Blockchain, AI, and Fintech in 2026: What Has Actually Changed for Businesses

Trend reports on blockchain, AI, and fintech tend to quote enormous projected numbers and treat every pilot as a revolution. This guide sticks to what has measurably changed by September 2026, which laws passed or stalled, and what a business should do about each area. Most of the change is real but narrower than the headlines, and much of it comes down to paperwork: accounting rules, tax forms, and compliance obligations.

Stablecoins: large, regulated, and still mostly used inside crypto

About $300 billion of dollar stablecoins were in circulation in September 2026, per trackers such as RWA.xyz, with Tether's USDT around 60% of the total and USDC most of the rest. The GENIUS Act, signed in July 2025, set the first US federal framework for payment stablecoins: one-to-one reserves in cash and short-term Treasuries, regular reserve disclosure, and a ban on issuers paying interest to holders. Regulators were still writing the implementing rules in 2026.

Mainstream payment companies now accept them. Stripe and PayPal both charge merchants 1.5% for stablecoin or crypto payments that settle in dollars, and Shopify offers USDC at its standard rate. Actual checkout use is still small; our guide to crypto payment and finance tools has the details and fees.

Cross-border business payments are where stablecoins make the most sense. The World Bank's Remittance Prices Worldwide put the global average cost of sending $200 at about 6.4% in late 2025, with banks the most expensive channel. Larger business payments cost less as a percentage but still carry wire fees and exchange spreads.

An illustration for a $50,000 payment to a supplier abroad:

The blockchain part is almost free. The conversions at each end are where the money goes, and in this example a good money transfer service is still cheapest. Stablecoins win when both sides already hold them, when the recipient's local banking is slow or unreliable, or when payments need to settle outside banking hours. Get quotes for your own corridor before assuming savings. For the exchange-rate risk itself, see our currency hedging guide.

Tokenization: real in money market funds, early elsewhere

Tokenized real-world assets excluding stablecoins reached about $38 billion by September 2026 per RWA.xyz, up from under $5 billion at the 2024 low. Roughly $15 billion of that was tokenized US Treasury and money market funds such as BlackRock's BUIDL and Franklin Templeton's BENJI. Commodities (mostly tokenized gold), private credit, and tokenized stocks made up most of the rest.

What works: tokenized money market funds are used as collateral and cash management inside crypto markets, where they settle around the clock. What is still early: tokenized real estate and private equity. The token may trade on-chain, but transfer restrictions, investor eligibility rules, and thin secondary markets mean the "instant liquidity" pitch rarely holds up. If you invest this way, the questions from our private real estate funds guide and real estate crowdfunding guide still apply: who holds legal title, what happens if the platform fails, and who will buy your tokens.

The US legal framework is unsettled. The CLARITY Act, which would divide oversight of crypto markets between the SEC and CFTC, passed the House 294 to 134 in July 2025, but a Senate motion to proceed failed 49 to 50 on September 15, 2026, largely over ethics provisions. In the meantime the SEC and CFTC are setting policy through exemptions and staff guidance, which can change with leadership. In the EU, MiCA has applied in full since December 2024, and transition periods for existing crypto service providers ended by July 1, 2026.

AI in financial services

Adoption is broad and returns are uneven. McKinsey's State of AI survey from November 2025 found 88% of organizations using AI in at least one function but only 39% reporting any effect on EBIT. In finance, the clearest uses are:

• Fraud detection, where machine learning has been standard for years; see our AI fraud detection guide.
• Forecasting, where models help only if they beat a simple baseline; see AI financial forecasting.
• Compliance work, such as alert triage and regulatory change tracking; see AI regulatory compliance.

Rules have not loosened because a model is involved. In the US, ECOA and Regulation B require lenders to give specific reasons for denying credit, and a complex model does not excuse vague explanations. Banks' model risk management expectations under the Federal Reserve's SR 11-7 apply to machine learning models as much as older ones. In the EU, AI used to assess individuals' creditworthiness is high-risk under the AI Act, and after the 2026 Digital Omnibus those obligations apply from December 2, 2027. Our AI governance guide covers how to organize this.

AI has also made payment fraud easier. In 2024 an employee of the engineering firm Arup in Hong Kong transferred about US$25 million after a video call in which the "executives" were deepfakes. IBM's 2026 Cost of a Data Breach report found that about one in four malicious breaches involved AI, often impersonation. The fix is procedural: verify any unusual payment request through a separately sourced phone number, and require two people to approve new payees and bank detail changes.

Open banking and banking-as-a-service

The CFPB's 2024 open banking rule under Section 1033 was supposed to start applying to the largest banks in April 2026. A federal court in Kentucky enjoined it in October 2025, and the CFPB sent a rewritten proposal to the White House for review in August 2026. The main fight is whether banks can charge data aggregators for access. Until a new rule is final, data access runs on bilateral agreements between banks and aggregators, some of which now include fees.

Embedded finance also carries partner risk. When the middleware company Synapse failed in April 2024, customers of several fintech apps found their funds frozen for months and a shortfall between what the partner banks held and what customers were owed. If your business holds money in a fintech account, find out which bank holds the funds, whether the account is titled so that pass-through FDIC insurance applies, and who keeps the ledger.

Security: signing, keys, and the quantum timeline

The largest crypto theft on record, about $1.5 billion from the exchange Bybit in February 2025, came from a compromised signing interface, not broken cryptography. Signers approved a transaction that looked routine. For any business holding digital assets, independent transaction verification matters more than the choice of wallet technology. Our zero trust guide covers the wider access-control picture.

Quantum computing is a planning issue. NIST published its first post-quantum cryptography standards (FIPS 203, 204, and 205) in August 2024 and has proposed deprecating today's RSA and elliptic-curve algorithms by 2030 and disallowing them by 2035. Bitcoin and Ethereum signatures use elliptic-curve cryptography, and both communities are debating migration paths. For a business, the first step is an inventory of where you use public-key cryptography and which data must stay confidential for ten years or more.

Central bank digital currencies

In January 2025 a US executive order barred federal agencies from establishing or promoting a CBDC. The European Central Bank decided in October 2025 to move its digital euro project to the next preparation phase, with a possible first issuance around 2029 if EU legislation is adopted. For most businesses this is a watch item, not a planning item.

What to do, by type of business

• Merchants. Offer stablecoin checkout only if customers ask for it. Fees are lower than cards, there are no chargebacks, and refunds are manual.
• Companies paying suppliers abroad. Get all-in quotes for your corridors from your bank, a money transfer service, and a stablecoin payout provider, and compare.
• Treasury teams. Tokenized money market funds are the most established use of tokenization, but for most companies an ordinary money market fund does the same job with fewer operational questions.
• Lenders and fintechs. Budget for model documentation, adverse action explanations, and, in the EU, the AI Act's high-risk requirements before December 2027.
• Everyone. Tighten payment approval procedures against deepfake requests, and start a cryptography inventory for post-quantum migration.

This article is for informational purposes only and does not constitute financial, investment, legal, or technology advice. Figures and regulatory status reflect September 2026 and change quickly; check current sources before making decisions.