On Monday, Magic Labs—the embedded wallet infrastructure company behind Polymarket and WalletConnect—sold its wallet business to Payward, the parent company of Kraken. Co-founder and CEO Sean Li made the announcement in a single post. The same post carried a second piece of news: the company is rebranding as Newton Labs.
A sale. A rebrand. Two clean announcements. Nothing about this story is clean.
Magic Labs was not a distressed startup. It was the rails. Polymarket, one of the most active prediction markets in crypto, uses Magic’s embedded wallet infrastructure. WalletConnect, the session layer that connects mobile wallets to dApps, is also a customer. This is not a fire sale. It is an infrastructure provider handing its most visible product to an exchange parent while the founders keep the company and a new name.
Follow the gas, not the narrative. The narrative is “Magic is evolving.” The gas is that a company that spent years building wallet onboarding for other people’s apps just sold that exact reach to Kraken. When infrastructure sells its flagship product, it tells you where the margin has moved. It is not in wallets.
Context: What Magic Actually Built
Magic Labs, originally Fortmatic, popularized the embedded wallet pattern: a non-custodial wallet accessible through email authentication, magic links, and social login. To a retail user, it removes seed phrases. To an application, it provides a JavaScript SDK that turns a website into a wallet frontend. This is not a hardware wallet business and not an exchange custody engine. It is an account abstraction play that became a login layer.
Kraken, through Payward, now owns the wallet business. That includes the technology, the user-facing components, and presumably the customer relationships with Magic’s existing wallet users. This is a distribution event. Kraken has its own exchange and its own custody infrastructure, but it did not control the embedded wallet layer used by Polymarket and other dApps. That gap is now closed.
Newton Labs is the remaining entity. The name suggests a shift toward something closer to protocol work, but the announcement does not provide a product roadmap. Do not fill in the blank with speculation. In 2021, I built a Dune dashboard that mapped NFT community wallets and found the same pattern over and over: the asset that matters is wallet-level behavior, not the logo on the website. Magic Labs’ sale is a wallet-level event. The rest is narrative.
Core: What the Exchange of the Wallet Actually Reveals
The core insight is not the sale. It is what the sale exposes about the wallet stack.
Wallet infrastructure is being split into two layers: custody and distribution. Kraken just bought distribution. Magic’s embedded wallets are non-custodial, meaning the private keys are held by the user, but the onboarding experience is controlled by the service provider. Whoever controls the SDK controls the first screen, the recovery fallback, and the data trail of sign-ups. That is more valuable than custody in a market where users do not know the difference.
Look at Polymarket’s user journey. A trader arrives, signs up with an email, and is immediately given a wallet via Magic’s infrastructure. From the user’s perspective, there is no “wallet” at all. From a data perspective, the wallet is the account. Whoever owns the SDK sees the creation events, the login frequency, and the integration points with the app. Kraken now has a direct window into a massive consumer finance application that is not its exchange. That is not wallet technology. That is market structure.
During my time at Dune, I learned to chase the transaction flow, not the tweets. The comparable event in this deal is not the change of control; it is the dependency graph. Polymarket and WalletConnect were Magic’s customers. After the sale, they become Kraken-adjacent components in a larger system. They may not feel like part of Kraken’s stack, but the gas flow says otherwise.
The truth in the tx is simple: this deal is not about wallets. It is about the first-party edge that comes from sitting inside the login flow of applications that are nominally outside the exchange ecosystem. Centralized exchanges are terrified of becoming dumb settlement layers. Magic gave them a way to buy the counter-party connection without building it from scratch.
Contrarian: Correlation Is Not Causation
The obvious read is that Kraken is buying a wallet to compete with Coinbase Wallet, MetaMask, and the rest of the embedded wallet field. That is correlation dressed up as causation.
Kraken does not need Magic’s code to build a wallet. The codebase was never the durable asset. The durable asset is the installed base of embedded wallet sessions across high-volume consumer dApps. Payward is buying the ability to appear in the login flow of apps that are currently viewed as competitors to centralized exchanges. There is no price tag for the ability to observe how a prediction-market user behaves when they think no exchange is watching.
The blind spot for most analysts is that they treat the wallet business as one business. It is not. The wallet business has at least four value streams: custody, private key management, front-end UX, and data access. Magic was selling custody and UX; Kraken is buying access. Newton Labs, free of the wallet business, is left with the company’s engineering talent and whatever new project it wants to build. But do not confuse the naming convenience with a roadmap. Newton Labs might be building toward cross-chain infrastructure, but the market should demand proof before it assigns a premium.
Data never lies, but deal announcements do. The announcement frames the transaction as a clean separation: wallet goes to Kraken, innovation stays with Newton Labs. That framing ignores the fact that the innovation is useless without distribution. Newton Labs is starting over with a blank page. Kraken is starting over with the login screens of Polymarket and WalletConnect on its side of the table.
Takeaway: What to Watch Next
If you want to know whether this deal is a good outcome, do not watch the press conference. Watch the wallet flows.
Over the next ninety days, check two things. First, does Polymarket’s embedded wallet continue to use the same SDK, or does it migrate to a Payward-controlled stack? Second, does Kraken start offering its own branded embedded wallet to its institutional clients? If the answer to both is yes, then the wallet business was never the product. The network effect was the product.
The deal closes a chapter for Magic Labs, but it opens a chapter for the wider market: exchanges are no longer buying exchanges. They are buying the login screen to the rest of the decentralized web. Follow the gas, not the narrative. Newton is the name on the door. Kraken is the hand in the drawer.