The eye-catching US$2 trillion projection for the potential growth of the stablecoin market that was often cited during the recent push to approve US regulation of the crypto market for the first time is “a little bit optimistic”, according to JPMorgan Chase. JPMorgan strategists wrote in a note to clients, per the report: “We find it hard to believe that the market could grow substantially larger over the next few years as the infrastructure/ecosystem that supports stablecoins is far from developed and will take time to build out. While adoption is poised to grow further, it might be at a slower pace than what some might anticipate.” despite the recent increase in interest for the assets, they still only account for less than 1 per cent of global money flows, suggesting the role of the digital asset in upending financial rails still has significant steps to take, the strategists wrote. JPMorgan pointed out that given the current growth trajectory, it was more likely that the market would double or triple, which is far lower than other estimates. “We suspect liquidity investors, whether retail or institutional, are not going to immediately jump into payment stablecoins as a cash alternative given their conservative nature in terms of how they manage their cash as a source of liquidity,” the bank said.
A multi-round AI coding challenge K Prize that tests models against flagged issues from GitHub to assess how well models can deal with real-world programming problems sees a top score of just 7.5% versus the industry benchmark of 75%
Nonprofit Laude Institute announced the first winner of the K Prize, a multi-round AI coding challenge launched by Databricks and Perplexity co-founder Andy Konwinski. The winner will receive $50,000 for the prize. The final score set a new bar for AI-powered software engineers; with correct answers to just 7.5% of the questions on the test. K Prize runs offline with limited compute, so it favors smaller and open models. It levels the playing field. Konwinski has pledged $1 million to the first open source model that can score higher than 90% on the test. K Prize tests models against flagged issues from GitHub as a test of how well models can deal with real-world programming problems. But while SWE-Bench is based on a fixed set of problems that models can train against, the K Prize is designed as a “contamination-free version of SWE-Bench,” using a timed entry system to guard against any benchmark-specific training. For round one, models were due by March 12. The K Prize organizers then built the test using only GitHub issues flagged after that date. The 7.5% top score stands in marked contrast to SWE-Bench itself, which currently shows a 75% top score on its easier “Verified” test and 34% on its harder “Full” test. Konwinski still isn’t sure whether the disparity is due to contamination on SWE-Bench or just the challenge of collecting new issues from GitHub, but he expects the K Prize project to answer the question soon.
Anthropic study finds all LLMs showed “performance degradation with extended reasoning” on complex deductive tasks, “suggesting difficulties in maintaining focus during complex deductive tasks
Artificial intelligence models that spend more time “thinking” through problems don’t always perform better — and in some cases, they get significantly worse, according to new research from Anthropic that challenges a core assumption driving the AI industry’s latest scaling efforts. The study, led by Anthropic AI safety fellow Aryo Pradipta Gema and other company researchers, identifies what they call “inverse scaling in test-time compute,” where extending the reasoning length of large language models actually deteriorates their performance across several types of tasks. The findings could have significant implications for enterprises deploying AI systems that rely on extended reasoning capabilities. The study reveals distinct failure patterns across major AI systems. Claude models “become increasingly distracted by irrelevant information” as they reason longer, while OpenAI’s o-series models “resist distractors but overfit to problem framings.” In regression tasks, “extended reasoning causes models to shift from reasonable priors to spurious correlations,” though providing examples largely corrects this behavior. Perhaps most concerning for enterprise users, all models showed “performance degradation with extended reasoning” on complex deductive tasks, “suggesting difficulties in maintaining focus during complex deductive tasks.” Major AI companies have invested heavily in “test-time compute” — allowing models more processing time to work through complex problems — as a key strategy for enhancing capabilities. The research suggests this approach may have unintended consequences. “While test-time compute scaling remains promising for improving model capabilities, it may inadvertently reinforce problematic reasoning patterns,” the authors conclude. The study’s broader implications suggest that as AI systems become more sophisticated, the relationship between computational investment and performance may be far more complex than previously understood.
More efficient financial markets, powered by unifying digital layers that operate 24/7, will free up billions of dollars in collateral held against operational risks and create new opportunities for financing and liquidity
The financial industry is rapidly transforming as blockchain technology enables real-time, global, and 24/7 markets. Major firms like BlackRock, Fidelity, and Franklin Templeton are investing in public blockchains such as Ethereum, Solana, and Avalanche to tokenize real-world assets—starting with stocks, bonds, and money market funds, and soon potentially expanding to real estate, art, and carbon credits. Unlike traditional systems, blockchain allows for instant settlement, composability, and continuous operation, eliminating time-based trading constraints and cross-border barriers. Legacy infrastructure like session-based markets and end-of-day fund valuation is being replaced by systems that enable second-by-second tracking of yield and daily payouts—even on weekends. However, not all tokenized assets offer true blockchain functionality; some merely provide digital receipts while relying on legacy systems. The real breakthrough comes when ownership and yield distribution are fully on-chain and managed via compliant digital wallets, empowering investors with peer-to-peer transfers, self-custody, and seamless access to a broad range of assets. This shift toward wallet-based finance promises to free up capital, reduce operational risk, and fundamentally reshape how global markets function. Institutions that fail to embrace this evolution risk falling behind.
Turning embedded payments into a profit center requires software platforms to participate in card economics by taking a slice of the acquiring fee or issuing virtual cards, route payments and control disbursements and monetize the payment network
Jay Dearborn, president of Corporate Payments at WEX described embedded payments maturity as a three-stage journey: enablement, customer value creation and monetization. Stage 3 is about turning embedded payments into a profit center. That’s where Dearborn said he sees the biggest gap and the biggest opportunity. “It almost takes someone to bring payments expertise onto the leadership team,” he said. Software companies are good at building great products, but payments is a domain business. Without deep expertise — or the right partner — the monetization piece rarely materializes, he said. That monetization comes in several forms. Dearborn said to start with participating in card economics, which means taking a slice of the acquiring fee, issuing virtual cards or receiving rebates through ACH+ networks. Next is participating in the funds flow itself. Companies that sit inside the accounts payable process, route payments and control disbursements aren’t just facilitators. They’re operators, and operators get paid. Then comes network monetization. Software platforms don’t just process payments; they connect buyers and sellers. That connection creates a closed-loop network that can be optimized, priced and monetized — without ever touching the payment itself, he said. That’s not the only monetizable layer. Speed matters too. Instant access to funds — on either side of a transaction — creates tangible value. Buyers are more likely to pay early. Sellers get paid faster. That value can be priced, packaged and monetized. “You can build an ecosystem around that account, that payment credential,” Dearborn said. “That’s the next unlock.” Embedded payments are table stakes, he said. Embedding them well and building a business around them is still a competitive edge.
The passage of the GENIUS Act to spur the supply of stablecoins by a “relatively modest” $25 billion-$75 billion in the near term, driven by product rollouts, infrastructure investment and competition from tokenized deposits and money market funds
Bank of America says the GENIUS Act, signed into law last Friday by President Donald Trump, marks a turning point for U.S. stablecoin regulation, laying the groundwork for infrastructure development and tokenized finance growth. Supply of stablecoins, crypto tokens whose value is pegged to real world assets such as fiat currencies or gold, will grow a “relatively modest” $25 billion-$75 billion in the near term, driven by product rollouts, infrastructure investment and competition from tokenized deposits and money market funds, the bank said. The total market cap for stablecoins is currently about $270 billion, according to CoinMarketCap data. Over the next 2–3 years, the bank’s analysts foresee stablecoin consolidation and broader adoption of these cryptocurrencies and other tokenized assets, supported by the enactment of the CLARITY Act. That act aims to establish a clear regulatory framework for digital assets in the U.S., distinguishing cryptocurrencies as either commodities or securities. The legislation has been passed by the House of Representatives and will now be considered by the Senate. Banks appear ready to issue their own stablecoins, with management teams leaning toward consortium-led models, including BofA. While cross-border use cases are gaining traction, most bank executives do not expect near-term disruption to domestic payments
J.D. Power’s Mortgage Servicer Satisfaction Study | Rocket Mortgage ranks highest; Guild Mortgage ranks second and Regions Mortgage ranks third
With the average 30-year mortgage rate in the United States continuing to hover near recent highs of 6.8%,1 homeowners might be expected to feel some tension with their mortgage company. However, high rates alone do not explain why customer satisfaction scores for mortgage servicers are significantly lower—and declining—than they are for mortgage originators. According to the J.D. Power 2025 U.S. Mortgage Servicer Satisfaction Study, customer satisfaction with mortgage servicers has plummeted in 2025, with an average satisfaction score that is now 131 points (on a 1,000-point scale) lower than the average score for mortgage originators. Increasingly, the difference between the two comes down to effective communication and customer service. Rocket Mortgage ranks highest among mortgage servicers with a score of 685. Guild Mortgage (677) ranks second and Regions Mortgage (656) ranks third. Key findings of the 2025 study:
- A fragmented customer journey: Overall customer satisfaction with mortgage servicers is 596, which is down 10 points from the 2024 study. Customer satisfaction with mortgage servicers declines across all dimensions year over year. This decline stands in stark contrast to customer satisfaction with mortgage originators, which reached a score of 727 in the J.D. Power 2024 U.S. Mortgage Origination Satisfaction Study.SM
- Service quality and responsiveness play major role in customer loyalty: While better interest rates and lower costs and fees are cited most frequently by customers as a reason to switch mortgage providers, service quality and responsiveness can be equally powerful drivers of customer loyalty and retention. Customers cite better/improved customer service (51%); easy access to loan information (36%); and flexible ways to make a mortgage payment (27%) among the top reasons to switch mortgage companies.
- Communication breakdown: Despite industry efforts to deliver more effective communications, just 31% of mortgage servicer customers gave an excellent or perfect rating to their servicer for messaging that got their attention. Attention-getting is rated higher when there is a level of personalization added to the communication. Among those who have received personalized communications, account alerts are the most frequently recalled form of communication at 46%. Just 32% of customers give their mortgage servicer a high overall communication rating, down 5 percentage points from 2022.
- Satisfaction decreases as escrow costs rise: Escrow costs—the fees typically rolled into a mortgage to pay annual property tax and homeowners insurance bills—are rising nationwide, with 57% of mortgage servicer customers experiencing an increase in escrow costs this year. Overall satisfaction is 67 points lower, on average, among those who experienced an escrow cost increase than among those who experienced no change.
By integrating Discover’s direct savings bank with Capital One’s “Digital First” model, the combined entity can now rival the “Big Four” banks, unlocking a critical advantage- Discover’s exemption from the Durbin Amendment
Capital One’s $35.3B acquisition of Discover Financial creates a top-2 U.S. credit card player with 19% loan share and vertical integration in issuing/processing. By combining Discover’s proprietary payment networks (PULSE and Diners Club) with Capital One’s data-driven underwriting, the merged entity now holds 19% of U.S. credit card loans and 22% of the customer base. That’s second only to JPMorgan Chase. But the real magic lies in vertical integration: the combined firm now owns both the issuing and processing sides of the credit card business, a rare feat in an industry dominated by intermediaries like Visa and Mastercard. This integration unlocks a critical advantage: Discover’s exemption from the Durbin Amendment. That means the merged entity can bypass interchange fee caps on debit transactions, a $1.2 trillion market. For investors, this isn’t just a line item—it’s a new revenue stream that can be reinvested into customer rewards or used to undercut competitors on pricing. The result? A playbook that challenges the status quo. The merger redefines the fintech landscape. By integrating Discover’s direct savings bank with Capital One’s “Digital First” model, the combined entity can now rival the “Big Four” banks. Discover’s 70 million global merchant acceptance points and Capital One’s cloud-based tech stack create a platform for embedded finance—think payments in SaaS platforms or fintech partnerships. American Express and JPMorgan Chase remain formidable, but the merged entity’s agility in targeting mid-tier spenders and underbanked segments is a game-changer. For example, the Discover Cashback Debit card—a rare offering in the industry—now serves 100 million customers, enabling Capital One to monetize this segment in ways competitors can’t replicate. For long-term investors, this deal is a high-stakes poker game. The short-term pain of integration costs is offset by long-term gains in market share, innovation, and financial inclusion. The key is patience.
WEX’s Benefits segment is a growth engine- revenue in Q2 rose 8.5% year over year to $195.1 million, driven by 6% growth in SaaS accounts and 11.4% growth in custodial investment income
WEX is transforming from a legacy fuel card provider into a diversified FinTech infrastructure company, now operating in three segments: Mobility, Benefits, and Corporate Payments. The mobility segment, which still accounts for about 50% of WEX’s total revenue, is navigating pressures. Same-store sales across local and over-the-road (OTR) fleets are down, reflecting both efficiency gains (i.e., fewer gallons per mile) and cautious spending by mid-market fleet operators. While the full revenue from BP’s existing card portfolio won’t hit until after its conversion — likely sometime in 2026 — WEX expects 0.5% to 1% of additional annual revenue from the deal once fully implemented. In the meantime, the company is seeing strong traction from increased investments in digital marketing aimed at small fleet operators. Historically, each dollar spent in this channel has generated $4 in revenue over two years, and early signs suggest the return profile remains intact. WEX’s Benefits segment may not make headlines, but it continues to be one of the company’s most stable growth engines. Revenue in Q2 rose 8.5% year over year to $195.1 million, driven by 6% growth in software-as-a-service (SaaS) accounts and 11.4% growth in custodial investment income. This segment — built on the complex infrastructure that powers HSAs, FSAs, and COBRA accounts — has both high margins and high stickiness. Switching providers in this space is “complex, time-consuming and disruptive,” which explains why WEX serves nearly 60% of the Fortune 1000 and powers more than 20% of the HSA market through its direct and channel partner offerings. WEX also launched a new AI-driven claims processing tool that slashes reimbursement processing from days to minutes — a rare moment where FinTech buzzwords meet real impact. The automation reduces costs while improving the user experience — an important differentiator as benefits become a battleground for attracting talent. If there’s a wildcard in WEX’s portfolio, it’s corporate payments. The segment, which includes both embedded payments (mainly virtual cards used in travel and other verticals) and accounts payable (AP) automation, saw revenue decline 11.8% to $118.3 million. WEX has increased its dedicated AP sales force by over 50% and is riding a wave of demand from mid-size and enterprise companies looking to digitize legacy payment workflows. With more than 140 new customers signed year-to-date and a record pipeline, this unit could quietly become a growth engine in its own right. Meanwhile, the company is expanding its embedded payments offering into new verticals such as media, eCommerce and expense management. Owning a bank (WEX Bank) gives it an edge in these scenarios, allowing end-to-end integration that many FinTech challengers struggle to offer at scale.
Morgan Stanley Wealth Management launches Equity Vulnerability Score a risk management tool providing important insights for investors and advisors who hold concentrated equity positions
Morgan Stanley Wealth Management’s Global Investment Office (GIO) has launched the Equity Vulnerability Score, a proprietary tool that can help clients and the financial advisors who serve them measure and rank the susceptibility of US stocks to potential future drops in value. As a risk management tool, this can help provide important insights for investors—especially those who hold concentrated equity positions, which Morgan Stanley defines as five or fewer stocks making up more than 30% of the risk in a portfolio. Concentrated equity positions often occur naturally for company founders, those who receive equity compensation, and early investors. As these positions grow over time, they can unwittingly expose the investor to underperformance, higher volatility and material drawdowns—when a stock begins to decline from its peak and can drag the rest of the investor’s portfolio down with it. Looking historically, the GIO found that among the individual stocks contained in the Russell 1000 Index, a stock market index that represents the 1000 top companies by market capitalization in the United States:
- Individual stocks were more than twice as volatile as the index itself (37% v. 15%) since 2014
- The average stock’s maximum drawdown was twice as large as the index’s (approximately 50% vs. 25%)
- Most individual stocks tend to underperform the index on any forward-looking basis, with the median underperformance clocking in at -2.6% per year
- Most stocks that outperformed the index over five years went on to then underperform in the following five years1
- The Equity Vulnerability Score can help flag the likelihood that a stock may soon drop in value, and can also be used to complement Morgan Stanley’s existing Tactical Equity Framework, which helps identify short-term opportunities to seek overall stronger performance.