Short Blog – Bullets

It’s getting harder to keep up with payments, a subsector that has not been great for payment investors, between the Saaspocalypse, AI, agentic commerce, stablecoins, the Genius Act, open bank charters, and COF buying Brex, we are deep into “what just happened?” era. While I see tremendous opportunities, not everyone is impacted the same (see 2025: The Great Decoupling).

You don’t read this blog for deep tech insight; you read it to understand where change is actually happening and where the money, risk, and power are moving. Today is a short recap of which stakeholders face the biggest near-term impact, where progress is being made and where investment is flowing.

Processors: Authentication > AI

As I’ve been beating the drum about in previous posts, perfect authentication is going to reshape payments way more than AI ever will. Here’s what’s happening:

• Risk reallocation is the big game – How we manage and allocate risk among stakeholders will fundamentally rewire merchant acquiring in eCommerce. Merchants and processors have invested billions in eCom fraud and risk management. Perfect auth renders all these investments meaningless. Or it COULD.
• Payments is never an all or nothing game. High risk merchants may go all in as they solved their 300bps fraud problem, big merchants may let some high risk payments flow through a new channel. With customers engaging in new behaviors on new platforms, the last thing anyone wants to do is introduce another new brand or partner.
• Merchants prefer cards on file and ownership of the checkout experience. PSPs/MSPs must lead in managing how deterministic IDs (ex Payment Account Reference) are passed between platforms, networks, PSPs and processors. The near term will be the expansion of payment optimization. The longer term is in demand optimization, PSPs need to invest in both.
• PSP VAS – Processors need to fundamentally rethink how they create value for merchants beyond just moving money. AI gives them a shot at becoming merchant growth partners (conversion optimization, demand forecasting, inventory management) but they’re still stuck selling payment rails when merchants want business outcomes. The ones who figure out how to tie AI-driven VAS to actual merchant revenue growth will win the next decade. Fir example, can they evolve beyond back-office plumbing to build a Shopify-style demand side agentic connector (ex Google UCP)?
• Their challenge? They have been a cost center dealing with a payments team. Can they gain credibility w/ CMOs? Stripe has alienated Google with Pay to Play for everything. There may be opportunities for Adyen and eCom specialists to be the partners of choice to connect merchants where Shopify can’t.

Banks and Issuers

• As I’ve written in the “Future of Retail Banking” series, banks remain the biggest investors in new tech—using DLT to make closed networks more efficient, not to blow them up. While the tech analysts espouse “on chain” these closed networks use of DLT is like talking about banks being “on database”. It’s DLT making existing systems and networks more efficient; it’s not about expanding the commercial constructs (in most cases).
• Stablecoin Issuers’ payment of interest is the #1 banking issue as yields threaten to drain bank deposits across the spectrum. The Genius Act prohibited interest on balances stablecoin, but Stablecoin issuers created and end run “rewards”. Last week Draft Clarity Act Legislation signaled that only rewards tied to activity are within bounds. That’s a quiet but material win for banks.
• US banks are still grinding away on PAZE, pivoting (again) from guest checkout to agentic via a liability shift and a new bespoke authentication solution. This is loaded with irony here, as Early Warning pushed heavily for a new auth solution (ZelleID) 8 yrs ago (by my good friend Eric Woodward) only to have JPM smash it. While banks have a regulatory responsibility to manage authentication (when they own the risk), others have a need. A bank only solution will not scale as an approach that meets all needs requires governance and commercial terms (intent and payment). 
• Banks’ approach to partnership … While banks can deliver a technically adequate solution, they don’t collaborate well. As a reminder, the week after ApplePay launched in 2014, TCH and 2 large banks told Google, “Forget about what Apple got (15bps), you’re not Apple, let’s start the conversation with what Google will pay for the right to have our cards in your wallet”. BTW Google hasn’t forgotten this.  Meanwhile, ApplePay already has the liability shift and the consumer traction (more on this below). Or remember when ChasePay wanted their Logo below Amazon One Click Checkout (seriously).

Visa and Mastercard

• If you’ve read my “Bull Case for Visa and Mastercard” and “Network of Networks” pieces, you know I remain very positive. V/MA have the best shot at turning incremental VAS into massive economic flywheels, with network effects no new entrant can match. The history of 3DS shows that the challenge of managing liability shift is not tech, its all about incentives. Networks must control the authentication of actors, instruments and transactions. Yes banks must have the primary say, but its not their problem in isolation. They don’t touch customers until the END of the purchasing process and no one chooses what they want to buy based on their bank’s capability. Banks must support what the consumer wants to do.
• Change in network is hard as each entity had invested based upon the status quo. Thus the future looks like a “network of networks” where issuers and merchants opt into arrays of services: stablecoin clearing between participating issuers, richer authentication rails, and more specialized data services. I see the dam breaking with a big platform commits to a given direction. Right now, it’s the merchants that have the keys to platform direction. This is what the banks don’t understand (and Shopify does).
• The only areas where I see Network headwinds are B2B, as it is subscale. However, cards still anchor the economics, with players like RAMP who could defect, but would be crazy to walk away from card-based terms.

Agentic Platforms

• If you haven’t tried buying through an agentic flow yet, you should—if only for the comedy value. This week, Google announced new integration with Etsy and Wayfair, which is a new best-in-class example, but try something else.
• Google is doing everything right, but its moving at light speed and all their internal teams have yet to connect everything. GC on federated models, Gemini on checkout, Android in authorization, identity and payment credentials, GPay in AP2 and UCP, partnerships (every stakeholder), … there are 15 others that I can’t name. The scale of their innovations is unmatched. They are the TOP partner that merchants want to lean into right now.
• OpenAI is considering their own wallet, but is consumed with conquering the Gordian knot of agentic monetization. As they struggle ads and preferences introduce bias. Current experience dumps users to merchant checkout pages, breaking the agentic flow entirely. You go from chatting with AI to suddenly being on Target.com filling out forms. It’s jarring and defeats the purpose
• Perplexity’s commerce play – They’re pushing hard on product search and “buy” buttons, but it’s the same story: redirect to merchant checkout. No risk assumption, no real payment flow ownership. It’s affiliate marketing with AI lipstick.
• Shopify is still the best-in-class merchant agentic connector, which I laid out in my “Agentic Commerce” and Money2020 recap posts: catalog in, payments out, and agents in the middle. Expect Google to close much of that gap over the next 5–6 months as the NRF announcements with Walmart, Target, Macy’s and friends turn into real flows.
• The fundamental problem – These platforms want to be discovery engines but don’t want to touch fraud liability. Can’t blame them, but it means they’re dependent on either: (a) merchants integrating them via Shopify-style connectors, or (b) partnering with someone who’ll take the risk (hint: nobody wants to).
• The 2027 question – Will any of these platforms build actual payment rails, or will they all eventually pipe through Google/Apple/Shopify? My money’s on the latter. Building payments infrastructure is hard, and these companies are focused on AI, not becoming payment processors.

Merchants

Merchants ultimately pay for everything in payments and are the group to watch to understand where the industry is actually going. They want to use their own data to create unique customer experiences. Google and Shopify seem to be the only ones that understands this.

• Toe-dipping into agentic – Google’s NRF announcement made clear that the biggest merchants are tip toeing into agentic and plan to own the risk (and control data) until they get comfortable (see UCP Embedded).
• Heterogeneous. Top merchants like Walmart and Amazon have have fraud rates around ~10 bps with blended credit costs of 50-60bps; others merchants like footlocker have fraud rates north of 200 bps, and card processing costs just shy of 300bps. The gap in cost-of-payments and risk isn’t a rounding error, it’s the strategy.
• Merchants want a network solution (not joking). My favorite quote last year: “We have negotiated, litigated, and legislated Visa and Mastercard into the devil we know. There is no way we are giving up what we have built for something new (particularly with banks).” That’s exactly the dynamic I discussed in “The Great Decoupling”: V/MA as governance layer, not just pipes.
• In addition to authentication, agentic data must flow. In this area, nothing has changed since my Data Games post: tech like UCP, 3DS and Visa TAP can move data, but control and pricing block it. For now, merchants keep the data and the risk, and everyone else keeps updating tech around an Intent Mandate, with the economics and governance stalled. My key point here is that AP2 will be constrained within the Google domain. A commercial construct (or government mandate) is necessary to manage cross domain trust.

Mobile: Apple and Google

• Long-time readers will remember my thesis that Apple Card served 2 primary purposes 1) it demonstrated the great customer experiences that Apple could deliver on a card and 2) provided access to credit to finance Apple products. Now that Apple Card has moved from GS to JPM, wil Apple get off the enemy #1 list?  Will JPM partner with Apple on experiences and entitlements?
• ApplePay remains a net win for networks and issuers: they earn their 15 bps in the US (7 bps in EU), and card volumes grow. Apple’s extension beyond Safari hammered PayPal, and we’re still early in that curve. In the latest development, Network have begun hitting merchants with a tokenization fee (1bps) to shift the incremental costs of tokenized transactions from Issuer. Now merchants are concerned with greater costs coming (indirectly) from ApplePay (and all tokenized wallets) which pushes them further into managing cards on file (see COF Blog) AND withholding all transactional data (see Device Graph blog).
• Apple has the best starting position in agentic commerce due to their DPAN liability shift and auth rates, but must solve the collaboration roadblock with both Issuers and agentic platforms: Apple AND Issuers AND banks need clear signals about what, exactly, the consumer authorized.
• Google, meanwhile, has adopted the Apple direction of device bound identity (away from cloud credentials). This is the Android identity wallet vision I’ve been writing about since “Identity Wars – Silicon vs. Sovereignty” work: Titan M2, Android Ready, eID, UPI, PIX, agentic, .and everything else a consumer wants to do. Card agnostic, Identity provider agnostic, federated credential store, all in one coherent experience. Issuer PAZE efforts look like a pimple on a giant compared to the innovations here. IMHO Issuer friction only pushes Google’s innovation focus to areas where it has more control.
• Net: Apple stays card-friendly, extending the rails of Visa/Mastercard. Google pushes the frontier on agentic customer experiences, identity, non-card rails, merchant data integration, the role of mobile in credential management and payments.

Strategic Innovation (And Why I’m Tempted to Stir the Pot)

• Across the blogs, I’ve kept returning to the same point: networks are necessary for value transfer and risk management because no single stakeholder can move the system alone.
• The most dynamic collaboration is merchant ↔ processor/MSP (Shopify et al) <-> platform. This is where new economics and agentic flows are actually being built, not just talked about.
• The bundle of payments and identity have placed V/MA as the incumbent identity infrastructure of the internet. Their networks distribute the value of Bank KYC. But as machines interact with other machines, bank level KYC isn’t always required. Banks are at risk if transaction risk can be managed in an different way, for example Google managed risk to Affirm BNPL (or paybybank). Banks need to make their solution work beyond their own needs. Only V/MA can do that.
• In the era of strategic innovation, it is large incumbents tackling these big problems together.

The tech isn’t the problem. We’re at the heart of a generational evolution—but the arteries are clogged by incentives, governance, and control. I’m seriously contemplating pulling some of you into a more formal construct to accelerate what we all know should happen. If you’ve made it this far, you’re probably on my short list.