The State of Crypto Research: Alpha or Paid Marketing?

The crypto research landscape is rapidly evolving, characterized by an increase in available data, the influence of AI, and a proliferation of research providers. This has made it challenging for investors to discern valuable insights from noise, leading to a growing need for critical evaluation of sources. Ishmael from Bitwise and Patrick from Serotonin discussed the current state of crypto research. Ishmael noted that Bitwise maintains a broad, top-down view of the market, incorporating macroeconomics and increasingly leveraging AI for data gathering, though emphasizing that AI is not a replacement for human analysis. Bitwise focuses on areas like stablecoins and tokenization due to client interest. Serotonin, primarily a PR and marketing agency, is expanding into research, having recently released its first report on on-chain credit. They aim to provide both public educational resources and white-label research for partners. A significant challenge in crypto research is managing conflicts of interest, given that many research providers have vested interests. Ishmael highlighted that Masari, where he previously worked, operated on a paid research model, emphasizing independence and data-driven analysis. They made it clear to clients that reports would be objective, allowing readers to interpret findings. Patrick echoed this, noting that firms have different incentives, some aligned with VCs or protocols, while others sell subscriptions. Bitwise anchors its research in transparent data, which they believe keeps it honest and credible. He also pointed out a trend where protocols or companies paying for reports include founder quotes, which can be opinionated, while the core research remains objective. Venture funds, while sometimes having stakes, can provide high signal research when they have liquid positions and can exit if their conviction wanes. The distinction between locked-in, illiquid VC positions and liquid fund positions is crucial for understanding potential bias, though this nuance is often lost on retail investors. This proliferation of conflicted research, similar to issues in traditional media, is undermining trust and driving investors towards doing their own research. Both Ishmael and Patrick shared their experiences with producing research that might not always be positive. At Masari, clients reviewed reports but material changes to data or framing were resisted. They maintained that including data on downtrends was essential for transparency and credibility, especially in down quarters, to showcase resilience and rebound potential. Bitwise follows a similar practice, not removing downward trending charts in their quarterly reviews to preserve credibility for when those charts are on the uptrend. The impact of AI on research generation is a double-edged sword. AI can easily produce lengthy reports, potentially leading to an influx of "slop" and low-quality content. Ishmael believes AI is most effective when combined with human expertise, as AI currently lacks critical thinking and may simply regurgitate existing narratives or draw incorrect conclusions. Bitwise uses AI for information gathering and to streamline processes, compressing research timelines, but human oversight is crucial for generating alpha and meaningful conclusions. Patrick emphasizes the "human touch," finding AI difficult to use for generating institutional-grade research from scratch. He notes that AI-driven content often lacks the narrative flow and intuitive reasoning of human-written reports. However, AI can be valuable as a sounding board for validating ideas and categorizing information, as Patrick used with Grok for his on-chain credit report. AI is also useful for creating visuals and speeding up manual, repetitive tasks. AI falls short in its ability to craft compelling narratives and complex reports. For shorter pieces, AI can be a useful starting point, but for comprehensive reports, human insight is needed to structure the narrative and tell a cohesive story. A purely AI-generated 30-page report would likely be less informative than one written by a human researcher. Regarding free tools, the accessibility of crypto data has increased with platforms like DeFi Llama and Token Terminal, though some are becoming less free. Raw on-chain data remains accessible but requires significant effort to process. Ishmael mentioned that crypto-specific AI LLMs or chatbots are more helpful than general tools like ChatGPT because they possess relevant context. Serotonin is actively implementing AI, using tools to directly plug data sources and facilitate data analysis. The demand for institutional-grade crypto research is growing, attracting major Wall Street firms like Fidelity and Goldman Sachs. This shift suggests that research reports are increasingly targeted at an institutional audience, as individual investors now have easier access to data and tools. Crypto-native firms like Bitwise and Serotonin compete with larger institutions by maintaining a deeper alignment with the crypto space, its ethos, and its community. While Wall Street firms focus on broad market research, crypto-native firms can offer more specialized, thesis-driven research and leverage their existing connections and understanding of decentralized principles. Serotonin's model involves producing white-labeled research for clients, helping them build their brands and distribute content through their own channels. This approach allows brands to become their own content creators, a trend mirroring the media landscape. Ishmael highlighted the proliferation of "stretch back stablecoins" and tokenized versions of "stretch," a preferred stock that has funded Bitcoin's ascent, as an underreported trend. Patrick remains bullish on on-chain credit, expecting it to offer higher, uncorrelated yields. He also mentioned the growing interest in "ownership tokens" that align more directly with protocol interests, differentiating between platforms focusing on token mechanics and those emphasizing transparency. For individual investors seeking data and alpha, Ishmael recommends data sources like Dune, Glassnode, and keeping up with crypto Twitter, while acknowledging the need to filter noise. Serotonin has launched a curated feed on its platform to address this, manually selecting insightful content. To improve trust in research, transparency is key. Disclosures about backers, ties to protocols, and whether a report is commissioned are crucial. Ishmael suggested that a 12-month lockup for tokens, as proposed in the Clarity Act, could also enhance trust. Patrick added that stating the purpose of a report (e.g., transparency, brand awareness) and understanding the entry price for venture fund positions would be beneficial. Reliable research outlets include data aggregators like Dune and Glassnode, news outlets like Blockworks, The Block, CoinDesk, and The Defiant. On the research firm side, Galaxy Research, Blockworks, Messari, and independent shops like Four Pillars and TLC are respected. Looking ahead, Serotonin will continue publishing primers on on-chain credit players. Bitwise anticipates a strong second half of the year with new research accompanying partnerships and product launches.