Week 22 · 2026

The core challenge in quantum computing is shifting from maximizing qubit count on a single chip to establishing scalable, functional connectivity. The concept of modular quantum computing proposes linking multiple smaller, specialized processors (modules) to function as a single, vastly larger machine. This approach mitigates the physical and manufacturing limitations encountered when attempting to create monolithic quantum chips.

There are two primary architectural strategies for modularity. The homogeneous view focuses on scaling a single qubit modality by networking many identical modules, as demonstrated by companies like Xanadu using photonic chips. Concurrently, the heterogeneous view—which is more ambitious—combines different qubit hardware types, such as superconducting qubits, trapped ions, and neutral atoms, to leverage the distinct physical strengths of each modality. For instance, trapped ions offer high fidelity, while neutral atoms excel at simulating physical systems.

The critical enabler across both strategies is the interconnect. The industry is treating the wiring and networking layer as foundational infrastructure. Major governmental efforts, such as DARPA's HARQ program, are actively working to develop both the physical interconnects and the software compilers necessary to assign tasks across diverse hardware types. This move signals a maturation of the field, mirroring the transition of classical computing from isolated machines to interconnected, high-performance networks. Ultimately, the focus is moving toward system integration, where various quantum components work together to solve complex, multi-faceted problems, such as simulating complex battery chemistry interactions.

The US government recently announced a significant $2 billion investment, allocating $100 million each to nine different quantum computing startups. These investments aim to support companies that are still years away from commercial products, making them critical for the future of quantum technology. However, the use of these funds has drawn immediate legal scrutiny from a member of Congress. Zoe Lofgren, a ranking member of the House Science, Space, and Technology Committee, stated that the deal is illegal because the money originated from the CHIPS and Science Act. This act specifically earmarked funds for microelectronics R&D, with a distinct focus on semiconductor technology. Lofgren pointed out that the technology used in quantum processors only overlaps partially with the intended scope of the funding. Furthermore, she argued that the money was meant to foster public/private research partnerships, which she believes the current deal does not constitute. The largest single recipient of the funding is a new entity called Anderon, which will receive $1 billion from both IBM and the government. Anderon is designed to function as a specialized foundry for fabricating quantum processing units, offering services to IBM and other interested companies. While Lofgren clarified that her objections are procedural and not a judgment on the technology's value, she maintains that the government must receive explicit Congressional allocation to proceed with such broad investments.

The article provides a technical analysis of internet scanning trends targeting AI infrastructure as observed by HoneyLabs. A significant portion of the observed activity involves scanning for misconfigured reverse proxies that expose Anthropic API paths, with a specific Dutch ASN identified as sending thousands of requests to enumerate these proxies. The report also highlights the rising prominence of port 11434, used by Ollama, as a target for discovery probes. The author notes that unauthenticated instances allow attackers to potentially pull arbitrary models via the /api/pull endpoint. A particularly noteworthy event involves a coordinated 45-minute sweep by a single IP address that targeted a wide array of AI-related credential files, including those for Claude, Gemini, DeepSeek, and DashScope, alongside traditional targets like AWS, Azure, and Docker. Furthermore, the analysis observes that the OpenAI-compatible API structure has become a standard target for reconnaissance across various self-hosted LLM stacks, such as vLLM, LM Studio, and LiteLLM. While the absolute volume of scans on AI-specific ports remains orders of magnitude lower than legacy targets like SMB or RDP, the author emphasizes the concerning upward trajectory of this activity. Finally, the analysis concludes that attackers are increasingly focusing on file paths where credentials are likely to be stored on disk, such as .env files and shell histories, rather than attempting to fuzz the API keys themselves.

The increasing flow of institutional capital into tokenized Real-World Assets (RWAs) is exposing a significant gap between the open, speculative nature of Decentralized Finance (DeFi) and the stringent requirements of traditional finance. Experts highlight that much of DeFi's existing infrastructure is not built to handle regulated assets, which necessitate robust identity frameworks and clear compliance protocols. While standards like ERC-3643 and the newly finalized ERC-7943 (uRWA) aim to standardize tokenized securities, the primary challenge remains interoperability across disparate compliance systems, custodians, and exchanges. Furthermore, institutional adoption is heavily driven by assets offering predictable cash flows and established legal structures, favoring systems that enable faster settlement and programmable collateral.

Beyond technical standards, privacy represents a major hurdle; large financial entities cannot afford to expose their entire portfolio activity on public blockchains. This has driven the adoption of specialized, permissioned networks, such as Canton, which allow data visibility to be restricted only to relevant participants. Looking ahead, the infrastructure must also prepare for machine-driven finance. As AI agents begin to move capital autonomously, they will require RWAs that possess a standardized digital form readable and actionable across multiple jurisdictions. Ultimately, while new tokenization standards are vital, the industry must build flexible, non-opinionated frameworks capable of connecting traditional financial processes with the decentralized ledger, thereby bridging the gap between crypto-native infrastructure and regulated global markets.

A recent report from the American Foreign Policy Council analyzes the strategic implications of placing mass drivers—electromagnetic catapults—on the Moon. These devices, which use powerful magnetic fields, could launch satellites and probes into deep space without the need for conventional, costly chemical propellants. While proponents, including SpaceX, view them as essential for bootstrapping a lunar off-world economy, the report stresses their inherent dual-use nature. This duality means that while they facilitate peaceful missions, they also carry the potent military capability to operate as unassailable, undetectable first-strike platforms.

The technology, rooted in concepts dating back to the 1970s, could enable a sustained, high-throughput logistics chain. However, the report warns that the United States must act proactively to shape the rules of the cislunar frontier, as failure to invest could allow competitors, notably China, to gain strategic dominance. Potential military applications include launching Kinetic Energy Impactors (KEI), Anti-Satellite (ASAT) weapons designed to disrupt other spacecraft, and even Nuclear Reentry Vehicles (RV). Although the United Nations Outer Space Treaty prohibits certain activities, the mixed-use nature of mass drivers complicates international regulation.

While the concept is advanced by major industry players, the technology is not yet industrially scalable. The primary barriers remain logistical and technical, limiting current deployments to small payloads. Nevertheless, the potential for lunar-based mass drivers to become an unparalleled source of space power makes them a critical focus in current space power and geopolitical strategy.

Blockchain infrastructure platform Paxos has announced that its subsidiary, Paxos Securities Settlement Company, has received SEC registration as a clearing agency. This status makes Paxos the only blockchain-native firm approved to provide central securities depository services in the United States, an approval deemed critical for the ongoing convergence of blockchain technology and traditional capital markets. Clearing agencies are essential for ensuring the clean execution of securities trades by verifying, matching, and settling the exchange of money and securities. The SEC-approved blockchain clearinghouse status removes significant infrastructure barriers for both banks and brokerages seeking to implement crypto-based services. Paxos's journey toward this registration involved a multi-year effort, starting with a no-action letter from the SEC in 2019, which allowed the company to pilot a blockchain-based settlement service for US equities. This pilot demonstrated the capability of blockchain post-trade infrastructure to deliver same-day settlement, operational efficiency, and cost reduction within a fully regulated framework. The company's regulatory history is noted for its resilience, having previously navigated enforcement actions and compliance orders from both the SEC and the New York Department of Financial Services (NYDFS) regarding stablecoin issuance. Despite these challenges, the successful registration solidifies Paxos's position as a key player facilitating the modernization of financial market infrastructure through distributed ledger technology.

The article outlines a philosophical and structural pivot for the Ethereum Foundation (EF), asserting that the EF must transition from being perceived as the central steward to acting as a specialized, opinionated node. The author critiques the general industry trend of sacrificing idealistic principles for financial gain and corporate expediency. The EF's limited resources necessitate making difficult choices to maintain focus on the core mission: ensuring Ethereum remains a censorship-resistant, open, private, and secure system.

Technically, the author argues that while scalability is important, the deepest dimension of improvement lies in the CROPS dimension. Key technical initiatives include pursuing provably bug-free Ethereum through formal verification, a goal previously deemed impossible. Furthermore, the network must maintain available chain consensus, uniquely combining traditional BFT safety under asynchrony with PoW-style safety under synchrony. Significant work is also focused on intermediary minimization, utilizing standards like FOCIL and EIP-8141 to secure the transaction sending process at the user layer, thereby mitigating current points of systemic fragility.

These goals, while ambitious, are presented as compatible with high TPS and are crucial for maintaining Ethereum's status as a fundamentally robust and impressive technology. The EF is therefore committing to a path of principled specialization, even if it requires making tough decisions regarding its operational scope and the involvement of external contributors.

The Depository Trust & Clearing Corporation (DTCC), a central institution in U.S. market infrastructure overseeing more than $114 trillion in assets, has announced a plan to connect its tokenized securities platform to the Stellar (XLM) network. This collaboration with the Stellar Development Foundation aims to support the issuance, settlement, and lifecycle management of blockchain-based versions of traditional securities, including highly liquid assets such as major indices and U.S. Treasury debt instruments. The integration is expected to make tokenized assets available on the Stellar blockchain during the first half of 2027. This initiative is a core component of DTCC's 'multi-chain' strategy, which seeks to create an open, interoperable digital infrastructure where assets can move across different blockchain networks rather than being restricted to a single platform. The move follows an SEC no-action letter granted in December 2025, which allows DTCC to tokenize a specific set of assets including ETFs and Russell 1000 stocks. The push toward tokenization is part of a larger movement within Wall Street to utilize blockchain technology to reduce settlement delays, optimize collateral, and enable markets to operate outside of standard trading hours. Other major financial players are also making moves in this space, such as Nasdaq's development of blockchain-based share infrastructure with Kraken's parent company, Payward, and Intercontinental Exchange's (ICE) partnership with OKX.

Bitwise has positioned Hyperliquid as a key platform poised to become foundational infrastructure for the future of finance. The firm is capitalizing on this narrative by launching HYPE ETF products, offering investors the benefit of staked HYPE to maximize yield. Furthermore, Bitwise aligns its own interests by allocating a portion of management fees toward purchasing HYPE tokens for its balance sheet.

According to Bitwise's research, Hyperliquid is expanding far beyond its initial niche, showing robust growth in perpetual futures, prediction markets, and spot trading. These trends, combined with the increasing use of stablecoins and tokenized equities, suggest the platform's long-term viability. The institution points to partnerships, such as the one involving USDC liquidity, as concrete signs of institutional momentum.

From a tokenomics perspective, the platform benefits from a buy-and-burn mechanism where nearly all generated fees are used to acquire and retire HYPE tokens. Bitwise compares this self-regulating model to traditional stock buybacks, creating an understandable narrative for mainstream investors. The firm believes that the current regulatory climate allows projects like Hyperliquid to launch with stronger incentives than in previous cycles.

Despite the bullish outlook, the analysis acknowledges significant risks. These include potential U.S. regulatory scrutiny over decentralized perpetual futures markets and macro-level concerns such as inflation and geopolitical instability. The article concludes by noting a positive shift in institutional dialogue, where wealth managers are increasingly discussing tokenization and stablecoins rather than questioning the fundamental legitimacy of crypto assets.

Quantinuum has submitted an amended S-1A filing to the U.S. SEC, significantly updating the details from its previous IPO filings. The most immediate financial update concerns the offering size, which involves 21,052,632 shares of Class A common stock, priced between $45.00 and $50.00 per share. This establishes a post-IPO valuation range between $11.43 billion and $12.70 billion.

A major operational highlight is the non-binding Letter of Intent with the U.S. Department of Commerce. Under the CHIPS Act of 2022, Quantinuum is slated to receive up to $100 million, disbursed in tranches. This funding is tied to developing critical quantum components, specifically the fabrication and testing of integrated optical waveguides and diffraction gratings, alongside custom-designed ASICs. The agreement mandates that funded innovations must be produced exclusively in the U.S. for a period of ten years.

In corporate governance, the board structure was formalized, moving from 'Director Nominees' to actively seated directors. Furthermore, the company updated its executive compensation policies, including increases to the CEO's base salary and performance targets, which will take effect upon the IPO closing.

Finally, the filing completed financial statement disclosures. Pro forma financial data is now fully populated, and the total registration and issuance expenses were finalized at $18,300,000. These updates provide a comprehensive view of the company's market readiness, technological focus, and strategic government backing.

Researchers at Multiverse Computing developed a breakthrough method to reduce uncertainty in Large Language Models (LLMs) by leveraging quantum computing. Their work marks the first demonstration of 'quantum enhancement' in a production-scale LLM, addressing the challenge of scaling classical AI infrastructure. The core of the technique involves using Cayley-parameterized unitary adapters (CUAs), which are small, trainable mathematical matrices inserted into the LLM's layers. These adapters are trained classically but executed on quantum hardware, specifically the 156-qubit IBM Quantum System Two superconducting QPU.

The researchers applied this hybrid model to Llama 3.1 8B, achieving a 1.4% reduction in perplexity while adding only a negligible number of parameters. Perplexity, a key metric, measures a model's ability to predict the next word; a lower score indicates higher capability. Crucially, the quantum-enhanced model not only showed improved metrics but also correctly answered complex questions in fields like astronomy and biology that the original, non-enhanced model had failed to address. This success suggests that quantum hardware can provide genuine advantages over purely classical paradigms. While the primary technical hurdle remains managing quantum noise, the research provides a clear, viable path toward developing quantum-hybrid AI systems capable of achieving a quantum advantage, thereby overcoming current development bottlenecks in classical computing.

The FROST attack is a novel side-channel vulnerability that enables malicious websites to spy on a user's device activity by analyzing SSD I/O interactions. While browser sandboxing prevents direct access to the host system, JavaScript can still measure the timing of I/O operations. The attack works by performing random reads from a large Origin Private File System (OPFS) file to measure SSD contention caused by other concurrent user activities. These latency traces are then processed through a pre-trained convolutional neural network (CNN) to classify and fingerprint the various applications and websites active on the host system. The researchers demonstrated the full attack on an M2 Mac and confirmed the underlying primitive works on Linux, though Windows was not tested. However, the attack faces significant practical limitations: it requires the creation of extremely large files, likely over one gigabyte, which could be detected by users, and it is restricted to the same physical SSD where the browser's OPFS is located. To mitigate this threat, researchers suggest that users close unnecessary tabs and recommend that browser developers implement maximum size limits for OPFS files. Currently, there are no reports of FROST being used in real-world attacks, and the findings are slated for presentation at the DIMVA conference.

The US Space Force, through its Space Systems Command, has announced a $2.29 billion firm-fixed-price agreement with SpaceX to develop the Space Data Network (SDN) Backbone. This project is intended to serve as the fundamental communications layer for a rearchitected sensor-to-shooter targeting network, ensuring that military sensors and weapons systems are connected continuously, globally, and securely. The technology powering this backbone will be based on the architecture originally developed for SpaceX's Starlink global Internet constellation and is expected to utilize the Starshield platform, which is specifically designed for military applications. The network will feature an expanded, optically interconnected mesh of satellites operating in low-Earth orbit (LEO), providing both tactical and broadband communication services. This strategic procurement marks a significant shift in approach, following the stalling of previous Pentagon-led initiatives, such as those managed by the Space Development Agency (SDA). By leveraging commercial innovation through SpaceX, the Space Force aims to accelerate the deployment of a highly resilient and high-speed communications infrastructure essential for modern space-based warfare and global situational awareness.