Artificial intelligence has attracted an extraordinary share of technology financing, but this reallocation of attention does not imply that cryptocurrencies have disappeared or that the two fields are economically interchangeable. Their underlying models differ. The cryptocurrency sector seeks to maintain transferable digital assets and transactions on shared ledgers, sometimes reducing dependence on specific institutional intermediaries. The dominant commercial AI sector instead relies on expensive computing facilities, proprietary models, power infrastructure, cloud services and distribution systems often controlled by a relatively small number of corporations. Both technologies may affect information, payments and organizational arrangements; they mobilize capital through very different channels.
The OECD measured worldwide venture-capital investment in AI companies at approximately $258.7 billion during 2025, or 61% of the $427.1 billion covered by its dataset. AI's share had been about 30% in 2022. Galaxy Research, using a separate dataset and different sector definitions, estimated roughly $20.3 billion of venture investment in private cryptocurrency and blockchain companies in 2025. These amounts cannot be converted into exact cross-sector shares without reconciling their methodologies, but they demonstrate a major difference in financing scale. [1][2]
The capital gap has multiple meanings. One concerns the difference between token price appreciation and the actual financing of productive businesses. Another is the gap between financial speculation and reliable economic applications. A third compares the current status of digital-asset venture financing with commitments to AI infrastructure. A fourth asks whether programmable money and distributed records could become complements to AI rather than competitors. These distinct interpretations require separate evidence instead of a single narrative about the rise or failure of cryptocurrencies.
1. Two intellectual origins: distributed verification and statistical prediction
Bitcoin was introduced in 2008 as a proposal for electronic payments that could be validated without delegating all trust to one financial intermediary. The arrangement combines cryptographic signatures, a distributed ledger, proof of work and incentives for participants to maintain transaction ordering. Its principal innovation was organizational: permitting people who do not trust one another to agree on a common record under predetermined rules. Its limitations include fees, throughput, governance disputes and the resources needed to maintain security.
Ethereum broadened blockchain development through programmable contracts and applications. This expanded the ambition from transferable money to financial infrastructure, digital property and organizational coordination. Some ventures sought to reduce reliance on banks and exchanges, while others replicated many functions of traditional intermediaries with a new technological substrate. The distinction between decentralization at the software layer and concentration at the ownership or service layer soon became central.
Artificial intelligence developed along another trajectory. Statistical learning, neural networks, research computing and natural-language processing evolved for decades before large models became widely commercial. Better chips, large datasets and scaling methods supported systems able to assist with writing, software, customer service and analysis. They were not principally designed to eliminate a trusted record keeper. Their commercial advantage derives from useful output, process integration and control of expensive infrastructure.
2. The speculative cycles of digital assets
Cryptocurrencies experienced repeated booms in price, adoption and fundraising. The first wave centered on Bitcoin as transferable digital money. Subsequent waves emphasized token issuance, decentralized finance, smart-contract platforms and digital collectibles. Investors frequently financed projects based on expectations of network effects and future utility, before many applications had mature commercial demand. This pattern combined real experimentation with intense speculation.
Initial coin offerings allowed projects to raise capital by selling tokens, sometimes without issuing conventional equity or clearly enforceable claims on future profits. Decentralized finance expanded automated lending and trading, but also exposed participants to coding failures, liquidation cascades, weak governance and dependence on price feeds. Many assets lost liquidity, and fraudulent projects damaged trust. The survival of particular protocols does not demonstrate that the financial returns generated by the whole sector were sustainable.
Prices and fundraising often moved together without representing the same economic process. A rising token price could improve the apparent wealth of holders and allow projects to finance operations from treasury assets. It could also attract leverage that reversed sharply in downturns. The crucial historical question is which projects built durable services, which depended on continuous speculative inflows and which transferred losses to later investors.
3. Generative AI and the visibility of commercial usefulness
Generative AI gained attention because many of its applications can be demonstrated immediately. A system that drafts code, analyzes a document or supports customer operations provides a recognizable service even before its long-term productivity impact is known. Large companies began testing AI within existing budgets and workflows, creating potential revenue through subscriptions, cloud usage, software bundles and enterprise contracts.
This visibility creates an advantage in capital markets. Investors can estimate potential spending on software, computing and corporate labor and imagine growth within familiar contractual structures. AI infrastructure also requires investment in tangible, securable equipment such as buildings, electrical systems and servers. Those assets allow forms of financing that are not directly comparable with purchasing digital tokens in a secondary market.
Nevertheless, adoption is not the same as economic profit. Model training, analysis, data security and human verification can be expensive. A system may accelerate one task but slow a full workflow or produce errors that require costly correction. Some applications increase productivity; others merely redistribute existing software spending. An economic judgment requires observed unit economics and customer behavior rather than extrapolation from impressive demonstrations.
4. What the investment figures actually measure
The OECD found that approximately $258.7 billion of venture capital went to AI businesses in 2025. Around 73% of this AI investment was in financing rounds greater than $100 million, and about three quarters was directed to US-headquartered companies in the dataset. AI capital formation was thus concentrated by field, enterprise size and geography. These patterns help explain why a handful of corporate ecosystems gained strong negotiating positions over infrastructure. [1]
Galaxy Research reported $5.683 billion invested through 384 cryptocurrency and blockchain deals in the second quarter of 2026, up 31% compared with the prior quarter. Its reported first-half total was approximately $10.018 billion. The report also discussed highly concentrated fundraising by specialist venture managers: approximately $3.9 billion raised across five new funds in that second quarter. This illustrates continued investment activity without establishing a return to the earlier financing environment. [2]
The surveys define their universes independently. Some firms combine blockchain and AI, while reporting conventions differ on corporate financing, private rounds, business classification and dates. A division of one total by the other would be a misleading precise market-share estimate. The broad conclusion remains substantive: AI attracted far larger conventional venture allocations in the observed period, while private cryptocurrency financing continued in narrower and often cyclical segments.
5. Market capitalization is not productive investment
The market capitalization of a cryptocurrency is generally calculated from a quoted price multiplied by a measure of circulating units. It is a valuation of outstanding assets, not a bank account containing the same amount of invested money. A marginal transaction can change the price assigned to units held by participants who did not trade. Therefore, an increase in capitalization does not establish an equal inflow of new financing to developers, businesses or miners.
Venture-capital investment is a transaction involving identifiable companies or investors. Depending on the survey, it may include new share subscriptions and selected secondary deals. Physical capital expenditure, or capital investment, represents spending on long-lived productive assets such as machines and buildings. Revenue is money paid by customers for a service or product. Profit and cash flow are yet other concepts, affected by costs, accounting choices and financing obligations.
The distinctions are essential when comparing blockchain with AI. A publicly traded AI-chip manufacturer may gain hundreds of billions in equity value without issuing new shares. A token may appreciate strongly while its associated protocol generates little revenue. A cloud provider may spend heavily on infrastructure even when near-term profits are uncertain. Investors who combine these magnitudes risk mistaking price discovery for an expansion of the productive capital stock.
6. Capital concentration within the AI economy
AI financing is concentrated not only in the sector as a whole but also among leading model developers, hyperscale cloud providers and specialized hardware firms. Frontier models require compute clusters, engineering talent, data pipelines and systems that can serve many users at acceptable latency. These demands create high fixed costs that may advantage well-financed participants, although software efficiency and open models can modify the competitive structure.
Investment concentration affects bargaining power. Firms with large financial resources can secure electricity supply, lease data-centre sites, negotiate capacity with chip suppliers and finance prolonged periods of uncertain returns. Startups lacking access to these resources often depend on established providers. The resulting ecosystem combines innovation at the application layer with possible concentration over the infrastructure on which applications run.
Scale does not prove monopoly power by itself. High investment may reflect high operating costs and genuine technological complexity. Competition between providers can reduce pricing, and open alternatives can constrain incumbents. Measuring market power requires evidence about customer switching, contractual exclusivity, margins, interoperability and barriers to entry. Large venture rounds should be treated as signals about access to finance, not definitive proof of superior future productivity.
7. The physical cost of a digital promise
Both cryptocurrency networks and AI systems have physical infrastructures, but their energy demands arise through different mechanisms. Bitcoin proof-of-work mining uses computational expenditure as part of network security. Many other blockchains use different consensus methods and energy profiles. AI training and analysis rely on chips performing statistical operations, with electricity requirements influenced by model size, efficiency, data-centre utilization and cooling.
The International Energy Agency's Energy and AI research estimated data-centre electricity consumption around 415 TWh in 2024 and published projections approaching 945 TWh in 2030 under its central scenario, with subsequent discussion giving a similar order of magnitude. These figures cover data-centre demand, not exclusively generative AI, and represent scenario-based forecasts rather than realized future electricity use. [3]
The physical scale of investment changes financial exposure. A data-centre project may involve land, power interconnection, construction financing, chip leasing and long-term customer commitments. Some assets may depreciate economically before their expected contractual life if processors improve quickly. The same infrastructure can create real services and financial vulnerability. High investment in physical equipment is therefore neither intrinsically productive nor proof of a speculative bubble; the test is future cash generation relative to total obligations.
8. Bitcoin as a monetary asset and speculative reserve
Bitcoin functions for some holders as a scarce digital asset with rules designed to limit issuance, and for others as a speculative exposure to cryptocurrency adoption and liquidity. These motivations can coexist. Scarcity is a property of its protocol's supply rule; it does not establish that purchasing power will be stable in fiat terms. Prices remain sensitive to market sentiment, leveraged demand, regulation and macroeconomic conditions.
The comparison with gold can be useful only when specific properties are distinguished. Both assets have limited supply relative to demand, but their histories, storage risks, volatility and institutional uses differ. Bitcoin transfers can be verified through a public ledger, while custody and access to traditional financial markets often depend on exchanges, funds and regulated service providers. Institutional investment may expand its investor base without making its price independent of monetary policy or risk appetite.
Bitcoin's persistence weakens the claim that AI has simply displaced the cryptocurrency sector. Yet price resilience alone is insufficient evidence of an economic ecosystem delivering broad productive innovation. Assessing Bitcoin requires separating its monetary hypothesis, network-security operation, concentration of holdings and the financial infrastructure that connects it to conventional markets.
9. Stablecoins and the strongest payment-use case
Stablecoins seek to maintain a relatively stable value against a reference asset, commonly the US dollar. Their usefulness derives from transferring claims or tokens through programmable networks, sometimes with rapid settlement and access beyond traditional banking hours. They can support remittances, trading, international transactions and selected payment services. Their economic significance therefore need not depend on generalized appreciation in token prices.
Yet stablecoins differ substantially in the rights holders possess. Reserve composition, redemption rules, custodian arrangements and legal jurisdiction shape the reliability of each product. A token labeled as stable may depend on banks, short-term sovereign securities or centralized governance. On-chain transferability does not eliminate credit, liquidity or operational risk at the issuer. The Bank for International Settlements has emphasized the distinction between innovation in payment technologies and the monetary arrangements necessary for stability. [5][9]
Issuers may receive income from assets backing tokens, while payment intermediaries earn fees for conversion or settlement. This revenue does not automatically accrue to every participant holding a token. Stablecoin adoption can therefore grow while many speculative cryptoassets continue to struggle. This separation is central to understanding why the sector's genuine applications may survive a decline in the popularity of broad Web3 investment narratives.
10. Tokenization and the institutions beneath digital claims
Tokenization represents economic claims or assets on programmable infrastructures. Possible applications include securities settlement, trade finance, collateral movement and automated contractual processes. The expected benefits include reduced reconciliation work, faster transfer and improved access to certain markets. But technical representation of a claim does not establish its legal enforceability, liquidity or value.
If a financial security is tokenized, regulators and market participants must still determine ownership rights, settlement finality, insolvency treatment and authorized intermediaries. The asset can become easier to transfer digitally while remaining dependent on a central issuer, registrar or custodian. This architecture may be efficient without being decentralized in the strong political sense associated with early cryptocurrencies.
The IMF has stressed that tokenization requires suitable institutions and legal coordination to achieve meaningful economic results. Technical feasibility alone cannot resolve fragmented regulation, data standards, cyber risk and the incentives of existing intermediaries. [6] The relevant benchmark is the cost, security and access of the complete system, not the fact that a ledger uses blockchain terminology.
11. The monetization problem of protocols
Many cryptocurrency protocols are open infrastructure rather than conventional businesses. A network may process transactions and create social value without distributing enforceable profits to token holders. Protocol fees, issuance, staking rewards and governance rights are not interchangeable with corporate dividends. The value that a network generates can accrue to users, validators, custodians, developers or competing application providers.
Token investors often presume that increasing usage will mechanically raise token prices. That analysis depends on specific token economics, supply rules, fee obligations and investor expectations. In some designs, services grow without requiring users to hold substantial token balances. In others, fees are shared with validators rather than passive holders. Regulatory restrictions may affect the ability to promise future returns.
These questions help explain why the cryptocurrency investment cycle cannot be measured through adoption statistics alone. A protocol may become technically important while its token is an unsuitable investment claim. Conversely, a token may appreciate through speculation without material use of the underlying infrastructure. Financial sustainability requires evidence of durable economic demand and clear mechanisms connecting that demand to revenue or asset value.
12. Switching costs and platform dependence
Digital businesses face different forms of dependency. A cryptocurrency user can sometimes transfer assets between compatible wallets, but changing exchanges, blockchains or applications may expose them to fees, security risks and fragmented liquidity. Developers may depend on specialized tools, user communities and infrastructure providers. Decentralized governance can coexist with practical dependence on concentrated services.
AI markets show another version of the same problem. Applications may rely on specific model interfaces, cloud features, retrieval systems and pricing schedules. Migrating to another provider requires testing output quality, latency, privacy protections and code compatibility. Firms able to control essential distribution or computing resources gain bargaining power over businesses built on top of them.
Neither problem is unique to digital technology, but network effects can intensify dependence. A platform is more valuable when users, developers and services cluster around it, making departure individually costly. Economic competition depends on portability, interoperability and the availability of credible alternatives. These mechanisms are relevant to both sectors even when their political language is sharply different.
13. Venture capital and investor exits
Venture capital finances risky businesses in exchange for claims on future value, typically seeking returns from a sale, acquisition or public listing. Investor funds have finite horizons and constraints on deploying new capital. When interest rates, liquidity conditions or exit opportunities change, the amount committed to emerging sectors may change without a corresponding collapse in their technical usefulness.
Crypto investment cycles have sometimes depended on token issuance and secondary trading as additional routes to liquidity. These channels can accelerate fundraising, but they can also allow speculative valuation before a business demonstrates sustainable revenue. AI companies increasingly use conventional equity rounds, strategic investments and commercial partnerships. Some also rely on extensive debt, leases or commitments negotiated across affiliated firms.
A meaningful comparison must therefore trace financing through its contractual destination. Money raised for a venture fund is not necessarily already invested in startups. A company's announced strategic partnership is not identical to cash paid. A financing agreement may create future payment duties, contingent obligations or credits for cloud services. These structures affect which risks ultimately fall on investors, lenders, counterparties or customers.
14. Growth, valuation and the timing of returns
Asset prices reflect expected future cash flows and discount rates, among other factors. When an industry is perceived to have extraordinary growth potential, investors may pay high current valuations for earnings expected much later. That process is not necessarily irrational, but uncertainty grows when market prices imply sustained expansion, dominant margins or low competitive pressure.
AI currently offers evidence of real commercial activity, including sales of computing hardware, cloud services and software products. It also presents uncertain future profitability for particular model developers and infrastructure ventures. The fact that suppliers receive revenue does not demonstrate that each downstream customer obtains sufficient return on investment. Complex financing arrangements can create obligations before end-user cash flow is firmly established.
Cryptocurrency valuations have a different challenge when token holders lack direct cash-flow rights. Their value may rely on monetary adoption, use in protocol operations or speculative willingness to pay. These mechanisms require different models, and ordinary price-earnings ratios may be inapplicable. A rigorous comparison examines the nature of each claim instead of assuming all technological assets should be priced by the same method.
15. Money, governance and the enduring power of states
Cryptocurrencies emerged partly from criticism of centralized monetary authority and financial intermediation. Yet sovereign currencies, tax obligations, legal enforcement and public payment systems continue defining a large proportion of economic activity. The state can regulate intermediaries, establish consumer rights and determine the legal treatment of financial claims. It can also create public digital infrastructures that compete with private payment networks.
Stablecoins linked to sovereign currencies may extend the reach of those currencies rather than replace them. Dollar-denominated tokens can increase the usefulness of dollars for cross-border exchange while creating regulatory concerns regarding reserves, run risk and consumer protection. Tokenization may improve settlement but still depend on public law and institutions governing title and finality.
The resulting economy is neither pure decentralization nor uninterrupted state monopoly. Market-based digital payment services can innovate within frameworks of monetary sovereignty. The Financial Stability Board's review of cryptoasset regulation emphasizes continuing implementation differences across jurisdictions, illustrating why technology alone cannot harmonize rules. [7] The practical question is how private innovation interacts with public responsibility for stable payments and financial integrity.
16. Geopolitics of digital infrastructure
AI compute and blockchain infrastructure depend on semiconductors, network connections, electricity, skilled labor and legally enforceable contracts. These resources are unevenly distributed. Advanced chip manufacturing and access to large data-centre clusters have geopolitical significance. Export restrictions and investment screening can alter the capacity of firms or countries to develop their own systems.
Cryptocurrency protocols can distribute record validation across jurisdictions, but user access often passes through exchanges, stablecoin issuers and payment processors regulated in particular states. Mining can move in response to energy costs and restrictions, while transaction verification relies on network incentives. Physical geography thus reappears even in technologies promoted as borderless.
For emerging economies, the policy trade-off involves infrastructure autonomy and affordable access. Building independent computing capacity may be expensive, while dependence on foreign providers creates strategic exposure. Licensing open software can reduce some dependency, but power, connectivity and security still require local investment. Digital sovereignty should therefore be measured through capabilities and contracts, not slogans about national tokens or universal AI independence.
17. Latin America, inflation and everyday crypto uses
Economic instability, inflation and cross-border payment barriers have encouraged interest in digital assets in parts of Latin America. Some households seek access to dollar-denominated claims or alternatives to costly transfers. Businesses experiment with international settlement. The region is diverse, however, and adoption indicators can differ dramatically between countries depending on banking access, capital restrictions and inflation experiences.
The economic utility of stablecoins is not identical to the investment thesis for volatile tokens. A resident seeking temporary dollar exposure faces risks related to issuer reserves, platform access and exchange rates. A worker receiving a remittance may prioritize total transaction cost and withdrawal availability rather than decentralization. Legal protections and safe custody can be more important than the underlying ledger's theoretical throughput.
Chainalysis has documented substantial regional cryptocurrency activity, while warning through its methodological choices that on-chain value estimates are not equivalent to national accounts of productive investment. [4] Policy should evaluate concrete payment outcomes, consumer losses, fraud and the functioning of local financial systems. The region provides evidence of demand for alternatives, not proof that every crypto project generates lasting economic value.
18. Fraud, custody and residual trust
Distributed verification can make selected transaction histories auditable, but users still need to protect keys, interact with software and sometimes trust custodians. Fraudulent investment offerings, compromised wallets, phishing and poorly designed smart contracts have caused substantial losses. Code transparency is not sufficient protection when participants cannot assess vulnerabilities or when governance can change the rules.
Centralized exchanges may concentrate customer assets and create counterparty exposures unlike those of a self-custodied wallet. Self-custody, conversely, places technical and security burdens on users. A transaction irreversible at the ledger level can make recovery from fraud more difficult. Intermediaries that promise convenience may reintroduce precisely the concentration originally criticized by cryptocurrency advocates.
AI also introduces security and governance challenges, including unreliable generated output, confidential-data leakage and dependence on providers. These risks differ from wallet theft but similarly demonstrate that technical systems do not abolish institutional responsibility. Trust is redistributed among software developers, hardware providers, custodians, auditors, regulators and users; it is not simply eliminated by innovation.
19. Potential convergence between AI and programmable networks
AI agents may need to purchase services, transfer payments, verify provenance or execute rules across organizations. Programmable payments and shared settlement infrastructures could become useful for such applications. The possible benefit arises when integration reduces administrative work or enables transactions that were previously impractical. Merely combining the labels AI and blockchain does not establish a commercial market.
Technical requirements include reliable identity, spending limits, authorization, dispute resolution and verification of external information. Smart contracts execute programmed rules but do not inherently know whether a physical shipment arrived or a service met quality standards. Data oracles and legal recourse remain necessary where economic claims depend on the external world.
Existing banking rails, card networks and conventional APIs may solve many of these problems efficiently. Blockchain-based arrangements must demonstrate benefits against those alternatives. A successful convergence could increase use of programmable records without reviving every speculative token issued during earlier cycles. Its economic consequence would depend on who operates the infrastructure and who receives the associated fees.
20. Open infrastructure and the recreation of intermediaries
An open protocol may reduce barriers to permissionless participation, yet commercially successful applications often introduce curated interfaces, hosted wallets and professional custody. Those firms provide support, compliance and risk management, which users may value. They also acquire power to set fees, restrict access and collect information. Decentralized infrastructure can therefore support centralized businesses.
AI displays a similar split between open and proprietary layers. Open model weights can lower some entry barriers, but analysis at scale still requires hardware, software operations and customer distribution. Closed providers may offer stronger support or security assurance in particular settings. The relevant policy choices concern interoperability, market power and safeguards rather than imposing a universal preference for one licensing model.
For both sectors, industrial organization matters more than the branding of the software architecture. Users may gain from coordinated services while losing bargaining power when switching becomes expensive. Evaluating open infrastructure requires examining effective access, rights and competition across the entire service chain.
21. Employment, expertise and organizational change
Cryptocurrency development supported specialized skills in cryptography, distributed systems, security auditing and financial software. AI expansion increases demand for data infrastructure, machine learning, computing hardware and application engineering. Their labor markets overlap in software and cybersecurity but differ in capital intensity and organizational requirements. A shift in financing can change wages and employment opportunities without erasing accumulated technical knowledge.
AI may automate portions of existing work, create new tasks and increase the need for evaluation and supervision. Productivity effects depend on how firms restructure processes. Cryptocurrency applications may lower the cost of specific financial operations while expanding responsibilities in compliance and risk management. Neither technology automatically creates employment proportional to its valuation.
Distribution also depends on geography. High-paying research positions and specialized capital often cluster in a few regions, while electricity infrastructure, mining, customer support and application work are more dispersed. Public investment in education and connectivity can widen participation, but claims about job creation require observed occupational data and local context.
22. Credit, liquidity and monetary policy
Technology financing takes place within the larger credit system. High policy rates can increase borrowing costs and reduce the present value of distant expected profits. More abundant liquidity can support risky assets, fundraising and corporate investment. These effects differ according to cash-flow maturity, balance-sheet strength and investor composition.
Cryptocurrency prices have displayed sensitivity to global risk appetite, though their relationships with other assets change across periods. AI infrastructure financing involves bank lending, corporate bonds, private credit and leasing alongside equity capital. These commitments can become vulnerable if forecast utilization or customer payments disappoint. A project backed by physical equipment may still be exposed to rapid obsolescence and concentrated counterparties.
The IMF's October 2026 financial-stability work examines broader vulnerabilities associated with technology investment and financial interconnectedness. [8] It is important not to equate the existence of financing risks with evidence that a systemic crisis is inevitable. The diagnostic variables are repayment schedules, collateral values, independent customer revenue and the capacity of investors to absorb losses.
23. Capital displacement and alternative explanations
The proposition that money left cryptocurrencies in order to fund AI is stronger than the evidence generally supports. Capital markets do not contain a fixed pool that must move dollar for dollar between sectors. Investment can expand through savings, credit, securities issuance and changing valuations. An AI company raising new equity does not prove that an equal amount was previously held in cryptoassets.
A narrower hypothesis is more plausible: venture managers, founders and institutional investors have reallocated attention and future commitments in response to perceived AI opportunities. Evidence could include sector-specific fundraising, the mandates of venture funds, changes in deal counts and investor interviews. But macroeconomic interest rates, regulation, market scandals and differences in business-model maturity may explain changes simultaneously.
Other explanations concern visibility and measurement. AI rounds are frequently very large, while cryptocurrency innovation may shift from equity-backed startups toward protocols or mature issuers. Some investors may hold both fields. The comparison should test actual financing records rather than treating competing headlines as a complete account of capital flows.
24. Three possible paths through the end of the decade
In an infrastructure-dominant path, AI continues attracting substantial capital, enterprise revenue grows and demand for data centres and computing equipment remains robust. Cryptocurrency applications persist in payments, asset transfers and specialized financial infrastructure, but much of the speculative token universe fails to develop durable use. Market structure becomes more concentrated around major computing and financial intermediaries.
In a convergence path, programmable settlement integrates into software used by AI agents, international firms and financial markets. Adoption is driven by operating efficiency and enforceable claims rather than merely token appreciation. Successful systems may be public, private or hybrid, and their benefits will depend on governance and competition.
In a correction path, expected AI returns disappoint while debt and replacement costs rise. Financial losses and consolidation affect parts of the sector, but technology adoption continues in economically useful activities. A correction in AI valuations would not necessarily cause a corresponding cryptocurrency boom. These scenarios are conditional mechanisms, not precise forecasts.
25. A measurement framework for distinct economic processes
A coherent analytical framework separates five groups of indicators. First are venture-capital flows: deal counts, sizes, company locations and investor concentration. Second are capital-expenditure measures: actual spending on servers, power facilities and software development. Third are market valuations and prices, which reflect expected returns and liquidity but not necessarily new financing. Fourth are service indicators: paid customers, payment volumes, quality and unit economics. Fifth are systemic risks: leverage, concentration, custody losses and unresolved legal claims.
Within cryptocurrencies, transactions should be separated by stablecoin transfers, exchange activity, decentralized applications and real-economy payments where measurement permits. Within AI, training costs, analysis expenses, application usage and customer profitability should be separated. Different surveys require transparent denominators and sample definitions.
Comparisons over time are also necessary. A single financing year does not establish a permanent industrial regime. A technology may reach commercial usefulness after a speculative downturn, while another can attract intense investment before productivity becomes measurable. The important change is the distribution of durable capabilities, contractual claims and cash flows.
26. From the promise of disintermediation to the infrastructure economy
The cryptocurrency movement sought to make verification and transfer less dependent on centralized gatekeepers. The present AI investment cycle directs extraordinary capital toward physical computing and software platforms that can increase concentration at critical points in the digital economy. Yet the opposition should not be exaggerated. Stablecoins, tokenization and selected distributed applications may continue evolving even as AI dominates venture financing.
The unresolved economic issue is who owns the infrastructure and who receives payment from its use. Protocol openness does not guarantee decentralized control over custody or market access. Large compute investment does not guarantee lasting monopoly rents or productivity gains. Governments, firms and households will experience these developments through fees, finance, labor conditions and access to services.
The most consequential outcome may be neither the victory of cryptocurrencies nor the displacement of all conventional finance by AI, but a hybrid digital economy. Its architecture will depend on legal rights, security, energy infrastructure, competitive alternatives and the conversion of promised technological improvements into verifiable economic benefits.
Comparative evidence
| Indicator | Period | Magnitude | Universe and qualification |
|---|---|---|---|
| Global venture investment in AI | 2025 | US$ 258.7 billion | OECD AI-company dataset |
| AI share of global venture capital | 2025 | 61% | OECD denominator: US$ 427.1 billion |
| Private cryptocurrency and blockchain VC | 2025 | about US$ 20.3 billion | Galaxy, separate methodology |
| Cryptocurrency and blockchain venture investment | Q2 2026 | US$ 5.683 billion | Galaxy, 384 deals |
| Cryptocurrency and blockchain venture investment | H1 2026 | US$ 10.018 billion | Galaxy, six-month total |
View data
| Indicator / period | Value (% of investment) |
|---|---|
| AI | 61 |
| Other firms | 39 |
- Visible progress in AI applications
- revenue and productivity expectations
- very large venture rounds
- demand for chips, cloud services and electricity
- concentrated suppliers and financial obligations
- Stablecoin adoption and programmable records
- wider payment and settlement use
- revenue for issuers and intermediaries
- economic effects that do not automatically increase the price of all tokens
- Models and applications
- Accelerators and infrastructure
- Data and distribution
- Bitcoin and speculative reserves
- Stablecoins and payments
- Tokenization and settlement
- Protocols and governance
- Security and trust
- Regulation and accountability
View data
| Indicator / period | Value (% of worldwide venture investment) |
|---|---|
| 2022 | 30 |
| 2025 | 61 |
Financing concentration is not the same as worldwide production
Adoption varies with inflation, bank access and regulation
Institutional models and rights vary between jurisdictions
Technical supply chains and capital flows differ
References
[1] OECD, Venture capital investments in artificial intelligence through 2025, February 2026: https://www.oecd.org/en/publications/venture-capital-investments-in-artificial-intelligence-through-2025_a13752f5-en/full-report.html
[2] Galaxy Research, Crypto and Blockchain Venture Capital – Q2 2026, September 2026: https://www.galaxy.com/insights/research/crypto-blockchain-venture-capital-q2-2026-deal-count-fundraising
[3] IEA, Energy and AI and subsequent updates: https://www.iea.org/reports/energy-and-ai/energy-demand-from-aI ; https://www.iea.org/reports/key-questions-on-energy-and-ai/executive-summary
[4] Chainalysis, 2025 Geography of Cryptocurrency Report: https://www.chainalysis.com/reports/2025-geography-of-cryptocurrency-report/
[5] BIS, Annual Economic Report 2025, monetary-system chapter: https://www.bis.org/publ/arpdf/ar2025e.pdf
[6] IMF, What Is Needed for Tokenization to Deliver, 8 October 2026: https://www.imf.org/en/blogs/articles/2026/10/08/blog-what-is-needed-for-tokenization-to-deliver
[7] FSB, Thematic Review on FSB Global Regulatory Framework for Crypto-asset Activities, 16 October 2025: https://www.fsb.org/2025/10/thematic-review-on-fsb-global-regulatory-framework-for-crypto-asset-activities/
[8] IMF, Global Financial Stability Report, October 2026: https://www.imf.org/en/Publications/GFSR/Issues/2026/10/13/global-financial-stability-report-october-2026
[9] BIS, Cryptocurrencies and decentralised finance: functions and financial stability implications, April 2025: https://www.bis.org/publications/paper-156-cryptocurrencies-and-decentralised-finance-functions-and-financial-stability-implications