
The internet has always had a payments problem. HTTP moved data. SMTP moved email. But money? Money got stuck behind proprietary rails, bank integrations, and checkout forms that were never really built for a digital-first world. That gap, which the industry has spent decades papering over with varying degrees of success, is now the target of something bigger than any one company: the x402 Foundation, launched today under the Linux Foundation, with Coinbase, Cloudflare, and Stripe among its founding backers.
The announcement, timed to April 2 (a nod to HTTP status code 402, "Payment Required"), marks a formal step toward turning x402 into a neutral, community-governed standard. And the list of companies signing on makes it hard to dismiss as just another crypto lab experiment. Adyen, Amazon Web Services, American Express, Ant International, Google, Mastercard, Microsoft, Shopify, the Solana Foundation, Visa, and more than a dozen other names from across fintech, big tech, and crypto all attached their names to the effort.
The protocol is simple. When a client tries to access a resource gated behind x402, the server responds with the 402 Payment Required status code along with machine-readable payment instructions: amount, asset, network, recipient. The client then attaches a payment authorization header and resends the request. A facilitator verifies the payment and settles the transaction. That is the whole flow. No accounts, no subscriptions, no API keys, no manual billing cycles.
Coinbase launched the first version in May 2025, quietly, with the 402 HTTP status code having sat largely dormant since it was first defined in the early 1990s. Within months the protocol had processed over 100 million payments across APIs, apps, and AI agents. By December, the team shipped x402 V2, which added multi-chain support by default, cleaner separation between clients, servers, and facilitators, and the architectural foundations for session management and identity. The reference SDKs are available across TypeScript, Go, and Python.
Transaction costs sit near zero, with Coinbase's facilitator offering the first 1,000 transactions per month free and charging $0.001 per transaction beyond that. For micropayments, the kind worth a fraction of a cent that credit card networks have never handled well, that matters enormously. The protocol currently runs on Base, Polygon, and Solana, with stablecoins like USDC as the primary settlement layer. Future versions are designed to accommodate traditional rails as well, including ACH, SEPA, and card networks, using the same payment model.
The timing is not accidental. The push into autonomous AI agents across the industry has exposed a glaring problem: agents need to pay for things. When an AI assistant browses the web to buy something, or a trading bot needs a real-time data feed, or a robot needs to procure compute on the fly, making a human stop and authorize each payment defeats the entire point. What the industry needs is a payment primitive that works the way HTTP works: in the background, at machine speed, without friction.
"The internet was built on open protocols," said Jim Zemlin, CEO of the Linux Foundation, in comments tied to the launch. The Foundation's involvement is a deliberate move to ensure no single company ends up owning the payment layer of the agentic web. Cloudflare CEO Matthew Prince echoed that logic in September when the two companies announced their intent to launch the Foundation together: the internet's core protocols have always been governed independently, and x402 should be no different.
That governance structure is a meaningful part of the pitch. The x402 Foundation is framed explicitly as stewardship, not ownership. No single company controls the standard. The membership body is open to developers, startups, and enterprises. Cloudflare's alignment with the effort also signals that x402 is being treated as infrastructure at the edge level, not just a crypto developer toy. Integrating x402 into Cloudflare's edge compute and CDN stack means payment requests can slot into everyday web workflows the same way SSL became table stakes for basic security.
Early use cases already live in production. Hyperbolic, an AI compute marketplace, uses x402 for AI agents paying per GPU inference session rather than committing to a monthly subscription. OpenMind has robots autonomously procuring compute and data. Cal.com embeds x402 for paid human interactions directly inside scheduling workflows. The scope of what a frictionless pay-per-use primitive unlocks is genuinely wide, and that is before the protocol adds broader identity support and more payment backends.
There are real risks worth naming. The protocol currently leans heavily on Coinbase's own facilitator infrastructure, which handles verification and settlement and is, today, the most mature option in the ecosystem. Cloudflare and others reduce protocol-level concentration, but early traffic still routes largely through Coinbase's stack. The facilitator is free now. That may not last indefinitely once network effects solidify. And unlike credit card networks, x402 has no network-level payment reversal. Refunds require a compensating transfer from the merchant, making the protocol closer to cash than to a reversible card transaction. For high-frequency API calls that is a feature. For consumer flows that expect buyer protections, it is a liability worth monitoring.
What x402 has going for it, beyond the technical architecture, is the coalition. Visa and Mastercard alongside the Solana Foundation and Polygon Labs in the same founding member list is unusual. Google Cloud's managing director for Web3 and Digital Assets called the shift toward agentic commerce a fundamental reason Google is joining, describing the need for cloud infrastructure that is as open as the protocols it supports. Whether that breadth translates into real interoperability or remains aspirational will be one of the defining stories to watch as the Foundation gets off the ground. If x402 does become foundational plumbing, the question will be who benefits most from having been at the table when the standard was written.

Tether is best known for issuing USDT, the stablecoin that underpins much of the crypto market’s daily liquidity. But over the past few years, the company has been quietly expanding far beyond stablecoins. Its latest move pushes it even deeper into Bitcoin’s core infrastructure.
Tether has launched an open-source operating system designed specifically for Bitcoin mining. The software, called MiningOS, is meant to compete with the proprietary platforms that currently run much of the global mining industry. Unlike those systems, MiningOS is free, open to inspection, and designed to operate without centralized control.
It is a technical release, but also a philosophical one. At a time when Bitcoin mining is increasingly dominated by large, well-funded players, Tether is positioning itself as a company willing to open the tooling layer and lower at least some of the barriers to participation.
MiningOS is software that helps miners manage and coordinate their machines. It handles things like monitoring performance, configuring devices, managing power usage, and scaling operations across large numbers of mining rigs.
That may not sound exciting, but in mining, software choices matter a lot. Most large mining firms rely on closed, proprietary systems that are licensed and often tied to specific hardware vendors. These platforms work well, but they come with fees, restrictions, and limited transparency.
MiningOS takes a different approach. It is modular, meaning operators can adapt it to different setups and environments. It can run on lightweight hardware for small operations, but it is also designed to scale to industrial mining sites with thousands of machines.
One of the more interesting aspects is its peer-to-peer architecture. Instead of relying on centralized servers, devices communicate directly with one another. That design choice can reduce infrastructure costs and make operations more resilient, especially in environments where connectivity or uptime is a concern.
By making the software open source, Tether is also allowing anyone to inspect the code, modify it, or build on top of it. That alone is a big departure from how mining software has traditionally been distributed.
This release fits into a much broader shift inside Tether. The company has been steadily moving into mining, energy infrastructure, and artificial intelligence, positioning itself as more than just a stablecoin issuer.
Timing also matters. Bitcoin mining has become a tougher business, especially after the most recent halving cut block rewards again. Margins are tighter, competition is intense, and efficiency is everything. For miners trying to stay profitable, cutting software costs and gaining more control over operations can make a real difference.
At minimum, MiningOS gives miners another option. At best, it could force existing software providers to compete harder on transparency, pricing, and flexibility.
The mining software market rarely gets attention, but it has real influence. Whoever controls the software often controls how hardware is deployed, optimized, and integrated with pools and power systems.
An open-source alternative disrupts that model. Miners can audit the code themselves, customize it for specific environments, or adapt it to unusual power setups. They are no longer forced to trust a black box or depend on a vendor’s roadmap.
For smaller and mid-sized miners, this could be especially valuable. Licensing fees may not be the biggest expense in mining, but when margins are thin, every recurring cost matters. Removing software fees lowers the break-even point and gives operators more room to adapt.
There is also a broader network effect to consider. Bitcoin’s security depends on distributed hash power. Anything that makes it easier for independent miners to stay online and competitive helps reinforce that foundation, even if the impact is gradual rather than immediate.
This is not a magic solution. Mining is still capital-intensive and energy-dependent. Open-source software does not solve access to cheap electricity, hardware supply, or regulatory pressure.
Adoption will also take time. Mining operators tend to be conservative with infrastructure changes, especially when uptime and reliability are critical. MiningOS will need to prove itself in real-world deployments, not just in theory.
There is also the question of trust. Tether remains a controversial company in parts of the crypto world. Even with open-source code, some miners and developers may be hesitant to engage until the project builds a track record and an active community.
Still, the fact that the code is open means the software can outgrow its origin. If it is useful, the ecosystem can take it in directions Tether itself may not control.
I like this move. Open-sourcing mining software feels overdue in an industry that depends so heavily on closed systems most people never see. For all the talk about decentralization in Bitcoin, a surprising amount of the mining stack has remained locked behind proprietary tools and vendor agreements. This at least pushes in the opposite direction.
Will this suddenly make Bitcoin mining accessible to everyone? No. Power, hardware, and capital still matter, probably more than software ever will. But removing one layer of friction does count, especially at a time when miners are under real pressure to cut costs and stay flexible.
I am also glad this is open source rather than another branded platform with a free tier and strings attached. Anyone can inspect it, improve it, or fork it if they want. That alone changes the power dynamic. Even miners who never run MiningOS may benefit if existing software vendors are forced to be more transparent or more competitive as a result.
Tether is a complicated company, and skepticism around anything it touches is fair. But good ideas do not stop being good just because they come from a controversial source. Open infrastructure tends to outlive the companies that release it, and that is sort of the point.
If Bitcoin is going to stay resilient over the long run, it needs more open tools at the base layer, not fewer. On that front, this feels like a step in the right direction, and I am genuinely happy to see it happen.