🏆 Industry Awards — 98.5% cost cut per loan approval! The intelligent anti-fraud system jointly developed by China Everbright Bank & Tencent Cloud won @TheAsianBanker 2026 Best Risk Data and Analytics Technology Initiative in Asia Pacific. An AI-powered anti-fraud system with real-time decisioning across retail lending — delivering great precision, efficiency, and scalability throughout the credit approval process. ✅ Precision: 3x+ risk control precision, over 50M customers approved to date ⚡ Speed: Seconds-level approval, down from days via automated decisioning 🔒 Scale: More than 600 anti-fraud indicators covering 30+ credit products 🌐 Impact: Over 10,000 high-risk applications blocked per year — hundreds of millions in fraud losses prevented Overall efficiency improved by more than 100x, while loan approval cost per application was reduced by 98.5% 🔗 Click here to explore more innovation: #TencentCloud# #TencentCloudRiskSolutions# #TABGlobal# #FinTech# #FraudPrevention# #RiskManagement# #DigitalBanking# #AIInFinance#

The regulatory framing on fraud is shifting under most institutions' feet. Omri David, Senior Product Manager for Fraud Prevention at Hypernative, fielded a Q&A question on this during our recent onchain fraud prevention webinar. 🔹 MiCA in Europe carries explicit duty-of-care and operational resilience provisions. 🔹 The UK's authorized push payment fraud reimbursement rules are setting a precedent likely to extend to crypto. 🔹 FinCEN's actions against Huione signal the same direction in the US. Pre-transaction fraud screening is not yet universally mandated, but regulators are increasingly treating the failure to prevent fraud flows as a compliance failure rather than a user protection issue. We discussed this and much more in our recent webinar on fraud prevention. Watch the full session on demand:

When Telegram shut down Huione's channels in May 2025, scam transaction volume didn't drop. It migrated within days to a related platform called Tudou Guarantee, with the same controlling parties. Alex Rabke, VP Americas at Hypernative, used the Huione case to make a structural point during our recent webinar on the onchain fraud prevention blueprint. The platform was a $27B criminal enterprise, larger than Silk Road, but the takedown only affected the venue. The infrastructure built to support thousands of fraud vendors, money laundering services, and scam-site-as-a-service kits kept operating. This is just one part of a deep dive into fraud prevention put out recently by our security team. Watch the full session on demand:

Mastercard Q2 2026: what the network's report tells the private market @Mastercard quarterly report gets read as a stock idea. For a fund, that is the least useful way to read it. A public incumbent doing $9.3B in quarterly revenue is not a ticker, it is the best available dataset on which private companies someone is willing to buy, at what price, and in which layer of the stack margin actually exists. There are four such signals in this report, and none of them is about EPS. The category ceiling has moved On March 17, Mastercard agreed to acquire @BVNKFinance for $1.8B - $1.5B plus up to $300M in contingent consideration. It is the largest stablecoin infrastructure acquisition on record. The previous marker, @stripe $1.1B purchase of Bridge (@Stablecoin), held as the category ceiling for two years. For anyone holding a position in payment infrastructure, this repriced the exit math. BVNK processes over $30B in payments annually, which means the comp is a multiple of volume rather than revenue, and that is the anchor both founders and buyers will negotiate against in the next round. The ceiling rose 60% in two years, in a category where the existence of exits at all was in question not long ago. There are fewer buyers than it looks The second piece of news matters more than the first, and almost nobody picked it up. In May, Mastercard walked away from a minority investment in @zerohashx, a company it had earlier discussed acquiring outright for as much as $2B . @zerohashx went out to raise on its own instead, above a $1.5B valuation. Read that as an architecture decision. Mastercard chose a single integrated stack over a portfolio of bets on competing providers. The practical consequence for a fund: this category just lost a strategic buyer, and that buyer has already spent its budget. The first company in the segment exits at $1.8B, the second raises, and the third and fourth compete for an acquirer who is no longer in the market. Exit concentration risk here is higher than the size of the market suggests. The margin is not in the rail Mastercard's payment network grows 8-12% currency-neutral. Value-added services - scoring, authentication, fraud prevention, data - grew 22% year over year in Q1, organically, and now account for roughly 40% of net revenue. So an incumbent running a 58% operating margin is showing you where its money is while simultaneously paying $1.8B not to build settlement itself. Both facts point the same direction. Settlement is being commoditized, and it is being bought. Software sold on top of the traffic is not. If your pipeline holds another cheaper, faster rail, that is a company acquired for its volume in the best case. A company selling risk data and compliance on top of someone else's rails gets acquired for its revenue. The authorization window is open, and closing Agent Pay is now enabled on essentially every Mastercard card globally, with Verifiable Intent layered on top as a tamper-resistant record of user authorization, plus a Crossmint partnership for blockchain execution. An agentic payment creates a problem of consent, not settlement: who authorized what, exactly, and how do you prove it in a dispute six months later. A stablecoin rail does not address that question at all. Mastercard is entering agentic commerce through authorization rather than through the transfer, and doing it with distribution across billions of cards. This is the most interesting open layer in payments right now and the one closing fastest. What to take from it Public markets are paying for the transformation: 25.8x forward earnings against 18.6x for the industry. That is the discount rate underneath the whole thesis - incumbents hold expensive paper and have every reason to buy infrastructure with it. So the useful question after this report is not whether to own $MA . It is which layer of the stack @Mastercard pays for next instead of building. This year the answer was settlement. Judging by Agent Pay, next time it will not be.