Bitcoin Miners Chase AI, Ethereum Plans for Quantum Risk, and Tether Gets Its Audit

Highlights

• Riot Platforms signed a 20-year AI data-center deal worth $9.1 billion
• Ethereum is putting quantum security, privacy and AI deeper into its roadmap
• Kalshi is bringing its market data onto Solana-based DoubleZero infrastructure
• Tether completed its first full independent financial audit with KPMG US
• SharpLink plans to stake about $200 million of ETH through Lido
• Hyperscale Data sold roughly 685 BTC to support its AI data-center plans
• Trezor disclosed a data breach affecting nearly 14,000 hardware-wallet customers

Crypto infrastructure is being pulled in several directions at once. Mining companies are finding new uses for power and data-center assets, financial platforms are testing faster market-data rails, and large holders are trying to make idle assets work harder rather than simply sitting on balance sheets.
‍

At the same time, the industry is confronting familiar control problems. Security failures still expose users outside the blockchain itself, while new infrastructure introduces fresh dependencies around custody, data delivery and operational resilience. The progress is real, but so is the cost of getting these systems wrong.

Riot Makes a $9.1 Billion AI Bet

Riot Platforms signed a 20-year agreement to provide 191 MW of data-center capacity to a major AI company, with Bloomberg identifying the customer as Anthropic. The reported value of the agreement is $9.1 billion, putting it far beyond the scale of a routine hosting contract and giving Riot a new business line built around the same resource that powers its Bitcoin mining operations: electricity.

For miners, the attraction is clear. Mining economics depend heavily on energy costs, network competition and block rewards. AI infrastructure offers a different revenue stream tied to long-term computing demand. A miner that already controls large sites, power connections and cooling has assets that can be repurposed for high-performance computing.

That does not mean the transition is simple. AI customers expect different uptime, networking and hardware standards, and long contracts create their own execution risk. Still, Riot's agreement shows how Bitcoin mining companies can turn access to power into something broader than a crypto-specific business. In some cases, their most valuable asset may be the infrastructure around mining rather than the mining itself.

Ethereum Looks Beyond Its Current Design

Vitalik Buterin outlined an Ethereum roadmap that gives greater attention to quantum resistance, privacy and AI-assisted formal verification. The plans also include native rollups and longer-term work that could eventually move parts of Ethereum beyond its existing EVM architecture.
‍

The interesting point is not any single feature. Ethereum is being designed on the assumption that today's technical choices will not remain adequate forever. Quantum computing creates a future security problem, privacy remains incomplete, and AI is being considered as a tool for verifying increasingly complex software.
‍

Native rollups are especially important because they would bring functions now handled largely by external scaling systems closer to Ethereum itself. That could reduce fragmentation and simplify some parts of the user experience, although it also raises difficult questions about how much functionality should move back toward the base layer.
‍

The roadmap remains long term, and some ideas may change before implementation. But Ethereum's developers are clearly thinking beyond incremental throughput gains and toward whether the network can handle another decade of changing cryptography, scaling demands and software complexity.

Kalshi Takes Market Data Off the Public Internet

Kalshi is integrating its order-book data with DoubleZero, a Solana-based network that distributes information over dedicated fiber rather than relying entirely on the public internet. The goal is to provide institutional traders with low-latency, machine-readable data closer to the infrastructure used by established financial exchanges.
‍

That matters because market quality depends on more than matching buyers and sellers. Professional trading firms also care about how quickly and consistently they receive data. If one participant sees an order-book update materially later than another, the difference can translate directly into execution quality and trading risk.
‍

Crypto markets have traditionally relied on internet connectivity that is flexible but uneven. Dedicated infrastructure brings them closer to the expectations of electronic market makers and high-frequency firms. It may also widen the gap between sophisticated trading operations and smaller participants.
‍

Kalshi's integration is therefore less about prediction markets alone and more about the professionalization of crypto-linked trading infrastructure. As digital-asset markets attract more institutional activity, data delivery is becoming part of the competitive stack.

Tether Finally Gets a Full Audit

KPMG US audited Tether's 2025 financial statements and issued an unqualified opinion. That goes further than the reserve attestations Tether had relied on for years and addresses one of the longest-running questions around the stablecoin issuer.
‍

For a stablecoin, trust rests heavily on the quality and availability of reserves. Users need confidence that the issuer can meet redemptions under pressure. A full financial audit gives a wider view of the business than periodic reserve attestations and brings a higher level of scrutiny to the issuer's reporting.
‍

There is still an important limitation: the full audited statements have not been made public. That reduces how much outside investors and users can independently examine the detail behind KPMG's opinion.

Even so, completing a full audit improves the quality of financial reporting around an asset used throughout crypto markets. Stablecoin infrastructure increasingly sits close to trading, lending, settlement and payments, so the standard of disclosure around the issuer matters well beyond the company itself.

SharpLink Puts $200 Million of ETH to Work

SharpLink plans to stake roughly $200 million of Ethereum through Lido and receive wstETH, with Anchorage Digital providing custody. The structure gives the company staking exposure while keeping the position liquid and usable across DeFi.

This is a different approach from simply holding ETH or locking it into conventional staking. Wrapped liquid-staking tokens can remain productive in other financial applications, allowing a treasury holder to earn staking rewards while preserving more flexibility around collateral and liquidity.

The trade-off is added complexity. SharpLink is not only exposed to Ethereum, but also to the operational and smart-contract risks associated with liquid staking, custody and any DeFi use of the resulting token. For large holders, capital efficiency is attractive, but every extra layer introduces another point that needs to be monitored.

Another Miner Trades Bitcoin for AI Infrastructure

Hyperscale Data sold around 685 BTC for approximately $43 million as it redirected capital toward its Michigan data-center operations and debt reduction. The sale is another sign that some miners are reassessing how they allocate capital between Bitcoin holdings and physical infrastructure.
‍

For years, holding mined Bitcoin was part treasury strategy and part statement of conviction. The economics change when capital has another use that may offer steadier revenue. Selling BTC to fund data-center expansion is not necessarily bearish on Bitcoin. It can simply mean power, land and computing capacity offer a better corporate return.
‍

The broader point is that miners are no longer tied to a single business model. Their access to power can serve Bitcoin, AI or a combination of both. That flexibility could make the sector less dependent on mining cycles, although it also means investors have to evaluate very different businesses under the same corporate name.

Trezor Breach Shows the Risk Outside the Wallet

Trezor said a breach at fulfillment partner ShipMonk exposed personal information belonging to nearly 14,000 hardware-wallet customers. The compromised data included information such as shipping addresses, creating a risk that goes well beyond ordinary spam.
‍

The private keys were not the problem. The weakness appeared in the surrounding operational chain. That distinction matters because hardware wallets are sold primarily on protection against digital theft, yet customer records held by suppliers can still identify who owns valuable crypto and where they live.
‍

For users, the likely threat is targeted phishing, impersonation and potentially physical intimidation. For wallet manufacturers, the incident is a reminder that self-custody security cannot stop at the device. Order systems, logistics providers and customer databases all become part of the security model once they contain information that can be linked to crypto ownership.

Bottom Line

The strongest common thread is that crypto infrastructure is moving closer to the way mature financial and technology systems operate. Miners are selling compute capacity to AI companies, trading platforms are investing in dedicated data networks, stablecoin issuers are adopting fuller financial reporting, and large asset holders are looking for ways to keep capital productive without giving up liquidity.

The weak points remain operational. A hardware wallet can protect keys while a logistics provider leaks the owner's address. Liquid staking can improve capital efficiency while adding smart-contract and custody dependencies. AI data centers can diversify miners while creating new execution demands. The industry is building more useful infrastructure, but the quality of controls around that infrastructure will determine how durable those gains are.