• Anthropic, OpenAI among large issuers to have filed for IPO
• European listing malaise continues after KNDS IPO pulled

With the first half behind us and US investors back at their desks following Independence Day celebrations on 4 July, American equity capital markets are beginning 2H in a celebratory mood.

The US IPO market has exploded with activity in the first half of 2026, driven by optimism around artificial intelligence.

European new listings, though, remain constrained by structural pressures and ECM bankers are licking their wounds again after the postponement of Franco-German tank maker KNDS at the beginning of 2H.

There have been 194 IPOs in the US so far in 2026 worth USD 155.8bn, as of 2 July, the day before the US broke up for the Independence Day holiday. At USD 86.2bn, the historic IPO of SpaceX accounts for more than half that figure.

But even if you were to remove the largest listing of all time from the figures, there still would be USD 69.9bn of IPO paper printed in the US up to 2 July, more than double the USD 31.6bn printed YTD in 2025.

The strength of US IPO markets means that Dealogic’s IPO Health Index for the Americas, based on several inputs including issuance volume and aftermarket performance, is now close to post-2021 highs.

Momentum has continued since SpaceX, with 19 IPOs priced on US stock exchanges, raising a cumulative total of around USD 6.9bn.

These deals include the USD 1.68bn listing of Italian digital holding company Bending Spoons, alongside the USD 1bn IPO of UK industrial manufacturer Doncasters Group, listed as DPC Holdings.

“We’re seeing great quality assets with stories of serious, sustained growth and even better credible assumptions around their opportunities with realistic TAMs,” said one US-based banker.

Hype grows for 2H

The question of course is: can this continue?

As laid out in Dealogic’s first half highlights for 2026, the remarkable start to US ECM this year is centred on the growth story behind AI.

If underlying markets remain strong, and investors continue to buy into equity stories around the AI revolution, then ECM Pulse would expect US deal flow to continue at this tearing pace.

Interplanetary expansion aside, Elon Musk’s vision for orbital data centres was seen by many SpaceX investors as a unique way to overcome the terrestrial limitations on AI compute capacity and data centre growth.

His decision to rent xAI’s earthbound compute capacity to Anthropic and Alphabet has also led many to call SpaceX the ultimate AI infrastructure investment. SpaceX’s equity story was enough to convince huge swaths of retail and institutional money to buy shares, with FOMO outweighing any sober assessment of the group’s financial fundamentals.

“The debate on everyone’s minds is over how long does this cycle lasts,” said a second banker. “Investors are saying they can’t model these businesses on normal parameters given any ROI will take five-to-10 years to play out.”

This banker noted that if institutional and retail investors were happy to continue to back that AI build-out story, there was enough US-based capital to support much more ECM issuance in 2H.

The key deal that investors are eying in 2H26 is now Anthropic which, like fellow private market giant OpenAI, has confidentially filed for an IPO.

It has been reported that OpenAI might delay its IPO until next year, but there remains comfort among most ECM practitioners that there is enough cash in the market to support both should they come this year.

“I think Anthropic comes first, and then OpenAI. Look just what has happened in the last month: Google came and raised USD 50bn in roughly a two-day market,” noted the first banker. “It all comes down to each individual company’s ability to demonstrate their ROI on that investment.

“If that continues, I think capital will be there to support that growth.”

The second banker cautioned that investors should keep an eye on SpaceX shares around the company’s first listed earnings in late July/early August, when concurrently several insider lockups are set to expire.

“If SpaceX starts to trade down that could close the door for other AI issuers.”

Also, any change in the investment thesis around premium LLM providers will of course limit Anthropic and OpenAI’s ability to go to public. They are both seeking to list at a time when scepticism towards large language model providers is growing, even from voices considered to be among the technology’s greatest supporters.

A US investor expressed concerns that, seen in this light, both Anthropic and OpenAI could be more vulnerable than they seem if sentiment shifts.

“Those companies don’t have the idiosyncratic features that the SpaceX setup had,” said the investor. “There is this narrative that they need to outrun each other, raising debt and getting a better cost of capital.

“I think that sounds terrible for the broader market if you have OpenAI and Anthropic both list in September and October.”

European price battles continue to sink deals

Unlike the general optimism in the US, Europe starts 2H under a cloud following the postponement of the most high-profile IPO expected before the summer break, Franco-German tank maker KNDS.

The transaction was beset by poor trading among its peer group, German defence spending muddles and the rise of asymmetric drone warfare as opposed to traditional tank-based conflict.

This misfiring in Europe’s IPO market suprisingly comes at a time when the bid for continental equity remains robust, as investors seek greater diversification away from a hot US equity market, given the previously discussed fears over AI.

But the issues impacting European IPOs as a product are structural, as evidenced by Dealogic’s IPO Health Index for EMEA which, in contrast its Americas counterpart, remains close to its post-2021 nadir.

KNDS is a microcosm of the wider issues plaguing European listings.

The family shareholders that own a significant part of KNDS wanted a valuation of at least EUR 12.5bn for the business at IPO, essentially an exit price for a long-term shareholding.

But investors didn’t want to pay it, demanding a more substantial discount against falling defence sector peers to allow for a greater chance of aftermarket upside.

Unlike deals in the US, Europe relies on a far smaller, and more selective investor base of mutual fund investors and hedge funds, alongside pockets of domestic demand, to see deals over the line.

This is the total opposite of something like SpaceX, which could rely on legions of retail investors and institutional money to create enough buzz to push the deal across the line.

Accelerated index inclusion also put more pressure on institutions to buy the deal.

There was a small window of collective exuberance at the beginning of 2026 around defence names like Czechoslovak Group (CSG), but this evaporated as defence ECM deals, and the wider sector, have disappointed.

Investors don’t feel obligated to buy any European IPOs like they did SpaceX and can hold out for at least a 30% discount to peers to offset risks around newly listed businesses.

Without wholesale changes to the European investor base, this isn’t going to change.

“Recent reforms to facilitate listings in Europe are great to speed up listing preparations but we now need to encourage pension funds and more retail investors,” said a banker close to the KNDS IPO.

While a deeper European market would certainly help European IPOs, this isn’t happening soon. In the interim, seller behaviour needs to change. European investors are happy to look at fresh listings but are demanding sizeable discounts to take part.

There is some hope for IPO activity before the summer break, including Digi Spain and Switzerland’s Infracore.

If sellers can accept a lower price at listing, then European IPOs are possible. If not, their deals won’t work no matter how much they dream of American-style investor FOMO.