Last November I argued that the only fintech theme for 2026 that really mattered was multi-layer money, the structural shift where software innovation outpaces the evolution of the underlying rails, and value migrates up the stack into programmable layers that sit above the legacy plumbing. Wallets surpassing bank accounts. Stablecoins and tokenized deposits stacking on top of card rails. Agents transacting across whatever layer clears fastest and cheapest.
That piece was about the macro shape of the shift. What I didn’t name was the specific places where value gets captured as the old stack gets unbundled.
This piece is about one of those places. And it’s one the fintech commentary circuit has mostly missed.
Go to github.com/juspay/hyperswitch. The facts:
~41,000 stars, ~4,600 forks, 150+ releases, 260+ contributors
Written in Rust, Apache 2.0 licensed, one-click deploy to AWS/GCP/Azure
Six modular components: intelligent routing, vault, cost observability, revenue recovery, reconciliation, and alternate payment methods, deployable standalone or stitched together as a full orchestration stack
Maintained by Juspay, a twelve-year-old Bangalore-based payments company that processes roughly 200 million transactions per day and an annualized payment value north of $1 trillion for the largest merchants in India, with global operations in the US, UK, and EU since early 2025
Juspay calls Hyperswitch “Linux for Payments.” Too cute to accept at face value, too pointed to dismiss.
The reason this matters is that Hyperswitch does to the orchestration layer of payments what Linux did to proprietary Unix: commoditizes the code and redirects the margin to everything wrapped around it. The consequences show up unevenly across the stack, and they arrive on different timelines for different players.
But they will arrive.
Juspay is not a Silicon Valley startup burning venture money to buy market share. They are the payments operating system of Indian digital commerce. Flipkart, Amazon India, Swiggy, Uber India, most of the large Indian marketplaces and airlines run on their infrastructure. They have been profitable or near-profitable for years. Earlier this year they raised $50m at a $1.2B valuation, making them India’s latest “unicorn”. They have been doing at scale, in one of the most fragmented payments markets on the planet, exactly the thing Primer, Gr4vy, Spreedly, Payrails, and every enterprise in-house payments team in the West has been trying to build.
When Juspay open-sourced Hyperswitch in October 2022 and then spent three years hardening it to the point where they could launch it into US, UK, and EU markets in early 2025, they were not running a developer-relations project. They were executing one of the more deliberate commercial open-source go-to-market plays in fintech history.
The commercial logic is worth stating plainly:
Juspay’s moat in India was never the code. The moat was the bank relationships, the UPI integrations, the fraud models tuned on a decade of Indian transaction data, the 24/7 operational capability, and the compliance posture with the RBI. None of that is in the GitHub repo. What is in the repo is the connector framework, the routing engine, the vault, and the control plane; expensive to build, but not strategic differentiators once competitors exist.
The problem was distribution, not product. A Bangalore vendor cannot, by brute force of sales motion, convince a US-based Fortune 500 CFO to put their production payments stack on a brand nobody in procurement has heard of. The enterprise sales motion required to compete against Stripe, Adyen, or even Primer in North America would have cost hundreds of millions of dollars and taken a decade, with no guarantee of credibility at the end.
Open source solves the distribution problem in one move. Give away the code. Let developers pull it down on a Friday afternoon. Let a global community do the credibility work your Bangalore sales team never could. By the time enterprise procurement gets involved, the technology has already been validated inside the merchant’s own engineering organization.
Then monetize downstream. Juspay already ships both an Enterprise Edition (self-host with paid support and SLAs) and a Cloud Edition (fully managed SaaS with two-to-three week onboarding). Not “coming soon”. Live.
Hyperswitch’s Github repo is one of the fastest growing finance repos on Github (for the week ending Apil 10, 2026, they added close to 667 stars). In the thirteen months since the US/UK/EU launch specifically, accumulation has run about 2,200 stars per month, roughly five times the pre-launch rate. Recent activity reflects an acceleration.
The obvious objection: don’t Stripe and Adyen also have big open-source footprints? They do, but it’s a different category of thing.
Stripe’s GitHub organization has around 85 repositories, and the largest, stripe-node at ~4,400 stars, is a Node.js client library for calling Stripe’s proprietary API. Adyen’s footprint is structurally identical: checkout frameworks, mobile SDKs, API specifications. Fiserv has 18 public repositories total, and the flagship is sample integration code.
Stripe and Adyen have open-sourced the integration layer and kept the platform proprietary. Hyperswitch has open-sourced the platform.
That is a deliberate strategic choice, and it reveals where each side believes the moat lives. The incumbents bet the platform is the moat. Juspay is betting the platform is the commodity and the moat is operations, data, and managed services. Both bets can’t be right.
For Orchestration Pure-Plays (Primer, Gr4vy, Spreedly, Payrails) Hyperswitch is an existential threat. The entire business model was to sell the connector library, the routing UI, and the retry engine as proprietary SaaS, exactly the layer Hyperswitch now ships for free with enterprise backing. AI coding agents compress the integration labor even further.
These companies have maybe 18 to 24 months to pivot from “we sell orchestration” to “we sell managed orchestration with operational excellence, compliance, and proprietary data.” The thoughtful ones will make the transition. The PE-backed / VC-funded ones who assumed SaaS gross margins on what’s becoming commodity code are in trouble. Consolidation in this segment is the highest-probability event in the next 24 months.
Legacy Scale Acquirers (Fiserv, FIS, Global Payments/Worldpay, Nuvei). Counterintuitively, they are the best-positioned archetype if leadership is honest about what they actually get paid for. Hyperswitch commoditizes the software rent. It does not touch the acquiring rent, licenses, bank sponsorships, risk underwriting, scheme relationships, settlement, float. Those moats remain. When an acquirer CEO says “scale matters” on an earnings call, that’s what they’re pointing at, whether they realize it or not.
Global Payments has the strongest thesis in the group. They were never competitive on gateway or developer experience, so they lose nothing they were getting paid for, and the open-orchestration wave levels the playing field on the layer where they actually compete. Also worth noting how profitable the acquiring volume is when you can “outsource” the development and maintenance of the gateway to an open-source project like Hyperswitch.
If the Board isn’t reviewing a strategy to implement the “hybrid play” in the next quarter, it would be a big miss. Becoming the managed-Hyperswitch-plus-acquiring offering for enterprise merchants who want the open layer but don’t want to run Rust services is a $10B+ revenue opportunity and plays to their operational DNA.
The more likely outcome is that legacy players like GPN drift for 12 to 18 months pursuing proprietary orchestration out of institutional instinct as they watch enterprise wins go to competitors who embraced open orchestration. They’ll then pivot with a splashy “open payments initiative” in 2027 or 2028. The pivot will be correct. The delay will be expensive.
Full-Stack Challengers: Stripe and Adyen. Not existential, but the margin compression is real, and they’re positioned very differently. Stripe has been quietly multi-acquirer at the enterprise tier for years, Connect, Payment Method Configurations, and enterprise pricing all allow sophisticated merchants to route volume to non-Stripe acquirers when they outgrow the bundle. That gives Stripe a credible retreat path: concede orchestration pricing at the top of the pyramid while defending developer brand, SMB/mid-market, and adjacencies (Atlas, Capital, Issuing, embedded finance) where the next decade of growth lives.
Adyen is the more exposed of the two. Their entire identity has been the “single platform” religion, gateway and acquiring welded together as an architectural principle. Not a product choice, but a theology. When a large merchant drops Hyperswitch in front of Adyen and routes a test slice to a specialist acquirer in one geography, the blended price gets decomposed for the first time. The gateway portion collapses toward zero. Adyen has to defend the acquiring portion alone, on price and auth rate.
Adyen has two real defenses: best-in-class auth rates (especially in Europe), and deep acquiring licenses across ~40 jurisdictions. Both are durable moats but around maybe 40% of current revenue, not 100%. Stripe has optionality in how it retreats. Adyen has to either reinvent its strategic narrative or watch its multiple compress from “integrated platform” to “best-in-class acquirer,” which is a meaningfully lower valuation regime.
Here’s the caveat that keeps the thesis honest.
Everything above describes the card-not-present orchestration layer, e-commerce, mobile app payments, subscription billing, marketplace flows. It is the layer where Hyperswitch can credibly displace proprietary orchestration in the next 36 months. It is also not the entire enterprise payments market.
The enterprise omnichannel buyer, Marriott, Delta, CVS, Kroger, Darden, evaluates payments infrastructure against a different set of criteria. PCI-validated point-to-point encryption (P2PE) to descope thousands of store locations from audit scope. Certified integrations with Oracle Retail, Oracle Hospitality (OPERA, Simphony), NCR Voyix, Micros, SAP Retail. Semi-integrated terminal architectures where the payment layer sits outside the POS system’s PCI footprint. Terminal fleet management. 24/7 enterprise support with indemnification.
Almost none of that is software. It is compliance labor, hardware certification, multi-year partnership programs with POS software vendors, and field-proven operational capability. It is where FreedomPay quietly built a business running payments for Hilton, Marriott, IHG, Hyatt, and most of the premium global hospitality brands. It is where Fiserv’s Carat, Shift4, and Adyen’s in-person product actually earn their margin. And it is work that open source does not accelerate, if anything, P2PE certification is tied to specific validated configurations, which complicates arbitrary self-hosted deployments.
Hyperswitch’s addressable market in the next 36 months is the online-first and multi-channel-digital segment, marketplaces, platforms, global e-commerce, subscription businesses, agentic commerce infrastructure, and the mid-market enterprise tier that does not have thousands of physical doors. That’s still a very large pool: tens of billions in addressable software rent over the next decade, dwarfing what Juspay could capture from India alone.
Adyen’s revenue comes from two different populations sitting at different levels of exposure. The large online-first customers (Uber, Spotify, Booking.com, eBay) are exposed to Hyperswitch displacement. The true omni-enterprise customers (H&M, McDonald’s, Microsoft Store) are not, or at least not without Juspay investing years in an entirely different kind of capability-building. The equity story should reflect that these are effectively two companies inside one ticker. It doesn’t yet. That’s where the multiple compression comes from when the market catches up.
And the smartest move for Juspay is probably not to try to build the omni stack themselves. It’s to partner with the enterprise omni incumbents FreedomPay, Fiserv, Shift4, NCR Voyix as the open routing layer inside their certified solutions. Red Hat didn’t build every enterprise integration itself. It partnered with the vendors who already had them. The Linux distribution pattern, extended to payments.
If you’re a CFO or Chief Payments Officer at a merchant doing more than a few hundred million in online volume, the ground has moved.
The engineering cost of running your own orchestration has collapsed. Hyperswitch plus AI coding agents means the team required to maintain a multi-processor payments stack is a fraction of what it was in 2022. The old rule of thumb “don’t in-house orchestration under $10B in GMV” probably now sits around $500M to $1B. A tenfold expansion of the rational in-housing population.
The strategic cost of staying bundled is going up. Agentic commerce, multi-rail settlement, and cross-border optimization all reward merchants who can route dynamically across acquirers based on real-time economics. A merchant locked into a single full-stack provider cannot participate in that optimization and will watch their costs drift higher relative to peers who can.
The right move for most enterprise merchants is not to rip out their existing provider that’s a five-year project with real execution risk. The right move is to run the decomposition exercise. Pull Hyperswitch down in a sandbox this quarter. Route 5% of volume through it to a second acquirer in one geography. Benchmark the unbundled economics against your current blended rate. Use the resulting data as leverage in your next renewal conversation, regardless of whether you actually migrate. The quote you get from your current vendor when they know you have real optionality will pay for the exercise ten times over in the first year.
For enterprise platforms, marketplaces, vertical SaaS, embedded finance, the implication is larger. The next generation of commerce platforms is being architected right now, and the default choice for their payments stack is no longer “pick Stripe or Adyen.” It’s “pick an orchestration layer and plug the best acquirers underneath it.” Every platform that adopts the open layer compounds the community and raises the switching cost for proprietary full-stack providers. The new workloads go somewhere else.
The slow repricing (~55%). Hyperswitch becomes the default choice for new online-first enterprise merchant architectures over 24 to 36 months. Adoption driven by platforms, global marketplaces, and merchants in fragmented multi-rail geographies. One or two marquee Fortune 500 logos publicly announce migrations in late 2026 or 2027 is the tipping point. Orchestration pure-plays consolidate or pivot. Adyen’s multiple compresses as the two-speed reality becomes visible. Stripe holds enterprise pricing through brand and adjacencies but concedes the orchestration tier. GPN drifts, then pivots. Juspay captures the managed-service layer outside India, building toward a multi-billion-dollar managed-service revenue line by 2028-2029.
The hyperscaler strike (~25%). AWS, Google, or Microsoft launches a managed Hyperswitch-based payments service, bypassing Juspay and capturing the managed rent the way AWS did to MongoDB with DocumentDB and Elastic with OpenSearch. Juspay responds with a license change (SSPL or similar) that’s defensible but controversial and fragments the community. The open-orchestration thesis still wins in aggregate; the margin capture moves to the hyperscalers.
The incumbent counter-move (~15%). Stripe, Adyen, or a coalition of scale acquirers launches a defensive fork or parallel open-source initiative to fragment the ecosystem. Lower probability because it requires incumbents to cannibalize existing revenue on a faster timeline than boards typically tolerate, but not zero. Stripe in particular has the engineering culture and balance sheet to execute if margin compression arrives faster than expected.
The thing that does not happen in any scenario is the status quo holding. Proprietary online orchestration as a standalone software rent is over. The only real questions are who captures the downstream managed-service margin, how fast the repricing hits incumbent equity values, and whether Juspay executes the commercial-open-source go-to-market well enough to capture the economics its own strategic play has unlocked.
Last November I wrote that multi-layer money would be the only fintech theme of 2026 that mattered, that value would migrate up the stack as software innovation outpaced the rails. I framed it as a macro thesis.
What I didn’t say clearly enough is that the same dynamic was about to play out inside the payments stack itself.
The rails under the orchestration layer, card networks, acquiring banks, settlement systems, remain durable and rent-producing.
The rails above, agentic commerce, programmable money, multi-rail routing, are where the growth lives.
And the orchestration layer itself, the thing that used to be the most valuable real estate in the entire stack, just got routed around by a free Rust codebase maintained by a profitable Indian company that figured out distribution beats differentiation.
The enterprise omni tier is a different story, compliance, hardware certification, and POS software integration remain real moats for whoever has invested the years to build them. That’s where Fiserv, FreedomPay, Shift4, and Adyen’s in-person capability keep earning their margin.
The pragmatic version of the thesis is not “open source wins everything.” It’s this:
The online orchestration rent is over, the enterprise omni rent is durable for another decade, and the smart money is on companies positioned to capture value at the interface between the two.
Linux didn’t kill Unix by being better. It killed Unix by being free, good enough, and architecturally honest about what actually deserved to be proprietary. Hyperswitch is doing the same thing to online payments orchestration in 2026. The incumbents still selling the whole bundle in 2028 will look the way Sun did by 2005, still shipping, still profitable, and trading at a fraction of their peak multiple because the market had figured out the category was melting before the revenue line showed it. The ones who recognize that the integration and compliance layers are where their real moats live, and who position accordingly, will be fine.
The merchants who capture the benefit of the shift are the ones running the decomposition exercise this quarter. The incumbents who win are the ones honest about which layer of their stack was actually producing the rent. And the company that captures the most value from the whole transition is probably not the one whose logo you’d guess. It might be the Bangalore-based operator that spent three quiet years hardening an open-source payments kernel while nobody was looking.
That’s the story the fintech commentariat isn’t telling yet.
They will be.