Tim Spence, Fifth Third Chairman, CEO and President said,” Fifth Third’s financial results once again underscore our strong balance sheet, diverse revenue streams, and disciplined expense management. We’ve expanded our net interest margin, improved credit metrics, and strengthened our efficiency ratio. Our ongoing investments in strategic growth priorities continue to drive robust results. In the second quarter, adjusted revenues and adjusted PPNR increased year-over-year by 6% and 10%, respectively, marking the highest growth rate in the past two years. Our balance sheet remains well-diversified and neutrally positioned. This quarter, we accreted 13 basis points of CET1 capital and grew tangible book value per share by 18% over the past year. By focusing on developing the capabilities to generate high-quality deposits, diversified loan originations, recurring fee revenue and consistent improvements in operating scalability, we expect to continue to generate strong, stable returns for our long-term shareholders during volatile environments. As we move forward, we will continue to adhere to our operating principles of stability, profitability, and growth – in that order.
Citizens 2Q25 reports continued strong Private Bank progress, contributing $0.06 to EPS and up $0.02 QoQ
Chairman and CEO Bruce Van Saun said, “We are pleased to report strong results today that came in ahead of expectations, paced by strong NII and fee growth, disciplined expense management, and credit results that are trending favorably,” said Chairman and CEO Bruce Van Saun. “We saw some sizable M&A advisory fees push out to July, but offset that with strong performance across other fee categories. Our key strategic initiatives, paced by the Private Bank/Private Wealth build out, continue to make good progress, and we have commenced work on a broad ‘Reimagining the Bank’ initiative that will become a multi-year TOP program and drive meaningful benefits from using new technologies to serve customers in new ways and run the bank better. We are wellpositioned to have a strong second half of the year and to sustain that momentum into the medium-term.”
- Private Bank build out tracking well; adding wealth teams and opening new PBOs; confident the business will deliver a ~20 to 24% return on equity for FY2025 and maintain this over the medium term
Building a premier Private Bank
Deposits
- 36% noninterest-bearing
- 2% total deposit cost
- Continued client growth with a $966 million increase in average deposits; spot stable reflecting timing of inflows/ outflows
AUM
- Added 3 advisor teams during 2Q25 in NYC metro, North New Jersey and Los Angeles
- 8 advisor teams added since launch; all located in key markets aligned with PBO offices: New York, San Francisco, Boston, Boca Raton, Naples, Southern California
Loans
- Loan growth driven by improvement in capital call line utilization, given increasing client activity, as well as growth in retail mortgage
- portfolio yield ~7%
Financial profitability
- Tracking well towards targets for ~$12B deposits, ~$7B loans and ~$11B AUM by end of 2025
- Expect 5%+ earnings contribution to total CFG in 2025
Confident the business will deliver a ~20 to 24% return on equity for FY2025 and maintain this over the medium term
SoFi partners Paychex, allowing SMEs to access discounted loans and one-on-one financial advice directly from their payroll portal
The integration places SoFi at Work, the employee-benefits arm of the digital bank, within Paychex Flex Perks, the online benefits marketplace inside the Paychex cloud-based HR platform. Through single sign-on, workers can access lower-rate personal loans, refinance student debt, schedule a free 30-minute financial planning call, and use SoFi tools to evaluate payoff strategies. “There’s incredible alignment between what we’re both trying to accomplish,” said Kelli Keough, SoFi’s SVP, highlighting Paychex’s reach among smaller businesses that often lack robust benefit offerings. Paychex processes payroll for around 740,000 businesses, and since launching Flex Perks last year, more than 230,000 employees have purchased at least one benefit through the platform. The SoFi integration brings “enterprise-level” financial wellness tools typically inaccessible to small companies. Keough noted that SoFi’s value extends beyond discounts, emphasizing education, planning, and digital tools to help workers take control of their financial lives. Employees can explore and apply for financial products without leaving the Paychex environment, with early incentives focused on workers—such as discounts on loans and access to financial tools. Student-loan relief is expected to attract interest, but SoFi’s broader goal is addressing holistic financial needs like credit-card consolidation. The partnership also helps Paychex differentiate in the competitive payroll space by expanding beyond lifestyle perks into financial services. SoFi will track engagement metrics including employer promotion rates, employee logins, and product uptake, using the data to refine its offerings. The integration is already live, with enrollment open immediately and expected to peak during fall’s open-enrollment season. For SoFi, this move not only supports financial wellness goals but also opens the door to future banking relationships, as Keough noted: “Of course, we do hope that they will become lifelong members.”
PNC is considering charging fintechs for access to valuable customer data, following the lead of JPMorgan
PNC Financial Services Group Inc. is considering charging financial-technology companies for access to valuable customer data, following the lead of JPMorgan Chase & Co., Chief Executive Officer Bill Demchak said. The bank is in “discussions” to determine what actions to take, Demchak said. JPMorgan has sent pricing sheets to data aggregators outlining proposed charges that could amount to hundreds of millions of dollars. “I applaud” what JPMorgan did, Demchak said. “There’s a big cost to keeping this data secure and producing it in a form that’s readable for our clients. So we’re thinking about it.” The charges could upend business models and potentially affect fintechs that rely on access to bank account information, like peer-to-peer payment platforms or cryptocurrency wallets. “American consumers, not banks, own their financial data, and big banks are taking advantage of this moment of regulatory uncertainty to crush competition and undermine innovation,” Lee said.
Google study shows LLMs could quickly lose confidence and abandon correct answers when presented with a counterargument, even if the counterargument is incorrect threatening multi-turn conversational agents
A new study by researchers at Google DeepMind and University College London reveals how LLMs form, maintain and lose confidence in their answers. The findings reveal striking similarities between the cognitive biases of LLMs and humans, while also highlighting stark differences. The research reveals that LLMs can be overconfident in their own answers yet quickly lose that confidence and change their minds when presented with a counterargument, even if the counterargument is incorrect. Understanding the nuances of this behavior can have direct consequences on how you build LLM applications, especially conversational interfaces that span several turns. This study confirms that AI systems are not the purely logical agents they are often perceived to be. They exhibit their own set of biases, some resembling human cognitive errors and others unique to themselves, which can make their behavior unpredictable in human terms. For enterprise applications, this means that in an extended conversation between a human and an AI agent, the most recent information could have a disproportionate impact on the LLM’s reasoning (especially if it is contradictory to the model’s initial answer), potentially causing it to discard an initially correct answer. Fortunately, as the study also shows, we can manipulate an LLM’s memory to mitigate these unwanted biases in ways that are not possible with humans. Developers building multi-turn conversational agents can implement strategies to manage the AI’s context.
Truist’s Merchant Engage combines banking and payments in one- delivers a rich, intuitive user experience that shows all of the processing, all of the settlements, all the dispute management in a seamless experience
Truist Financial Corp. has launched a new integrated merchant-services platform called Merchant Engage, aimed at meeting the banking and payment needs of businesses ranging from small enterprises to large merchants. The free service allows Truist clients with both banking and merchant accounts to view all transaction data in one centralized platform. Developed in partnership with London-based fintech Pollinate International Ltd.—marking its U.S. debut—Merchant Engage offers a comprehensive digital experience covering dispute management, new-location onboarding, and transaction processing. According to Chris Noe, Truist’s head of wholesale payments, the bank covers Pollinate’s service fees, making the offering cost-free for users. Pollinate provides customizable modules across the merchant lifecycle, delivering a seamless experience without requiring users to manage processor tie-ins or aggregate data themselves. “The client is able to get a rich, intuitive user experience that shows all of the processing, all of the settlements, all the dispute management, and anything that they would possibly need,” Noe explained. For instance, onboarding a new merchant location can now be done entirely within Merchant Engage, avoiding the need to navigate multiple platforms. While Truist may be first with this level of integration, other banks like U.S. Bank are also expanding their merchant services, such as with April’s launch of U.S. Bank Essentials.
While MCP offers agent identification, absence of guardrails for communication, lack of audit trail and inability to do KYC on agents to authenticate them present adoption challenges for regulated companies
Model Context Protocol (MCP) has begun amassing a large number of users, all but guaranteeing the mass adoption needed to make it an industry standard. While many regulated companies, such as banks, financial institutions, and hospitals, have begun experimenting with AI agents, these are typically internal agents. Regulated companies do have APIs. Still, so much of the integration these companies undertake has taken years of vetting to ensure compliance and safety. However, it doesn’t mean financial services companies want to jump into the MCP and Agent2Agent (A2A) bandwagon immediately. Since MCP is open source and new, it is still undergoing constant updates. Sean Neville, cofounder of Catena Labs said that while MCP offers agent identification, which is key for many companies, there are still some missing features, such as guardrails for communication and, most importantly, an audit trail. These issues could either be solved through MCP, A2A or even an entirely different standard like LOKA. He said one of the biggest problems with the current MCP revolves around authentication. When agents become part of the financial system, even MCP or A2A, there’s no real way to do “know-your-customer” on agents. Neville said financial institutions need to know that their agents are dealing with licensed entities, so the agent must be able to point to that verifiably. John Waldron, senior vice president at Elavon, doesn’t discount the possibility that financial institutions may work towards supporting MCP or A2A in the future. “Looking at it from a business perspective and demand, I think MCP is a very critical part of where I think the business logic is going,” he said.
Eudia’s acquisition of Johnson Hana to enable it to build a network of AI-augmented human workforce by embedding the deep industry and institutional knowledge of 300+ elite legal professionals into its AI platform
Eudia, the Augmented Intelligence platform for Fortune 500 legal teams, announced its acquisition of Johnson Hana, bringing 300+ elite legal professionals into the world’s first AI-augmented human workforce. The move creates an entirely new type of company—one that fuses human expertise and artificial intelligence to fundamentally reinvent how legal work is performed. While many believe AI is simply software, Eudia has proven that AI’s true potential lies in human partnerships. Human + AI teams consistently outperform humans or AI working alone—a principle Eudia calls “Augmented Intelligence.” This acquisition brings to life a new type of workforce that accelerates outcomes and drives unprecedented business value across Eudia’s customer base. Johnson Hana represents a fundamental departure from traditional alternative legal service providers (ALSPs). This acquisition enables Eudia to build the world’s first team of AI-augmented humans—legal professionals who accelerate both outcomes and transformations within their Fortune 500 clients. The Johnson Hana acquisition accelerates Eudia’s vision of building a comprehensive network of augmented legal professionals. These experts bring deep industry knowledge that is directly embedded into Eudia’s AI systems, creating compounding intelligence that benefits all clients. These operators are trained in Eudia’s proprietary process of capturing institutional knowledge and context, enabling their clients to compound the effectiveness of AI for both internal and external teams.
The memecoin market is moving away from creating hype to building trust backed by open AMAs, clear roadmaps, and community engagement, evolving into a broader ecosystem and pivoting to super app models
The memecoin market, once the playground of viral trends and overnight riches, is entering a new phase. Thousands of tokens now flood platforms like Ethereum and Solana, fragmenting liquidity and thinning investor focus. Coins often hold only 20–40% of their market cap in liquidity. That leaves little margin for volatile assets. The crowded market has sharpened investor expectations. No longer will a meme and a mascot suffice. The winning tokens now build trust—through transparency, accountability, and community engagement. CAPTAINBNB is one such example. Its 100% circulating supply and renounced contracts signalled integrity, helping it build a loyal base. This kind of trust—backed by open AMAs, clear roadmaps, and genuine developer commitment—often sustains projects through downturns. In contrast, countless memecoins launched with fanfare in 2023–24 are now abandoned, unable to survive a single market dip. Key Opinion Leaders (KOLs) once ruled the memecoin narrative. But by 2025, skepticism has caught up. In short, the influencer model is no longer a guarantee. In many cases, it’s a liability. Where hype is fading, utility and grassroots support are taking its place. Shiba Inu’s transformation offers a blueprint—evolving into a broader ecosystem with ShibaSwap and Shibarium, giving holders reasons to stay beyond the meme. Some projects are pivoting to super app models that empower user decisions and foster participation. This bottom-up governance reflects a maturing memecoin scene, where communities are not just holders but stakeholders.
Enterprises in regulated industries are embracing sovereign cloud with baked-in compliance, governance, and operational control across regions for scaling AI driven by pressure to support composable architectures, low-latency performance, and policy-based data residency
Sovereign cloud is rising in prominence: It ensures compliance, governance and operational control across geopolitical regions while still enabling innovation at scale, according to Kevin Cochrane, chief marketing officer of Vultr, a registered trademark of The Constant Company LLC. The two-fold nature of sovereignty lies in national infrastructure and enterprise-level governance, according to Cochrane. “The agents that they’re deploying are only going to be able to communicate in many cases within that geographic boundary. There’s all sorts of controls. The issue of sovereignty actually also matters for the enterprise.” This need for localized governance is especially urgent in healthcare, financial services and other heavily regulated industries. Vultr’s approach includes prebuilt infrastructure templates tailored for compliance — allowing developers to deploy AI-native applications with speed and confidence. By productizing infrastructure-as-code stacks, Vultr helps organizations reduce the time from proof of concept to production, Cochrane explained. Vultr’s sovereign cloud capabilities are not an add-on; they’re embedded into the architecture. The company’s on-demand infrastructure model and open ecosystem approach allow enterprises to customize deployments while maintaining strict control over data, networking and compute layers. With compliance baked in, organizations avoid the trap of retrofitting governance after the fact — a risk that can delay or even derail AI projects, according to Cochrane.