Robinhood Launches a Blockchain, Loblaw Buys into a Bank and Singapore Publishes an AI Agent Governance Framework

Three stories from the past week ask the same underlying question in three very different settings: who stays in control when money starts moving through AI agents, blockchains and cross-ownership deals rather than familiar human-led channels. Robinhood used a London keynote to launch its own public blockchain and a suite of AI-driven trading tools. Canada's EQB completed a landmark acquisition that makes grocery giant Loblaw a near 20 percent shareholder in a bank. And Singapore's financial regulator published a governance framework for AI agents built with input from Mastercard, Visa and OCBC. Here is what is happening, and why it matters for merchants, partners and financial institutions.

Robinhood Launches Robinhood Chain and AI Trading Tools in London

On 1 July, Robinhood used a keynote called “Robinhood Presents: The World is Flat,” streamed live from the Old Royal Naval College in London, to unveil one of the broadest product and international expansion pushes in the company's history. The centrepiece was the public mainnet launch of Robinhood Chain, a Layer 2 blockchain built on Arbitrum and designed natively for tokenised real-world assets and decentralised finance, first tested publicly in February at Consensus Hong Kong. Day-one ecosystem partners include Uniswap, which is deploying a dedicated automated market maker, alongside Alchemy, BitGo and Chainlink for custody, developer tooling and oracle services. Robinhood also rolled out Stock Tokens through its self-custody Wallet app in more than 120 countries outside the US, structured as tokenised debt securities that track underlying share prices and can be traded around the clock, deposited into DeFi lending pools, or used as collateral.

Alongside the blockchain launch, Robinhood introduced Robinhood Earn, a self-custody lending product letting US users lend dollar-backed USDG stablecoin at an estimated 7 percent annual yield through the Morpho protocol, insured against cyber and smart-contract exploits through Lloyd's of London and RELM. On the AI side, the company is extending its Agentic Trading feature, previously limited to US equities and options, into crypto: eligible traders will be able to connect a third-party AI model of their choice to Robinhood's Trading MCP to scan market data and execute strategies automatically, while retaining control over capital allocation and safety guardrails. To underline the moment, Robinhood secured an official Guinness World Records title for the most items purchased by an AI agent in three minutes using its Agentic Credit Card, certified on-site at the London event.

The company also confirmed UK crypto trading plans, an official Canadian launch following its WonderFi acquisition, and that Robinhood Singapore has received a capital markets services licence from the Monetary Authority of Singapore. “Decentralized finance unlocks possibilities beyond what traditional finance can offer, but historically, it has required technical expertise to navigate,” said Johann Kerbrat, Robinhood's SVP and General Manager of Crypto and International. “We're bringing the best of traditional finance and DeFi together.” Robinhood now serves nearly 28 million customers across 38 countries.

Why it matters: Robinhood's announcements bundle three distinct trends, tokenised equities, self-custody lending and AI agents wired directly into a retail brokerage's execution layer, into one live, production platform rather than three separate pilots. As one Curve Finance founder observed of the launch, tokenisation genuinely widens access to financial products, but it does not automatically confer ownership, and the legal and product infrastructure will need to mature alongside the technology. Expect regulators in the UK, EU and across Asia to scrutinise how tokenised securities and AI-driven trading interact with existing investor protection rules, even as Robinhood pushes deeper into markets like Singapore that have already built the regulatory rails to support this kind of product.

The Debia angle: Robinhood's launch is a vivid illustration of a pattern Debia watches closely: the fastest-growing edge of finance now sits at the intersection of tokenisation, self-custody and AI-driven execution, all stitched onto a single retail-facing platform. The lesson for payment infrastructure providers is that consumers increasingly expect this level of integration as standard, not as a novelty. Merchants and partners should think of blockchain settlement, stablecoin lending and AI-agent commerce not as three separate roadmap items but as one connected capability that the underlying infrastructure needs to support cleanly and safely.

Loblaw Becomes a Near 20% Shareholder in EQB After PC Financial Deal

On 1 July, Canada's EQB Inc., parent of digital challenger EQ Bank, completed its acquisition of PC Financial from Loblaw Companies Limited, first announced in December 2025 and cleared by Canada's Competition Bureau and Federal Minister of Finance earlier this year. The deal folds President's Choice Bank, PC Financial Insurance Agency and related entities into Equitable Bank, EQB's wholly owned subsidiary and Canada's seventh-largest Schedule I bank by assets, immediately expanding EQB's reach to nearly four million Canadian customers. Consideration was structured at 1.15 times PC Financial's book value: Loblaw received 7.2 million newly issued EQB common shares plus 234.5 million dollars in cash, against an implied purchase price of roughly 800 million dollars when the deal was first announced.

The transaction reshapes EQB's ownership structure as much as its customer base. Following completion, Loblaw beneficially owns approximately 8.46 million EQB shares, representing close to 20 percent of shares outstanding, with an investor rights agreement permitting Loblaw to acquire up to 25 percent over time. Two Loblaw nominees, chairman and CEO Galen G. Weston and CFO Richard Dufresne, joined EQB's board as part of the arrangement. EQB also becomes the exclusive financial services partner for PC Optimum, Loblaw's loyalty programme with more than 18 million active members, giving the bank direct access to PC Financial's roughly 180 in-store banking pavilions and 600-plus ATMs across Loblaw retail locations. EQB expects the deal to be mid-single-digit accretive to adjusted earnings per share in its first full year, with about 30 million dollars in annual run-rate cost synergies against 105 million dollars in one-time integration costs. “Today, as we celebrate Canada Day, we mark a turning point for Canadian banking,” said EQB President and CEO Chadwick Westlake.

Why it matters: This deal creates one of the clearest loyalty-linked banking ecosystems in North America, tying everyday grocery spending directly to banking products and rewards for millions of customers, while giving a major retailer meaningful equity influence over a regulated bank rather than a simple commercial partnership. It signals that the boundary between retail loyalty programmes and core banking infrastructure is becoming genuinely porous, not just cosmetically branded. Expect other large retailers with substantial loyalty ecosystems, particularly in grocery and pharmacy, to examine similar deep-equity banking partnerships as a way to convert everyday purchase data into a durable financial services relationship.

The Debia angle: The EQB-PC Financial deal is a strong example of a theme Debia sees accelerating across markets: the most durable financial relationships increasingly grow out of everyday, high-frequency spending rather than banking products marketed on their own. Loblaw did not build a bank; it bought meaningful ownership of one that already had regulatory infrastructure, digital capability and rewards flywheel. For merchants and platforms with strong loyalty ecosystems of their own, the practical lesson is that payment infrastructure and rewards infrastructure work best when treated as a single connected product, so that a customer's everyday purchases and their financial relationship reinforce each other rather than living in separate systems.

MAS Publishes SAFR, a Governance Framework for AI Agents in Finance

On 3 July, the Monetary Authority of Singapore published an industry white paper called Safeguards for Agentic Finance at Runtime, or SAFR, developed jointly with leading financial institutions and fintechs under MAS's BuildFin.ai initiative. The framework addresses a problem that has become urgent as AI agents take on more autonomous work in finance: existing governance processes were largely designed around human decision-making and are not built for systems that can propose and execute financial actions at machine speed and scale, well beyond what practical human intervention can keep pace with. SAFR introduces governance checkpoints that verify and record an AI agent's proposed action, whether initiating a payment, submitting a trading order, approving a credit application, filing a regulatory report or settling an insurance claim, before that action is allowed to execute, built around four pillars: policy-bound execution, real-time validation, auditability and interoperability.

The paper builds on MAS's earlier Project Mindforge AI Risk Management toolkit, focusing specifically on how safeguards can be operationalised at the exact point an AI agent takes action rather than only through model-level guardrails, which the paper notes are necessary but insufficient, since an agent's outputs can still be well-formed while proposing an action it was never authorised to take. Case studies included in the white paper, contributed by Mastercard, Ant International, Visa, Circle, OCBC, Bank of Singapore and Manulife, cover use cases including agent-assisted payments and treasury operations, wealth management and advisory document review, and client engagement drafting, all executed within narrowly scoped, pre-approved task boundaries. MAS has explicitly framed SAFR as an industry reference model rather than binding regulatory guidance, inviting further industry partners to join the BuildFin.ai working group to shape future iterations, with the newly announced Future of Finance Institute set to support pilot testing and sandbox experimentation of SAFR-aligned solutions.

Why it matters: SAFR is one of the first practical, publicly documented frameworks anywhere for governing AI agents at the exact moment they take financial action, rather than only at the model or policy level. With participation from card networks, custodians and banks already spanning payments, wealth management and treasury, it gives Singapore's financial sector a shared vocabulary and technical reference point just as agentic commerce and agent-initiated payments move from pilot to production globally, including through initiatives like Visa Intelligent Commerce and Mastercard Agent Pay. Expect other regulators, particularly in Hong Kong, the UK and the US, to study SAFR closely as they develop their own approaches to agentic AI oversight in finance.

The Debia angle: SAFR reflects exactly the kind of infrastructure-first thinking Debia applies to payments: safeguards that live inside the transaction flow itself, checked and logged at the point of action, are far more durable than governance that exists only in a policy document. As AI agents increasingly initiate payments and financial actions on behalf of consumers and businesses across the region, the institutions and infrastructure providers that can demonstrate real-time, auditable control over what an agent is permitted to do will be the ones merchants and partners trust with that traffic. Frameworks like SAFR give the whole ecosystem, including providers like Debia, a clearer target to build toward as agentic commerce becomes mainstream across Southeast Asia.

At Debia, we track these changes because the future of payments will be shaped by speed, trust, interoperability, and smarter financial infrastructure. We do not just process payments. We understand the infrastructure, regulation, technology, and market shifts behind the future of digital commerce, and we build for where the ecosystem is heading next.