Tokenization split into two camps that don't talk to each other. Public-equity tokenizers anchored on broker-dealer integrations. Private-market tokenizers anchored on transfer-agent rails. Path-dependent, not principled.
I had a long call last month with a wealth-product team in São Paulo who had spent a year integrating two separate tokenization vendors. One for tokenized AAPL exposure. A second for tokenized treasuries and a private credit fund. Two contracts, two compliance programs, two reconciliation pipelines, two on-call rotations. They asked me, half rhetorically, why they had to do this. The honest answer is that they didn't, but the category had taught them they did.
Tokenization split into two camps. On one side, tokenized public equities. Apple, Tesla, the S&P 500. Tracked through SEC-registered broker-dealers, custodied at Alpaca or Interactive Brokers, redeemable 1:1. On the other side, tokenized private market exposure. Pre-IPO secondaries, private credit funds, venture LP interests, tokenized treasuries. Issued through transfer-agent platforms or fund-administration rails.
Builders in each camp spent the last three years optimizing within their own constraints. Dinari, Ondo, Backed, Superstate Opening Bell on the public side. Securitize, INX, Centrifuge, Ondo's USDY and OUSG products on the private side. The platforms don't talk to each other. The APIs don't talk to each other. The settlement rails don't talk to each other.
I think this bifurcation is path-dependent, not principled. The end state for tokenized capital markets is a single API surface for any security a fintech wants to expose to its users. Public or private. Listed or unlisted. Equity or credit. US or non-US.
The public-equity tokenizers anchored on the broker-dealer integration. To tokenize Apple, you need a custodian holding actual Apple shares. The natural starting point is Alpaca's Broker API. Once you've built that, the entire stack is shaped by the public-equity workflow. Order routing. Fractional handling. Dividend pass-through. Corporate actions.
The private-market tokenizers anchored on the transfer-agent integration. To tokenize a fund interest, you need a Form D-eligible offering, accredited-investor verification, transfer-agent records, and lock-up handling. The natural starting point is a transfer-agent platform like Securitize. Once you've built that, the entire stack is shaped by the Reg D workflow. Issuance. Investor onboarding. Lock-up and resale logic.
These are not the same stack. Different custodians, different KYC graphs, different ledger designs, different redemption flows. A fintech that wants to offer tokenized exposure to both has historically had to integrate two separate vendors, run two separate compliance programs, and reconcile two separate ledgers internally. Only the largest can afford to do both. Most users see one or the other, not both.
A unified capital-markets SDK isn't one call that magically works for both sides. It's a thin developer surface over multiple underlying integration models, with the abstraction designed so the developer doesn't have to know which is which.
The mint endpoint takes a security identifier. Behind it, three different paths. Public US equity routes through a prime broker (IB or Alpaca), custodies the share, issues the token 1:1, settles the broker-to-exchange leg T+1 against DTCC while the user-facing token is live immediately. Tokenized fund interest routes through a transfer-agent integration, verifies investor eligibility, records the subscription, mints the token against the cap-table position, applies any lock-up logic. Tokenized treasury or money-market exposure routes through the underlying issuer's primary subscription rail, mints the token, and applies the relevant yield-distribution logic.
Done well, the developer integrates once and supports the union of all three paths. Done poorly, it's a leaky abstraction where the developer still has to know whether they're dealing with a public stock or a private fund. Most attempts at this so far have been the latter, in my experience.
Three things are converging that make unification possible in 2026 in a way it wasn't in 2024.
The regulatory infrastructure is catching up. The GENIUS Act of July 2025 codified stablecoin treatment for the first time. The Clarity Act, expected in 2026, will narrow the perimeter on tokenized securities. Both make it easier to build a stack that handles public and private market tokens in one ledger.
The broker-dealer infrastructure has matured. Alpaca's ~94% market share in tokenized public equities (Alpaca, December 2025) and Dinari's SEC-registered broker-dealer subsidiary mean the picks-and-shovels for the public side are mature enough to abstract over. Same on the transfer-agent side, with Securitize at $3B+ tokenized AUM, including BlackRock's BUIDL fund (~$2.6B as of April 2026, RWA.xyz).
User demand is pulling toward unified. Fintechs in MENA, LatAm, and SEA building wealth products for retail want the user to hold AAPL, OUSG, and a tokenized private credit fund in the same wallet, with the same UX. Asking the fintech to integrate three vendors to deliver that is a tax that gets arbitraged away in a competitive market.
Tokenized public equities and tokenized private fund interests live in different regulatory regimes, with different custodial requirements, different redemption mechanics, and different settlement timelines. A unified API surface doesn't erase those differences. It absorbs them.
The broker-to-exchange leg of a public equity transaction settles T+1 against DTCC. Redemption of a tokenized private fund interest may have a 30-day window. A tokenized treasury position may have intraday redemption with a yield accrual that updates every block. Each underlying has its own physics. The API hides the integration complexity, not the settlement complexity.
The vendors who survive are the ones whose marketing language matches what their architecture can actually deliver.
I might be drawing the public-private line in the wrong place. The category is moving fast and the abstraction that's right today may not be right in twelve months. But the São Paulo team I talked to last month shouldn't have had to integrate two vendors to ship one wealth product. That's the gap I'm building against.
Note on figures: the Securitize AUM number reflects public disclosures as of writing and includes BUIDL (~$2.6B as of April 2026 per RWA.xyz). Verify against current sources before quoting.
