Stack Articles

Bringing Rigor to Crypto Asset Valuations

Bitcoin is a digital commodity and a decentralized alternative to fiat currencies. It is often described as “digital gold” and, in our view, should be valued that way. But many other crypto assets can be valued with cash flows. At Grayscale, we believe that rigorous fundamental valuation will play a defining role in the next chapter of digital asset investing.

Onchain applications typically operate identifiable business models, generating fees from trading, lending, and other financial activities. After paying out supply-side costs1, remaining earnings can be measured and compared across protocols. Much like traditional businesses, these economic fundamentals provide a foundation for determining digital asset valuations.

Crypto protocols also have several mechanisms for returning value to tokenholders. Buybacks reduce circulating supply, token burns permanently reduce total supply, and staking distributions reward tokenholders directly.

At the same time, new token issuance and unlocks can offset these benefits through dilution. Understanding the relationship between capital returns and token supply is therefore critical to evaluating the economic value accruing to each token.

Grayscale Research evaluates tokenholder economics through measures such as Tokenholder Yield2, which captures distributions to tokenholders less supply inflation, and Adjusted Earnings Yield3, which measures protocol earnings relative to market capitalization after adjusting for inflation. Together, these metrics can offer a more complete picture of how onchain protocols generate economic value and how much of that value may ultimately accrue to tokenholders.

Key Takeaway: Crypto valuations don't have to rely on narratives alone. Onchain applications generate measurable economic value through earnings and capital returns, providing investors with a foundation for fundamental valuation.

Exhibit 1: Tokenholder yield and adjusted earnings yield vary widely by protocol

Fees paid to parties providing services or capital to the applications (e.g., liquidity providers, lenders or other suppliers).

Tokenholder Yield = (Staking Rewards + Buybacks − Issuance) ÷ Circulating Market Cap

Adjusted Earnings Yield = (Total Earnings ÷ Circulating Market Cap) − Circulating Supply Inflation

Compute Stocks Slide, But Demand for Compute Is Poised to Stay

AI compute infrastructure stocks have taken a beating recently as investors fretted about rising borrowing costs against the companies’ high financing needs. An index of these equities—which includes CoreWeave and Nebius—has lost roughly 27% of its value since the end of September.1 Most of these companies are early stage, capital-intensive businesses2, so investors’ concerns are not unfounded.

But we would like to stress one key factor of which investors, in our view, should not lose sight: We believe demand for compute is going up, not down. Even as stock prices corrected, demand (and prices) for compute remained.

To be sure, financing costs are a key consideration, but these businesses have many other growth drivers (Exhibit 1). As we mentioned, we believe the most important factor is demand for compute. We expect compute demand growth to remain high as falling token prices and rising use of AI agents result in much higher inference needs. If compute demand growth and compute prices remain high, we believe these companies could be well positioned to benefit.

There are also other factors relevant to the energized compute layer, including slower depreciation rates for Graphics Processing Units (GPUs) (often an asset on balance sheet) and potentially rising Bitcoin prices (due to the overlap between the AI compute and Bitcoin mining industries).

Key Takeaway: Higher borrowing costs have pressured compute infrastructure stocks. But the most important fundamental trend in our opinion—rising demand for compute—remains intact. We think the latest pullback should be considered an opportunity to add exposure to the energized compute layer.

Exhibit 1: Demand for compute remains a key business driver of the energized compute layer

Source: Bloomberg, Grayscale Investments. Based on the Indxx High Performance Computing Index. Data as of October 19, 2026. Past performance not indicative of future results.

The capex outlays of certain neoclouds are multiples higher than their current revenue. Source: Bloomberg, Grayscale Investments. Indxx High Performance Computing Index top five holdings have a capex-to-sales ratio of over 230% over the past 12 months. Data as of October 19, 2026. For illustrative purposes only.

The Solana Blockchain: A Class Apart in Speed and Scale

When it comes to volume, speed, and scale, the Solana blockchain is in a category of its own.

Exhibit 1 shows average transactions per second (TPS) for a range of blockchain networks. Solana (SOL) has averaged more than 1,400 TPS year-to-date. The next largest, BNB Chain, averaged less than 200 TPS, while other blockchain networks only processed around 40 TPS.

Solana executes this huge volume of transactions at low cost and incredibly fast speeds. Transaction costs on Solana have averaged just $0.005 this year1 and are currently considered final in roughly 13 seconds.2 After next week’s Alpenglow upgrade, Solana transaction finality should fall to just 150 milliseconds.3

Solana’s high performance has helped it attract a diverse ecosystem of applications, and has made it a leader in certain categories, including stablecoin payments. Over the last three months, Solana has processed about 25% of all stablecoin transactions, more than any other blockchain.4

Key Takeaway: Solana operates at a scale much greater than other blockchain networks, all while maintaining low costs and fast transaction finality. These strengths have supported diverse application development and wide adoption, including for stablecoins.

Exhibit 1: Solana processes much higher transaction volume than other networks

Source: Artemis, Grayscale Investments. Data as of October 13, 2026.

Source: Chainspect. Data as of October 13, 2026.

Source: Anza.

Source: Allium, Grayscale Investments. Data as of October 13, 2026, for the three months ending September 2026. Data are adjusted for bots, high frequency trading, and certain other distortions.

Putting Bitcoin’s Volatility in Context

Bitcoin’s reputation for high volatility can make it seem far removed from traditional investments. But comparing its price fluctuations with those of US large-cap stocks offers an unexpected perspective. Over the past year, Bitcoin’s realized volatility fell within the distribution of S&P 500 constituents, with a meaningful number of large-cap stocks exhibiting far greater volatility (Exhibit 1). Bitcoin’s volatility has declined to 43% over the past year, from an average of 64% over the past 10 years and 81% five years ago. It is now lower than that of 19% of S&P 500 constituents.1

In fact, Bitcoin's volatility has come down meaningfully not just over the past year, but over the past decade. As the asset has matured, with deeper liquidity, broader institutional participation, and more regulated access venues, its price swings have narrowed considerably from the extremes of its early years. Its volatility today is above that of most stocks in this sample but in our view no longer sits in a separate category of risk.

That shifts the question from whether to own a volatile asset to how to size an allocation. Investors already manage volatile holdings by sizing positions to match their risk tolerance and by considering how each one behaves alongside the rest of the portfolio. The same tools apply to Bitcoin.

One year of data cannot establish Bitcoin’s risk across market cycles, but the comparison gives investors a concrete reference point. For those comfortable owning volatile equities, Bitcoin’s price fluctuations may be less daunting than assumed.

Key Takeaway: Bitcoin’s volatility has continued to trend lower2 and is now familiar territory for investors in individual stocks, particularly growth-oriented ones—and a potential investment warrants the same disciplined assessment of risk and opportunity.

Exhibit 1: Bitcoin's volatility has been well within the range of S&P 500 stocks

Source: Artemis. Data as of October 14, 2026.

Based on average volatility of the past 3, 5, and 10 years.

Three Takeaways from Token2049

The Grayscale Research Team was on the ground in Singapore last week for Token2049, the crypto industry’s largest annual conference. Here are three quick takeaways:

1. Positive crypto market sentiment. Conference participants from around the industry—including investors, market makers, startup founders, and others—seemed to agree that the bear market is over and that crypto is in the early stages of a renewed bull market.

2. A focus on the tokenization megatrend. Public blockchains have many use cases (e.g., stablecoin payments, decentralized exchanges, digital scarcity), but the buzziest topic at this year’s conference was the tokenization of capital markets. Participants were especially focused on the benefits of tokenized equity markets (Exhibit 1), which include global access, 24/7 trading, and near-instant settlement. A key highlight was the announcement of a new tokenized asset platform from crypto exchange OKX and NYSE parent ICE.1

3. Caution over AI-powered cybersecurity risks. Partly offsetting the optimism was caution over digital security due to advancements in AI. OpenAI’s announcement of breakthroughs in mathematics last week did not include anything directly related to cryptography, but some observers fear advanced AI may eventually find vulnerabilities. Public blockchains (and other systems using conventional cryptography) may need to be hardened to ward off these threats.

Key Takeaway: The annual Token2049 conference featured positive market sentiment, enthusiasm for tokenized equity markets, and caution around cyber risks.

Exhibit 1: The value of tokenized equities onchain has been taking off

Source: Bloomberg

Cheaper, Agentic AI Means Greater Demand for Compute

Demand for Artificial Intelligence is on the rise, boosted by declining costs and the technology’s increasing capability to work on its own. That means increased appetite for compute and the physical infrastructure powering AI. Anthropic’s latest Sonnet model, for example, costs up to 30% less per task, according to the company, while OpenAI’s recently introduced “dots”—always-on agents that can work toward users’ goals around the clock1—could increase the need for compute dramatically.

Lower costs can unlock more AI activity. Competition from lower-priced open-weight models is putting pressure on token prices—the fees developers pay for the text that AI models process and generate. As we discuss in Investing for the Compute Bottleneck, cheaper intelligence can make more applications and automated workflows economically viable.

Agents could amplify this demand. A chatbot answers a question; an agent works through multiple steps toward a goal. Preparing for a client meeting could involve reviewing notes, analyzing portfolio changes, and drafting an agenda. These workflows can consume 5–50x as many tokens as a chatbot interaction, depending on the task. Always-on agents could extend this processing beyond individual conversations.

Inference—the computation required to run AI models—is already growing rapidly (Exhibit 1). Goldman Sachs projects that token consumption could increase 24x by 2030, driven largely by agents. Even as AI becomes more efficient, total compute demand can rise if usage expands faster.

That digital activity still depends on physical infrastructure. Power, data centers, and GPUs take time to deliver. Cheaper AI can therefore coexist with tight compute supply: CoreWeave recently reported signing new contracts at higher compute prices. In our view, cheaper intelligence and more capable agents could expand AI use cases and usage enough to outweigh efficiency gains—supporting demand for scarce compute capacity and benefiting its owners.

Key Takeaway: Cheaper AI and more capable agents will likely accelerate compute demand, benefiting owners of the scarce physical infrastructure supporting the expansion of artificial intelligence.

Exhibit 1: From chatbots to agents, inference demand is surging

OpenAI. “Introducing dots.” September 29, 2026. https://openai.com/index/introducing-dots/

BTC’s Hidden Risk: The “Out-of-Market” Opportunity Cost

Bitcoin (BTC) returned approximately 225% over the last three years, compared with 109% for the Nasdaq.1 However, Bitcoin’s gains were not distributed evenly across the period. Its strongest trading days contributed disproportionately to the overall result, highlighting a) the difficulty of timing BTC’s exposure; b) the potential cost of being on the “wrong side” of volatility; and c) the price of foregoing potential long-term upside.

Removing BTC’s five best trading days reduces its three-year return from 225% to 95% (Exhibit 1). Without its 10 best days, the return falls to 27%, and removing the top fifteen turns the three-year gain into an 11% loss. Comparatively, Nasdaq returns were less concentrated: removing its fifteen best days reduced its cumulative return from 109% to 21%. Bitcoin’s performance history demonstrates its capacity for powerful upside bursts that can materially reshape long-term investment outcomes.

Risk is typically defined as the possibility of loss, but for an asset with Bitcoin’s return and volatility profile, remaining uninvested carries an opportunity cost of its own, in our view. Fewer than 0.5% of the days in this period accounted for enough upside to cut Bitcoin’s cumulative return by more than half when removed. Investors waiting for volatility to subside or the outlook to become clearer may find that much of the repricing has already occurred. Bitcoin’s best days cannot be reliably predicted—capturing them requires being invested when they arrive.

Key Takeaway: Given BTC’s return and volatility profile, investors looking for long-term capital appreciation can avoid trying to time the market and instead seek consistent, long-term exposure to the asset.

Exhibit 1: Impact of removing top trading days of last three years

Source: Bloomberg. Data as of September 23, 2026. BTC represented by spot BTC/USD. Nasdaq represented by Nasdaq-100 Index. Index and spot price returns are presented for illustrative purposes only and exclude management fees and expenses. Indexes are unmanaged, and investors cannot invest directly in an index. Past performance does not guarantee future results.

The AI Trade Reaches Crypto with Enthusiasm

The Artificial Intelligence Crypto Sector meaningfully outperformed its peers in September, gaining 54% compared with 24% for the broader crypto market (Exhibit 1).

Artificial intelligence could be the most consequential technology of the 21st century, but its development remains concentrated among a handful of companies. Public blockchains can complement centralized AI by providing open infrastructure for agent payments, identity, private compute, and verifiable records.

Crypto’s AI-focused market segment remains remarkably small. Even after September’s rally, AI is the smallest of the six Crypto Sectors at roughly $15 billion.1 We believe the AI sector could produce one or more major winners over the next five years. Tokens leading the sector today include:

NEAR (NEAR, +183% in September): Platform for agentic commerce, co-founded by a researcher behind the transformer architecture.

Bittensor (TAO, +37%): Open ecosystem of specialized AI networks spanning inference, agents, data, and compute.

World (WLD, +47%): Proof-of-human network co-founded by OpenAI CEO Sam Altman.

Venice (VVV, +70%): Privacy-preserving consumer AI chatbot application providing private access to AI models without storing users’ prompts or responses.

Key Takeaway: The AI crypto sector remains small, but recent outperformance suggests investors are beginning to price in the role blockchains could play as infrastructure for the emerging agent economy.

PS: To learn more about crypto AI, check out Grayscale Institute’s recently released class on the Artificial Intelligence sector.

Exhibit 1: AI crypto assets led digital asset markets in September

The Grayscale Crypto Sectors is the taxonomy framework that Grayscale uses for its asset class. It has six sectors: Currencies, Smart Contract Platforms, Financials, Consumer and Culture, Utilities and Services, and Artificial Intelligence.

Zcash Has Room to Capture Market Share

Zcash—a Bitcoin-like digital currency with additional privacy features—has appreciated sharply over the last year, rising from about $60 per token to roughly $1,500 today.1 Despite the large gain, we do not think the token has reached any kind of valuation ceiling. Rather, the magnitude of the rise can be seen as justified, as it reflects a low starting point and Zcash’s large addressable market, in our view.

Zcash (ZEC) and Bitcoin (BTC) are both members of the Currencies Crypto Sector—the blockchains focused on the digital currency use case. Bitcoin is the king of the category, but there are 10 other qualifying tokens including ZEC, XRP, LTC, and others.

On a market capitalization basis, ZEC was valued at less than 0.1% of BTC at this time last year and today is worth ~1.5% of BTC. This is a substantial change, but not extreme in the context of the Currencies Crypto Sector (Exhibit 1). For context, XRP is currently valued at nearly 6% of BTC’s market cap. And during the 2017-18 market cycle, four different cryptocurrencies—XRP, BCH, LTC, and DASH—reached at least 3% of BTC’s market cap. (For an in-depth view of the sector, check out our Grayscale Institute class on Digital Currencies.)

Demand for privacy-preserving technologies will continue to grow, in our view. If Zcash continues to offer the most effective privacy features among competing cryptocurrencies, we think it can capture additional market share.

Key Takeaway: Zcash has appreciated sharply, but history shows there is room for more competition among cryptocurrencies. We think Zcash can continue to capture market share.

Exhibit 1: Zcash has continued to capture higher market share

Source: Artemis. Data as of September 28, 2026. Past performance not indicative of future results.

Gen Z’s Longevity Advantage

When it comes to US investors, the younger, the earlier: Gen Z began investing at an average age of 19, compared with 25 for Millennials, 32 for Gen X, and 35 for Baby Boomers. Assuming a retirement age of 65, that gives Gen Z a 46-year investment horizon—more than 50% longer than the 30-year horizon for Boomers when they started investing (Exhibit 1). Starting earlier is typically considered a smart move due to the benefits of compounding, but it may also expand an investor’s capacity to take risks.

An investor’s economic wealth can be thought of as the combination of financial capital—assets already accumulated—and labor capital, or the value of future earnings. Early in life, labor traditionally represents a much larger share of one’s total wealth. As investors age, future earnings decline while accumulated financial assets generally rise. A longer investment horizon therefore provides more time to recover from periods of volatility and more future income with which to continue saving and investing. In that sense, starting earlier expands an investor’s lifetime risk budget.

Digital assets may fit particularly well within a longer-horizon, more-flexible framework because the asset class’s return profile is both volatile and potentially asymmetric.

For an investor starting early, the cost of near-term volatility can be less consequential than the opportunity cost of having too little exposure to assets with significant long-run upside potential. With decades to rebalance, add capital, and compound through multiple market cycles, investors who start investing early may be able to allocate more of their portfolio to digital assets than investors who start later—while keeping overall lifetime risk in balance.

Key Takeaway: Starting earlier gives investors more time to absorb volatility, potentially expanding the utility function digital assets can have in a long-term portfolio.

Exhibit 1: Gen Z has started investing early, expanding investment horizons