A trading firm, three banks, and a sovereign fund all wrote checks into the same settlement layer this week.
Curated by Rob Saunders, WhoFiled. Interested in launching an industry intelligence report for your audience? Contact rob@whofiled.com.
Digital Asset, which runs the Canton Network, raised $355 million led by a16z crypto at roughly a $2 billion valuation, oversubscribing a $300 million target. Canton lets institutions issue and settle tokenized bonds, loans, and funds on shared infrastructure while keeping each transaction private, the one feature that has kept regulated firms off public chains. Look at who joined: Citadel Securities, BNP Paribas, HSBC, ABN Amro, Apollo, the Abu Dhabi Investment Authority, S&P Global, Tradeweb, Broadridge, iCapital, and SoFi. CEO Yuval Rooz framed it as plumbing: “For capital markets to move onchain, institutions need infrastructure that reflects how they actually operate.”
The instinct is to read that roster as the market pricing in tokenization. The more useful read is at least partly the opposite. None of these firms bet the balance sheet here. When you cannot tell which standard becomes the rail, you buy a seat at every table and let the option expire if it has to. The tell is that the buyers are the incumbents who lose most from a settlement layer they do not own. The clearest confirmation came the same week: seventeen of the largest US banks, many of them on Canton’s cap table, committed to build their own shared settlement network through The Clearing House, and pointedly have not chosen which blockchain to run it on. They are hedging external rails while building the one they would rather own. The counterargument is that several of them already run live workflows on Canton, Goldman’s and HSBC’s tokenization platforms among them, so this is commitment, not a hedge. But running production volume on a network while funding a rival you control is the hedge, not the refutation of it. Look at one firm: Goldman backed both AI implementation labs five weeks ago, sits among the seventeen banks in the Clearing House network, and runs a live platform on Canton. That is one firm, hedged three ways. The standard is not set.
Other notable investment tech raises
Capsa AI raised $18 million in a Series A co-led by TX Ventures and Pivot Investment Partners, building an AI operating system for private capital. The number that matters here is retention, not the round: 100 percent renewal, net dollar retention above 122 percent, 14x revenue growth. A fund that expands spend after adopting a tool has rewired a workflow around it, which is the difference between a vendor and a dependency. Private-capital back offices are proving stickier than the front-office AI tools that raised larger rounds earlier this year.
Vinyl Equity raised $20 million in a Series A led by Jump Capital, with MUFG Innovation Partners and Index Ventures participating. It is an SEC-registered transfer agent replacing legacy shareholder recordkeeping, paying agency, and transaction workflows with one real-time system across private and public markets. The license is the moat. Anyone can build recordkeeping software; a regulated transfer agent that also owns the workflow layer is positioned to be the system of record, the layer above the rail where control actually sits.
On the radar
The settlement-layer land grab was not one deal. EDGE Markets raised $29.2 million led by CoinFund to build settlement rails for regulated gaming and prediction markets, pursuing broker and FCM registrations so market makers can settle across pools without pre-funding.
Figure agreed to buy Kiavi for $717 million to fold loan origination into its blockchain-native marketplace, and Temenos agreed to acquire Additiv to put an orchestration layer over its core banking system. They all want to sit between the transaction and the books.
The incumbents are not waiting to buy in; they are making the move themselves. Citi began rolling out a product this week that lets clients trade tokenized shares of private companies, with the bank itself acting as issuer and custodian, which is the system-of-record position a startup like Vinyl is selling, claimed by a balance sheet instead. The same week, Field launched out of iAltA, which spent the spring buying BridgeFT and Precept and fused them into one operational database for an advisory firm’s back-office records, client accounts, and custodian data. It is the system of record again, this time built by acquisition.
The venue itself is in play too: Green Impact Exchange, an SEC-registered exchange preparing to launch later this year, filed to raise fresh capital for a market that trades traditional and tokenized securities side by side. This is not a tokenization wave cresting. It is a land grab for the chokepoint, happening before anyone knows which version of it is legal.
Away from the settlement story, the week’s other investment-tech raises clustered in advisor tooling. SyntheticFi crossed $2 billion in regulatory assets and disclosed more than $13 million raised since 2023, giving registered investment advisers access to portfolio-backed financing that used to sit inside private banks. Titan raised $3 million for AI trained on banking’s own compliance language, WealthReach $1 million for AI client-acquisition tools, and HyperNorm AI $2.2 million (Capital 2B, SenseAI Ventures) for decision-intelligence software that flags which client portfolios need attention across asset classes. In the UK, Vestd took its first institutional round, from Foresight, to expand its share-management platform and stand up as a licensed operator under the Financial Conduct Authority’s new private-company share-trading regime. The rest of the week ran from research to capital-markets plumbing: LinqAlpha ($7 million) for a multi-agent AI research platform for hedge funds and asset managers, TacticalMind AI ($1.5 million) for an AI tool that hands advisers buy/sell/hold signals with the reasoning attached, and CapConnectPlus ($2.5 million) for fixed-income issuance between issuers, dealers, and institutional investors.
The strategic takeaway
The money this period did not flow to a new asset. It flowed to the layer underneath, the one that decides whose books are the real ones. You can rebuild (vibecode) software. You cannot rebuild a license. The companies winning here are not the ones with the cleanest interface, they are the ones holding a license a competitor cannot replicate. The moat is permission to operate, and permission does not ship in a release note.
The trap is mistaking participation for commitment. Eleven of the largest institutions in finance funded the same settlement network, and outsiders read that as the standard being set. It is not. The standards fight is still open, and the window to take a position, influence a rail, or back a winner has not closed. The people you would expect to have closed that door are themselves still hedging.
The harder question is whether the chokepoint pays. We have seen this before: in other infrastructure shifts, the layer everyone races to own eventually becomes plumbing while the money moves to the layer above it. If a settlement layer owned in common by the institutions that use it becomes a utility rather than a tollbooth, the margin would sit a layer out, in the proprietary data priced across the rail and the client relationship that decides which rail to use. Own the layer if you can. But the real question is whether the prize everyone is racing for is the one that pays.
So a few things follow. In any infrastructure decision, ask what is protected by regulation rather than by code, and weight accordingly. When evaluating any AI or software vendor, lead with net revenue retention, not round size, because retention is the only number that tells you whether a tool has become a dependency or is still a trial. And decide deliberately whether you are a buyer of a rail or an owner of one, because in three years the firms that waited for certainty will be paying rent to the firms that did not. Optionality is cheap today. It will not stay that way.
Data sourced from SEC Form D filings, developer activity, and alternative signal tracking by Rob Saunders at WhoFiled. Reporting on this period’s deals draws on coverage from FinTech Global, FinTech Futures, CoinDesk, The Wall Street Journal, InvestmentNews, GlobeNewswire, PR Newswire, Wamda, and Temenos. Rob Saunders is exclusively responsible for its accuracy. If you have any feedback, please contact us.