What still gets funded in Web3

Web3 did not run out of venture capital in 2025. It did, however, become a more conventional and much less forgiving market.

Galaxy Research recorded $20 billion of venture investment in private crypto and blockchain companies across 1,660 deals during 2025. That was the largest annual amount since 2022 and more than twice its 2023 figure. Yet the recovery was neither broad nor evenly distributed. Later-stage companies represented 57 per cent of recorded capital, the highest annual share in Galaxy's series, while a small number of outsized transactions accounted for much of the increase.

The pattern continued in 2026. Galaxy recorded roughly $4 billion across 355 private-company deals in the first quarter. CryptoRank, using a different universe and classification system, counted $4.99 billion across 218 venture rounds in the second quarter. The two series cannot be combined into one continuous total. Both nevertheless describe a market in which announced capital remains available but is increasingly concentrated.

The useful question is not whether Web3 transactions are still being announced. It is which businesses can convert a crypto proposition into an investable company.

First, define what the numbers measure

Funding reports often place unlike transactions beside one another. Venture equity, token warrants, private token allocations, public token sales, debt facilities, acquisitions and purchases of existing shares can all appear beneath a general heading such as crypto fundraising. They do not give a company the same resources or impose the same obligations.

For this analysis, venture funding means a database-recorded private investment into a company or project, normally through equity, a convertible instrument, a private token allocation or a combination of these. Deal value means the amount attached to a reported transaction. It is not automatically equivalent to cash received and available for expenditure on the announcement date.

A financing may close in tranches, depend on conditions or include secondary shares sold by existing holders. A debt facility can state maximum borrowing capacity that has not yet been drawn. An acquisition price principally measures consideration offered for ownership of the target; it is not operating capital delivered to the acquired business. A fund announcement represents commitments to an investment vehicle, not money already invested in portfolio companies.

This article therefore uses terms such as recorded capital, reported investment and announced deal value. Even where a source says capital was invested or deployed, public private-market disclosure rarely provides evidence of settlement, drawdown, secondary components or unrestricted cash available to the company.

Token sales are treated separately. A public buyer purchasing tokens is acquiring a different instrument, with different liquidity, governance, disclosure and regulatory characteristics from venture equity. Adding public sales to venture rounds would exaggerate the capital represented by conventional start-up financing and obscure who bears the risk.

2025 was a recovery dominated by later-stage deals

Galaxy's quarterly reports recorded $4.8 billion in the first quarter, $1.976 billion in the second, $4.65 billion in the detailed third-quarter section and $8.5 billion in the fourth. These database figures indicate that the annual result was shaped by two exceptional quarters rather than a steady rise.

The first-quarter result included MGX's reported $2 billion investment in Binance. Galaxy calculated that, without that transaction, the fall from the first to the second quarter would have been 29 per cent rather than 59 per cent. Binance was already the world's largest crypto exchange, so the transaction said more about institutional appetite for a mature trading platform than about the availability of start-up capital.

Concentration became even more pronounced in the fourth quarter. Galaxy reported that 11 transactions above $100 million represented $7.3 billion, or 85 per cent of the quarter's $8.5 billion. Its list included Revolut at $3 billion, Touareg Group at $1 billion, Kraken at $800 million, Ripple and Tempo at $500 million each, and six other nine-figure transactions. These are reported transaction values and should not be interpreted as proof that each recipient had the entire amount available as unrestricted cash at announcement.

This creates two valid but different readings of 2025. On aggregate, Galaxy's recorded private financing was more than twice its 2023 level. For a seed-stage founder, however, that did not mean twice as many investors were competing to finance experimental protocols. Much of the increase came from larger reported transactions involving businesses with scale, licences, users, revenue or a credible route into regulated financial markets.

Deal count reinforces that interpretation. The 1,660 deals recorded in 2025 remained below the busier 2021 and 2022 market. Recorded capital rose faster than the number of financed companies. In the fourth quarter, Galaxy reported a median deal size of $4 million and a median pre-money valuation of $70 million. It cautioned that valuation data covered only 10 per cent of the quarter's deals and skewed heavily towards later-stage companies. The valuation is therefore a statistic for the disclosed sample, not a benchmark for every founder.

Stage matters more than the market-wide total

Pre-seed and seed: active, but no longer the centre of gravity

Early-stage funding did not disappear. In the fourth quarter of 2025, pre-seed transactions represented 23 per cent of Galaxy's deal count. The proportion fell to 19 per cent in the first quarter of 2026. During that first quarter, Galaxy classified 57 per cent of recorded capital as later stage and 43 per cent as going to younger companies. It also reported that later-stage transactions represented one-quarter of completed deals.

What changed was the burden of proof. A general claim that blockchain will disrupt an industry is no longer enough. Fundable early-stage companies tend to identify a specific customer, a measurable cost or risk, and a reason that blockchain or digital assets are necessary to the product. Infrastructure that improves settlement, compliance, custody, security or developer operations fits this test more readily than a consumer application dependent on speculative user acquisition.

The first-quarter 2026 category counts illustrate the breadth beneath the dollar headline. Galaxy recorded 56 infrastructure deals, 39 in its combined Web3, NFT, DAO, metaverse and gaming category, 33 in payments and rewards, 25 in tokenisation, 23 in DeFi and 22 in privacy and security. These are category counts across Galaxy's full quarterly dataset. They do not establish that seed investors alone examined this pipeline or that every category received equal capital.

Series A and B: evidence replaces narrative

At Series A and B, investors increasingly expect evidence that the product works as a business. Relevant evidence can include net revenue, retained transaction volume, contracted enterprise customers, reliable unit economics, audited technology or progress towards regulatory authorisation. Gross on-chain volume is less persuasive when it is subsidised, circular or generated by incentives that end with a token campaign.

This stage remains relevant to payment infrastructure, stablecoin orchestration, institutional trading technology, security, compliance and tokenisation platforms. These businesses can sell software or financial services while using crypto rails underneath. Their customers need not adopt a new cultural identity or hold a volatile asset to receive the benefit.

Series A also remained meaningful in CryptoRank's second-quarter 2026 dataset. It reported $1.72 billion for Series C and later, but $1.2 billion came from one Series F round for Kalshi. Removing that reported round value left $520 million across six later-stage rounds, below the quarter's Series A total. The apparent stage leader was therefore heavily dependent on one transaction.

Later stage: large reported values for scarce assets

Later-stage companies represented 57 per cent of Galaxy's recorded 2025 capital and the same approximate share in the first quarter of 2026. Their attraction is practical. Established exchanges, brokers, payment networks and custody providers possess customer relationships, licences, liquidity, compliance systems and operating histories that are expensive to reproduce.

Strategic investors were also prominent in CryptoRank's second-quarter 2026 venture data. It counted 61 strategic rounds, compared with 50 seed rounds and 35 Series A rounds. CryptoRank defined this line as capital coming from exchanges, protocols and other companies rather than dedicated venture funds. Corporate participation can provide distribution or regulated access, although it can also create commercial dependencies that a conventional financial investor would not require.

Large later-stage totals still require caution. Some databases classify broad financial-technology companies as crypto because digital assets form part of their product set. Revolut's inclusion materially affected Galaxy's 2025 result. Category boundaries are analytical choices, and two databases can produce different totals without either making an arithmetic error.

What categories are still attracting capital?

Trading, exchanges and financial intermediation

Galaxy reported $5.5 billion for trading, exchange, investing and lending businesses in the fourth quarter of 2025. In the first quarter of 2026, it recorded roughly $2.6 billion, about three-fifths of quarterly capital, and 74 deals in the category.

This is the clearest destination for large reported transaction values, but not necessarily the easiest category for a new entrant. Capital is concentrating around platforms with liquidity, regulatory permissions, distribution and multiple revenue lines. A new exchange offering little beyond another matching engine faces a different market from an established venue adding custody, derivatives, payments or institutional brokerage.

Stablecoins and payments

Stablecoin infrastructure is fundable because it connects blockchain settlement with identifiable financial activity. Investable layers include issuance technology, treasury management, cross-border settlement, merchant acceptance, cards, compliance, liquidity and connections between bank accounts and tokens.

An Artemis study published in 2025 analysed more than $94.2 billion of payments settled by 31 participating stablecoin payment firms between January 2023 and February 2025. The sample is not the whole market, and the figure measures payment value rather than company revenue or venture funding.

Funding reports also show activity in payments. Galaxy counted 33 payments and rewards deals in the first quarter of 2026. CryptoRank counted 40 payment transactions and $1.48 billion in its broader second-quarter fundraising dataset. That CryptoRank category can include venture, debt and acquisitions, so its value must not be compared directly with Galaxy's venture-only category figures. The count nevertheless shows that activity extended across many transactions.

Infrastructure, security and compliance

Infrastructure continues to receive numerous rounds. The category includes developer tools, interoperability, node and data services, institutional settlement systems, key management and operational software. It is attractive when the product removes technical complexity for a paying customer rather than simply adding another layer to a fragmented stack.

Security and compliance have a similarly durable case. Custody controls, transaction monitoring, identity systems, smart-contract assurance and regulatory reporting are recurring requirements. Demand can grow as institutions enter the market and regulation becomes more specific. These products can still fail commercially, but their funding proposition is anchored in an identifiable cost rather than discretionary speculation.

Tokenisation and institutional settlement

Tokenisation attracts investment where it improves issuance, servicing, collateral mobility or settlement for an existing asset. The strongest propositions identify who owns the asset, which legal claim the token represents, how cash and records reconcile, and which regulated entity is responsible at each stage.

The category is less convincing when tokenisation is treated as an end in itself. Putting an asset record on a blockchain does not create liquidity, legal enforceability or investor demand. Fundable companies generally combine technical rails with custody, transfer restrictions, administration, distribution or regulated market access.

AI and crypto

AI-related transactions were prominent in CryptoRank's second-quarter 2026 data, but its classifications require care. It counted 41 AI transactions. It also assigned $4.71 billion across five transactions to mining, while explaining that all five funded AI compute activity. That category total included IREN's $3.65 billion debt facility and its $625 million acquisition of Mirantis, so it was not a measure of venture investment in mining start-ups.

An investable intersection between AI and crypto therefore needs more than two popular labels. Possible cases include verifiable compute, machine payments, data provenance and infrastructure serving autonomous software. The financing instrument, customer and source of revenue matter more than the category assigned by a database.

DeFi, gaming, NFTs and consumer Web3

These categories still receive funding, but some indicators weakened. Galaxy said the share represented by its combined Web3, NFT, DAO, metaverse and gaming category had been waning. CryptoRank recorded $246 million across 28 DeFi venture rounds in the second quarter of 2026, its lowest quarterly capital figure since the fourth quarter of 2023. Its recorded DeFi round count had fallen from 111 in the first quarter of 2024.

This does not mean no DeFi company can raise. It means CryptoRank recorded fewer financed companies in the category. Products with differentiated risk management, sustainable fees, institutional distribution or infrastructure value retain a case. Forked protocols, temporary yield schemes and applications whose economics depend principally on token appreciation face a narrower audience.

Consumer Web3 has a comparable problem. Engagement must survive after incentives are removed. Investors can test retention, acquisition cost and spending behaviour against mature consumer software benchmarks. Token distribution is no longer accepted as a substitute for product-market fit.

Token sales returned, but they are a separate market

CoinList reported running sales for 21 projects in 2025. It said more than $95 million worth of tokens were purchased by more than 70,000 participants across 110 countries and that 18 of the 21 sales were oversubscribed.

Those self-reported figures describe purchases through one token-sale platform, not venture equity. Buyers purchased or became entitled to tokens under sale-specific terms. Supply, vesting, governance rights, network utility and legal treatment vary by project and jurisdiction. Oversubscription measures orders relative to the offered allocation; it does not prove durable demand after trading begins.

Token sales can finance network development and broaden ownership. They can also move market and regulatory risk directly to purchasers. Editorial analysis should therefore report the sale amount, instrument, eligibility restrictions, vesting terms and distribution schedule separately from private rounds. A token's fully diluted valuation must not be added to capital raised because valuation is not cash received.

Announced money is not always available money

CryptoRank reported $12.86 billion across 271 completed transactions in the second quarter of 2026. Its components included $4.99 billion across 218 venture rounds, $4.36 billion across nine debt transactions, $3.33 billion across 40 acquisitions, $76.4 million across two public-equity transactions and $104.1 million across two private investments in public equity. These categories use different instruments and cannot be treated as one pool of start-up capital.

One $3.65 billion IREN debt facility represented 84 per cent of CryptoRank's quarterly debt total. CryptoRank also reported that the ten largest transactions accounted for 67 per cent of all disclosed capital. A founder should not read the $12.86 billion headline as money available to start-ups. Acquisition consideration generally compensates owners of the target. Debt must be repaid and may be subject to drawdown conditions. Public equity belongs to a different market. Venture announcements can also include secondary components or staged closings that are not publicly itemised.

The most defensible market reading uses several measures together:

• Recorded venture deal count indicates how many transactions entered the database, subject to its inclusion rules.
• Median disclosed round size is less vulnerable than a total to one billion-dollar transaction, although missing values still bias the sample.
• Stage distribution shows whether recorded capital is reaching younger companies or concentrating around mature ones.
• Category deal count reveals breadth that can disappear beneath large financial-services transactions.
• Instrument type separates venture financing from debt, acquisitions, public placements and token purchases.

What investors appear to require now

The 2025 and 2026 reports point to a market with announced capital but little patience for undefined utility. Companies remain plausible funding candidates when they can demonstrate several of the following:

1. A specific customer with a costly problem, rather than a general claim about decentralisation.
2. Revenue or usage that survives the removal of token incentives.
3. A defensible advantage in regulation, distribution, liquidity, data, security or technical performance.
4. A clear explanation of why blockchain settlement or digital assets improve the product.
5. An instrument and token structure that do not conceal dilution, liabilities or conflicting incentives.
6. Milestones sized to the round, with enough runway to reach the evidence required by the next stage.

For early-stage teams, the opportunity is real but narrower. Infrastructure, payments, tokenisation, security and selected DeFi products continued to record rounds. For later-stage businesses, databases recorded very large financings where licences, revenue and distribution made the asset scarce. Between those groups, companies relying mainly on category momentum were exposed.

That is what still gets funded in Web3: not the broad promise of a new internet, but businesses that can state precisely what they provide, who pays for it, which risks they control and what the announced capital is intended to achieve.