BNPL Just Walked Into the Credit Bureau. Dubai Walked Into the Secondary Market.
A quiet October rewired two corners of credit: short-term consumer loans finally start showing up on a lender's screen, and tokenised real estate graduates from pilot into tradable product.
As of this week, the lending stack looks different from how it did at the start of the month. Credit infrastructure is absorbing buy now, pay later data in a way bureaus have been promising since 2022. Dubai has turned its real estate tokenisation experiment into something that resembles an actual exchange. And a handful of underwriting startups in Riyadh and London are quietly proving that the next underwriter is a model, not a human.
Here is a snapshot of what moved, where the money went, and why your credit file is about to get noisier.
BNPL finally shows up on the lender's screen
For years, the gap in the credit file has been staring everyone in the face. A consumer with four open Klarna plans, three Afterpay installments, and a maxed Affirm balance could still walk into a loan interview looking debt-free. That gap is why every risk officer at a regional bank has been circling TransUnion with questions.
As of October, TransUnion has begun making a limited set of BNPL and point-of-sale attributes available to lender partners. According to TransUnion's own blog, those attributes are available to a single lender partner today and do not currently affect consumer credit scores. A fuller version, with tradeline-level detail, is targeted for 2027.
In parallel, FICO has introduced FICO Score 10 BNPL and FICO Score 10T BNPL, two scores designed to pull short-term installment behaviour directly into the primary credit grade. CardRates reported the launch as the clearest signal yet that issuers can no longer price risk off the old file alone.
The holdouts matter. PaymentsJournal has reported that Klarna and Afterpay declined to feed data into the earlier credit-scoring experiments. No public update in 2026 has reversed that. For risk teams, that means the headline number from FICO still reflects partial coverage, which is exactly why TransUnion's bureau-level approach is more consequential than it looks.
Why it matters
One clear read: 2027 is the year the invisible debt stack becomes visible. Lenders that have been pricing SME overdrafts and credit cards without knowing a borrower's BNPL exposure can finally underwrite the full liability picture. Expect delinquency models to be rebuilt, expect approval rates on near-prime borrowers to wobble for two quarters, and expect at least one BNPL provider to make peace with the bureaus once the pricing penalty for staying out becomes obvious.
Embedded lending gets its SME moment
While the consumer side is being re-plumbed, the SME side is being rebuilt by a different cast of characters. Several are new names, most are infrastructure, and all of them are AI-native.
Abwab.ai, a Riyadh-based credit engine, raised a $4 million seed round in September 2026 co-led by Speedinvest and Middle East Venture Partners. According to Renascence, the platform pulls bank statements, VAT records, and bureau data into a single underwriting view, and is already used by more than 13 financial institutions. The company says it has processed over SAR 1 billion in SME loans. For a region where SME credit penetration has lagged consumer credit by a decade, that is not just a seed round, it is a template.
The UK side has been louder. Allica Bank closed a $155 million Series D earlier this year, per FintechFutures, valuing the digital SME lender at roughly $1.2 billion and funding its first international expansion. On the embedded side, Equipifi raised a $34 million Series B led by Left Lane Capital in May 2026 to extend bank-native BNPL into credit unions, as reported by Fintech Global.
The underlying bet
Each of these raises is a vote for the same thesis. SMEs do not want a loan product, they want a credit line that lives inside the software they already use (accounting, payroll, procurement). The embedded lending category, which Dealroom's newsletter now puts at over $2.3 billion deployed between 2025 and early 2026, is quickly becoming the main way working capital gets delivered to a sub-$10 million business.
Dubai turns tokenised real estate into a market
PropTech's biggest story this autumn is not a flashy app launch, it is market plumbing. Dubai's Real Estate Tokenisation Project moved into Phase II in early 2026, as reported by Zawya, and the follow-through this quarter has been striking. Investors can now resell previously issued real estate tokens, which converts a pilot into a secondary market.
PRYPCO Mint, the regulated marketplace built in partnership with the Dubai Land Department, said its first phase attracted investors from more than 50 nationalities, per KuCoin's news desk. That matters because the problem with real estate tokenisation has never been issuance, it has been exit. Without liquidity, a fractional title is just a dressed-up spreadsheet. Dubai is now trying to answer the exit question in regulation, not in press release.
Elsewhere, SettleMint is quietly building national blockchain infrastructure for property registries in Saudi Arabia. According to FF News, four Saudi PropTechs (Sahl, Madek, Ghanem, and Jozo) ran live tokenisation transactions in a controlled environment at the start of 2026. On the standards side, Tokeny's ERC-3643 standard, which embeds compliance rules directly into the token, has become the de facto rail for European issuers.
The watch item
If Dubai's secondary market records real volume this quarter, expect London, Paris, and Zurich to accelerate their own regulatory sandboxes. The UK already has a Dubai-linked experiment running: an April 2026 Pulse report described the first regulated tokenised build-to-rent offering in the UK, launched by Tokinvest. Call it 2026's answer to REITs: smaller tickets, programmable rules, and (for the first time) a plausible way out.
The three threads to pull
Pulling the three stories together, a pattern shows up. Lending is becoming a data problem (bureau coverage), a distribution problem (embedded rails), and a liquidity problem (tokenised collateral). The surprise in October is not that any one of these moved, it is that they all moved in the same month.
For bank risk teams: the FICO change rewrites the near-prime underwriting playbook before year-end models get locked for 2027. For fintech founders: the embedded SME lending category has deep-pocketed infrastructure buyers for the first time, and defensibility now sits in the credit engine, not the UI. For asset managers: Dubai's secondary market is the first real test of whether tokenised real estate can carry an institutional allocation. The right question is not whether to issue, it is whether you can underwrite the exit.
One more thing worth saying out loud. Every one of these shifts, bureau coverage, embedded SME, tokenised property, pushes credit decisions further away from a human analyst and closer to a model. Governance teams who treated model risk as a 2027 problem now have an uncomfortable 2026 to-do list.
