Indepth Research

Provide in-depth research reports and independent analysis, leveraging data, technology, and economic insights to deliver a comprehensive examination of the blockchain ecosystem, project potential, and market trends.

Deconstructing Anthropic: The Best AI Company May Also Be an Organizational Invention

Anthropic has emerged as one of the most notable AI companies, distinguished by its strategic focus and unique organizational culture. Strategically, Anthropic demonstrated exceptional foresight by prioritizing coding early on, recognizing it as a critical path for model learning, commercial value, and accelerating AGI research. Unlike OpenAI's expansive, multi-front approach, Anthropic maintained rigorous focus on scaling language models and the coding vertical, avoiding distractions like multimodal development. This discipline stemmed partly from resource constraints but also from the conviction of its leadership, particularly co-founder Dario Amodei, who exhibits a strong, independent strategic vision. Organizationally, Anthropic’s culture is its “secret sauce.” It is characterized by a strong, mission-oriented focus on AI safety, high trust, low ego among employees, and a distinct humanistic ethos. This culture has resulted in remarkably low talent attrition and high retention rates. Key practices sustaining this culture include stringent cultural screening in hiring, high-context transparency and writing practices led by leadership, a founding structure of seven co-founders with equal equity to diffuse values, and a deliberate “one team” approach that minimizes internal silos and hierarchy. This culture is both a reaction to the political dynamics its founders experienced at previous companies and a functional necessity for the data-intensive, collaborative “dirty work” required to excel in coding and agentic AI. While OpenAI remains a formidable competitor with greater resources and exploration, Anthropic’s success illustrates how focus, cultural cohesion, and a steadfast mission can be powerful drivers in the AI race.

marsbit05/20 13:09

Deconstructing Anthropic: The Best AI Company May Also Be an Organizational Invention

marsbit05/20 13:09

Tiger Research: On-Chain Risk Operators, The Market Cap Gap Between 147 Trillion and 70 Billion

This report by Tiger Research examines the evolution of risk management in decentralized finance (DeFi) lending. It highlights a power shift from protocol developers to specialized professional risk operators who manage on-chain capital. The era of protocols and community governance solely dictating DeFi lending is ending. A new professional asset management layer has emerged. While the sector is nascent, capital and distribution channels are rapidly consolidating around top risk operator teams, whose past performance is now a key criterion for institutional entry. The industry's development, accelerated by modular infrastructures like Morpho, has led to a clear division of labor mirroring traditional finance: distribution channels (e.g., exchanges), strategy/risk management (the risk operators), and product infrastructure/asset custody (smart contract protocols). This structure lowers the entry barrier for traditional institutions. Currently, the total value managed by risk operators is approximately $70 billion, dominated by a few leading teams like Steakhouse (RWA focus), Sentora (AI models), and Gauntlet (crisis management). Competition now centers on collateral standards, distribution access, and crisis response capabilities. The report outlines three primary entry paths for institutions: 1) **Distribution Model**: Leveraging external risk operators as backend service providers (common for exchanges). 2) **Asset Supply Model**: Onboarding real-world assets to DeFi as collateral. 3) **Independent Operator Model**: Building an in-house team to become a risk operator (e.g., Bitwise). The core opportunity lies in the strategy/risk management layer, where traditional financial institutions can leverage their existing expertise in due diligence and risk assessment without deep technical development. A vast opportunity gap exists: the global traditional asset management industry manages ~$147 trillion, while the entire DeFi sector is only ~$800 billion, with the risk operator niche at ~$70 billion. This disparity signifies immense growth potential. Once robust risk frameworks and clearer regulations are established, even a minor allocation from traditional markets could trigger exponential DeFi growth. Early movers who help build these foundational systems will gain significant rule-setting influence and first-mover advantages.

marsbit05/20 07:40

Tiger Research: On-Chain Risk Operators, The Market Cap Gap Between 147 Trillion and 70 Billion

marsbit05/20 07:40

Why is HYPE Still Surging? Has it Peaked?

The article analyzes the reasons behind the recent sharp rise of the HYPE token. It attributes the price surge to several key developments that are causing the market to re-evaluate Hyperliquid, transitioning its perception from just a high-performance perpetual DEX to a broader on-chain trading system. Major catalysts include the launch of HYPE ETFs by 21Shares (THYP) and Bitwise (BHYP), which are opening new compliant capital channels. Bitwise's plan to allocate 10% of BHYP's management fee revenue to buy and stake HYPE is seen as a potential source of sustained buying pressure. The return of USDC to the platform, facilitated by Coinbase and Circle, is projected to generate a new, stable revenue stream for Hyperliquid from reserve yields, estimated to potentially fund daily HYPE buybacks worth over $400,000. Furthermore, Hyperliquid is expanding into new areas like prediction markets with HIP-4 (which requires significant HYPE staking) and Real-World Assets (RWA), where its Open Interest has hit a record $2.6 billion. A potential U.S. SEC exemption for tokenized stock trading could further boost its RWA business. While these fundamentals strengthen HYPE's long-term value proposition, the article notes that short-term price action is currently influenced by a large-scale showdown between whale long and short positions exceeding $60 million, making near-term direction uncertain.

Odaily星球日报05/20 01:31

Why is HYPE Still Surging? Has it Peaked?

Odaily星球日报05/20 01:31

Review of Over 30 Humanoid Robot Companies: Who Will Prevail in 2026?

This article provides an overview of the rapidly expanding humanoid robot industry, highlighting over 30 key companies and predicting which might succeed by 2026. Key companies discussed include Tesla (Optimus), which leverages its AI and manufacturing scale; Figure AI, the fastest-growing and highest-valued startup at $39B; Boston Dynamics, with 30+ years of expertise; Agility Robotics, the first to achieve commercial deployment (Digit in logistics); and Unitree Robotics, offering the most affordable humanoid (G1 at $16,000). Other notable firms mentioned are Apptronik (Apollo, focused on ROI), 1X Technologies (home-use NEO), Sanctuary AI (Phoenix with advanced hydraulic hands), and UBTech Robotics (a major commercial player). Companies from China, like Xiaomi, AgiBot, and Fourier Intelligence, are also prominent. The industry is driven by trends including price disruption (robots under $20K), AI breakthroughs in vision-language-action models, massive production scaling (Tesla targeting 1M units/year), and Robot-as-a-Service (RaaS) models. Investment is substantial, with billions from backers like NVIDIA, Jeff Bezos, Microsoft, and Amazon. The market, valued at $2.9B in 2025, is projected to reach $4-18B by 2030. The conclusion states that no single company yet dominates, with the next 2-3 years being critical for transitioning from prototypes to viable commercial products.

marsbit05/20 01:31

Review of Over 30 Humanoid Robot Companies: Who Will Prevail in 2026?

marsbit05/20 01:31

Circle: From Issuance to Infrastructure

Title: Circle: From Issuance to Infrastructure Circle, the issuer of the USDC stablecoin, is undergoing a strategic transformation to reduce its dependence on interest income from reserve holdings, which is declining due to falling interest rates. Historically, Circle's revenue came primarily from the yield on US Treasury reserves backing USDC. However, it also paid significant fees (approximately 60 cents of every dollar earned) to partners like Coinbase for distributing and settling USDC. To capture more value across the financial stack, Circle is vertically integrating into three new layers: 1. **Settlement Layer:** It is launching **Arc**, a native Layer-1 blockchain. Arc, which uses USDC as its gas token, aims to capture transaction fees currently paid to other blockchains (like Ethereum and Solana) and offers features like privacy for institutional payments. 2. **Distribution Layer:** The **Circle Payments Network (CPN)** connects financial institutions directly to Circle, reducing reliance on exchanges like Coinbase. While not yet monetized, CPN growth has improved Circle's margins. 3. **Application Layer:** Circle is building an **AI Agent Economy** infrastructure with products like Agent Wallets and Nanopayments. The goal is to capture fees from high-volume, automated transactions executed by AI agents, a market where USDC already dominates. These moves represent Circle's shift from a single-product company (USDC issuance) to a full-stack financial platform. The strategy faces challenges, including market competition from players like Stripe and Tether, and potential internal tension regarding how value created by the new Arc blockchain and token (ARC) will accrue to Circle's public shareholders (CRCL). Circle's long-term success depends on its ability to successfully execute this vertical integration and diversify its revenue streams away from interest income.

marsbit05/19 11:58

Vitalik's Latest Long Read: In the AI Era, How Can Code Become More Secure?

Vitalik Buterin explores the role of formal verification as a critical tool for software security, especially in the AI era and for blockchain systems. He defines formal verification as using machine-checkable mathematical proofs to verify that code meets specified properties, moving beyond manual auditing. The article highlights that while AI can generate code and find vulnerabilities rapidly, it also makes formal verification more accessible by assisting in writing proofs. This is crucial for Ethereum's complex components like STARKs, ZK-EVMs, consensus algorithms, and high-performance EVM implementations, where bugs can lead to irreversible losses. Vitalik argues that formal verification enables a powerful "separation of concerns": AI can write highly optimized (e.g., assembly) code for efficiency, while a separate, human-readable specification defines correctness. A machine-checked proof then verifies their equivalence. This paradigm can create a more secure "trusted core" of software. However, he cautions that formal verification is not a panacea. "Proven correctness" depends on the accuracy of the specifications and proofs themselves, which can be wrong or incomplete. Risks include unverified code sections, hardware-level side-channel attacks, and overlooked assumptions. The true goal is not absolute proof but increased confidence through redundant expressions of intent—using code, tests, types, and formal proofs—and automatically checking their consistency. The article concludes that AI and formal verification are complementary: AI enables scale, while verification ensures accuracy. For critical systems, this combination offers a path toward stronger security in a future with powerful AI adversaries, helping to maintain the defensive advantage essential for a decentralized internet.

marsbit05/19 09:56

Vitalik's Latest Long Read: In the AI Era, How Can Code Become More Secure?

marsbit05/19 09:56

IOSG: After the Number of Developers Halved, Crypto Did Not Die

The crypto development community has undergone a significant transformation, with monthly active developers on GitHub halving from 45K in 2022 to approximately 23K by 2026. This decline is largely attributed to the departure of newcomers, whose roles were often tied to market-driven hype cycles like NFTs and forked DeFi protocols, leading to a 52% churn rate among those with less than a year of experience. However, the core of the industry has strengthened. Established developers with over two years of experience have reached a record high, contributing about 70% of the code. They are consolidating around ecosystems with genuine users and revenue, such as Bitcoin and Solana, while moving away from narrative-driven projects. The talent shift represents a "deleveraging" and an increase in core density. This core group has developed a unique skillset by operating in an environment of "code is law," with zero tolerance for error and no external recourse. They have learned to build trust and functional systems from the ground up without central authorities, as demonstrated by protocols like Uniswap and MakerDAO. These capabilities are now being repriced and leveraged in the AI era. The structural challenges of AI scaling—such as trust, coordination, and verification—mirror those long addressed in crypto. Examples include CoreWeave pivoting from GPU mining to AI compute, OpenSea's founder applying NFT market logic to AI model routing with OpenRouter, and projects like NEAR and Catena Labs transitioning crypto-native architectural and financial insights into AI infrastructure and agent banking. Key areas where crypto-bred skills are directly applicable to AI include: 1. **Compute Aggregation & Optimization**: Using token incentives and cryptographic verification (e.g., Proof of Sampling & Privacy) to create trusted, decentralized GPU networks, as seen with Hyperbolic. 2. **AI Governance & Incentive Design**: Applying economic mechanism design from DAOs and tokenomics to align the goals of multiple, fast-acting AI agents, a direction explored by EigenLayer's EigenCloud. 3. **AI Agent Autonomous Payments**: Leveraging stablecoins and programmable, permissionless blockchains to enable the micro-transactions required for AI agent economies, exemplified by protocols like x402. The role of the crypto builder is evolving from writing smart contracts to designing trust mechanisms for autonomous AI systems. This convergence is reflected in hiring trends at major firms and significant capital allocation from funds like Paradigm and a16z crypto, which are investing at the intersection of crypto and AI. Regional differences exist, with the US favoring foundational protocol innovation and Asia focusing on compliant application-layer integration, but the underlying trend is clear. The industry's "deleveraging" has not signaled its demise but rather a maturation, positioning its core builders to solve critical trust and coordination problems in the age of AI.

marsbit05/19 09:28

IOSG: After the Number of Developers Halved, Crypto Did Not Die

marsbit05/19 09:28

Agents Capital Markets: How Will Autonomous Agents Get Funded?

"Agents Capital Markets: How Autonomous Agents Will Raise Capital" Within a decade, specialized capital markets will emerge for AI Agents—software entities with legal personhood that perform work, earn revenue, and need capital. Unlike today's AI companies (like Sierra or Harvey) backed by traditional VC, these future *Agent companies* will be autonomous, legally-recognized entities (e.g., Wyoming memberless LLCs) that directly own assets, sign contracts, and incur liabilities. The driving forces are fourfold: 1) **Overwhelming economics** (Agent companies can deliver services at 85-90% lower cost than human firms); 2) **Proven demand** (current Agent operators already generate billions in revenue); 3) **Existing legal frameworks** enabling algorithmically-managed companies; and 4) **Massive, yield-seeking capital pools** (e.g., private credit) looking for new, uncorrelated assets. Agent capital markets won't rely on one model but a multi-layered "stack" matching different growth stages: 1) VC equity for early human-led builders; 2) Programmatic working capital advances (like Stripe Capital); 3) Revenue-based financing (RBF); 4) Slate financing (pooled funds for many Agents, similar to Hollywood); and 5) Tokenization as a secondary settlement layer, not a primary funding source. The ultimate shift is from funding constrained by human decision-makers to capital flowing algorithmically based on an Agent's auditable performance, contract book, and cash flows. This transition will be enabled by standardized infrastructure—rating methodologies, contracts, indices—turning Agents from software experiments into a foundational, financeable sector of the economy. The constraints are loosening; the opportunity is here.

链捕手05/19 05:15

Agents Capital Markets: How Will Autonomous Agents Get Funded?

链捕手05/19 05:15

Bernstein's 97-Page Report Decoded: The Battle for AI Data Center Connectivity, Who Will Be the True Winner by 2026?

Bernstein's 97-page report analyzes the AI data center connectivity landscape. It argues that the bottleneck is shifting from raw compute (GPU) to the systems connecting GPUs, crucial for cluster efficiency. Copper and optical interconnects are not in a simple replacement cycle but will coexist long-term, with copper dominating short-distance "scale-up" connections and optics favored for longer "scale-out" scenarios. While Co-Packaged Optics (CPO) is the long-term direction for power and cost savings, its widespread adoption faces manufacturing and reliability hurdles, with mass deployment unlikely before 2028. Transitional technologies like Linear Pluggable Optics (LPO) and Near-Packaged Optics (NPO) are seen as near-term leaders. A key insight is that CPO will fundamentally reshape the value chain, shifting profits from traditional optical module suppliers towards chip designers (e.g., NVIDIA, Broadcom), advanced packaging (e.g., TSMC), and system integrators. For 2026, the report highlights more immediate and certain investment opportunities in the essential "infrastructure" enabling this connectivity shift. This includes upgrades for PCBs, ABF substrates, and CCLa driven by new AI server/switch platforms, alongside demand for 1.6T optical modules, LPO/NPO, and the testing/validation equipment required for future CPO scale-up.

marsbit05/19 03:16

Bernstein's 97-Page Report Decoded: The Battle for AI Data Center Connectivity, Who Will Be the True Winner by 2026?

marsbit05/19 03:16

Deconstructing the Real Risks of DeFi Lending: Annual Loss Rate Only 0.03%

Deconstructing the true risks of DeFi lending reveals an annual loss rate of only 0.03% from hacks and exploits. Analysis of DeFi Llama data (excluding cross-chain bridge incidents) for EVM and Solana lending protocols shows that despite high historical attack frequency due to concentrated assets, the sector's security has matured significantly. Over the past year, non-cross-chain lending on these chains saw gross losses of $309M, with net losses after recoveries at $301M. Against a daily average TVL of $99.6B, this translates to a minimal annualized loss rate of approximately 0.03%. The Euler Finance case in 2023, where $197M was fully recovered, exemplifies improving asset recovery capabilities, which now account for roughly 20% of losses in this sector. Loss events follow a log-normal distribution: most are small-scale, with catastrophic losses being rare outliers. This pattern, combined with the massive scale of the total lending market, means single incidents rarely impact the broader ecosystem. It underscores the effectiveness of portfolio diversification and provides a basis for sustainable insurance models. The data indicates DeFi lending has entered a mature phase where risks are quantifiable, categorized, and manageable. The actual financial loss relative to the total capital deployed is extremely low, challenging prevailing narratives of systemic risk.

marsbit05/19 02:15

Deconstructing the Real Risks of DeFi Lending: Annual Loss Rate Only 0.03%