professor Harry First explained the DOJ's investigation into the NFL's streaming deals. The key legal point: the Sports Broadcasting Act exemption that lets the league pool its TV rights was written in 1961 for advertiser-supported broadcast, and it "is pretty hard to see how they're within the wording" for today's streaming deals. The league already lost the Sunday Ticket case (a $4.7 billion jury verdict before trebling, later tossed on a damages technicality). The political motive is not subtle: the House report opens with a Trump quote, "There's something very sad when they take football away from many, many people." But First is skeptical anyone follows through: "I don't really see any government group doing anything here."

Beer, a warning from a decree that just expired. On Brewbound (Aug 12), Chris Jones of the Main Street Competition Coalition explained what happens when the guardrails come off. The consent decree from the 2016 AB InBev and SABMiller merger expired in July 2024. For ten years it stopped the world's largest brewer from buying distributors or punishing them for carrying rivals, the conditions that let White Claw, Truly, and Athletic non-alcoholic beer explode. Now there's no independent monitor, and Jones flags 150–200 ABI distributors "seen as out of compliance" and a beer market already "highly concentrated" under the 2023 federal guidelines. His analogy for where an unwatched market ends up: the soda aisle, dominated by two players, where "you kind of have to accept what you get."

Why it matters: Bill Anderson said antitrust is "okay generally," and at the federal level, he's right. But the states are running their own enforcement, and they are winning temporary restraining orders and forcing settlements. For merger-arbitrage desks, the deal risk has moved from Washington to Sacramento and Albany.

6. The Macro Backdrop, an Open Window and a Wall of Issuance

Why is all of this happening now? Because financing is cheap enough and confidence is high enough.

On The Markets (Aug 14), Goldman Sachs Treasury trading head Mike Mitchell read the week's inflation print as friendly: softer core PCE (up 21.5 basis points) "should give the Fed comfort ahead of September's meeting," with markets pricing about 9 basis points of cuts. His longer-term worry is fiscal, not monetary: deficits pushing the term premium and long-end yields higher, with the 10-year auction printing its highest yield in years.

The most important macro observation for capital markets came from Morgan Stanley's chief US equity strategist Mike Wilson on Bloomberg Surveillance (Aug 14). He admitted the firm underestimated the earnings recovery, and, crucially, has been surprised by "the market's absorption of credit and equity issuance." That is the whole game right now: there is a massive wall of new stock and bond issuance coming (much of it to fund AI infrastructure and now these mega-IPOs), and so far the market is swallowing all of it without indigestion.

The question hanging over everything, raised across RiskReversal (Aug 14) and Prof G Markets (Aug 12), is whether the AI-financing loop (Nvidia's (NVDA) $500 billion program, private credit funding data centers) is a source of strength or the next credit problem. For now it is fuel. The moment the market stops absorbing the issuance, the IPO window that just opened will slam shut again.

The Bottom Line

This was the week the cycle turned. Anthropic's attempted $2 trillion listing is the signal flare, but the substance is underneath it: IPO volume doubled to $130 billion, mega-cap consolidation is cascading through sectors, prediction markets went from novelty to $40 billion overnight, and the exchanges are rebuilding their plumbing for a 24/7, tokenized, retail-driven world. The single biggest risk isn't the Fed and isn't valuation, it's a 12-state coalition of attorneys general who have decided that if Washington won't block deals, they will. Watch how Anthropic prices, watch whether the SpaceX lockup breaks the stock, and watch that March trial date in California.