Crypto venture capital is entering a slower and far more selective phase as speculative momentum fades, according to a long-form article published by TechFlowPost. The piece argues that the sector has not disappeared. What has changed is the basis on which capital is willing to underwrite new projects.
Written by Saurabh Deshpande, Joel John and Siddharth, and translated by Saoirse for Foresight News, the article places today’s crypto funding market inside a longer arc of technology booms and busts. Its central claim is straightforward: token-led liquidity once masked weak business quality, but that model is breaking down, and capital is starting to move toward projects with revenue, legal clarity and longer exit horizons.
Strong industry metrics have not translated into easy early-stage financing
The article opens with a contrast. Institutions now hold more than $175 billion in crypto assets through exchange-traded products. Onchain projects generated $11 billion in fees over the past 12 months. The GENIUS Act is nearing the end of what the authors describe as a decade of regulatory suspension. Exits are also running hot, with $8.6 billion in M&A and 11 IPOs.
Even so, founders in the market today still struggle to raise capital. Citing Galaxy Research, the article says only eight new VC funds were launched last quarter, the lowest number since 2020. Quarterly investment fell to $4 billion, which annualizes to around $16 billion, about half the $31 billion invested in 2021.
The authors say this points to a structural mismatch. Institutions increasingly accept crypto as an asset class, but the risk capital needed to incubate new sectors and new assets has not returned at the same pace. Capital that can buy mature assets is available. Capital willing to take early execution risk is far harder to find.
From the dot-com crash to crypto excess, bubbles often leave overbuilt markets behind
To frame the current moment, the article reaches back to the aftermath of the internet bubble. Between 2000 and 2002, global markets erased $10 trillion in value. It references a 2002 New York Times column that argued quality hardware companies such as Herman Miller could become attractive once venture money returned.
Crypto, the authors argue, does not offer a neat parallel. Their example comes from the metaverse cycle. A virtual plot next to Snoop Dogg in The Sandbox sold for $450,000 at the 2021 peak. It is now listed for just over $1,000, a 99.8% drop. In other words, bubbles do not always leave behind discounted quality. Sometimes they leave behind artifacts of narratives that no longer have economic support.
The piece says the retreat of metaverse, NFT and WAGMI narratives has forced the industry into a more uncomfortable discussion. Crypto now looks increasingly like fintech. In many cases, it is just backend infrastructure. Tokens need to generate revenue, and that revenue needs to flow back into tokens. The industry also has to confront a harsher question: was much of this a bubble from the start?
Technology investing used to ration liquidity; crypto pulled exits forward
The authors compare today’s market with earlier periods in technology investing. The median time for a technology company to go public is now 14 years, they write, while the dot-com period compressed that cycle to five years. Apple went public in 1980 at an $1.8 billion valuation. Meta, then Facebook, went public after just over eight years in operation, with 901 million monthly active users, $3.7 billion in annual revenue the prior year and a $104 billion valuation.
Earlier private technology markets operated under a different capital structure. Venture firms absorbed the risk of founder selection, category creation and company building, then exited through IPOs. The article notes that in 1980, Sequoia Capital had to sell its entire Apple position to deliver a 40x return to LPs, cashing out just $6 million. Long holding periods were not the norm. Technology investing itself was treated as a high-risk activity.
The piece then strings together a series of historical warnings about capital excess. In 1987, the New York Times floated the idea that pizza shops might divert money away from tech. In 1991, Time magazine published “My so-called Silicon Valley.” In 2002, NBC said the category had too much capital. By 2013, venture had become a lifestyle identity, and “zombie VCs” had started to appear in force.
ICO merged fundraising and listing, and distorted price discovery in the process
The article argues that crypto venture investing formed at the tail end of this broader technology cycle. Between 2013 and 2016, traditional venture structures were barely present in crypto. Retail users mainly participated through exchange accounts, investing directly in networks. Pantera’s 2013 fund bought Bitcoin at an average entry price of $65.
At that stage, launching a token was close to launching a blockchain. Teams had to fork code, operate infrastructure and persuade exchanges to list the asset. ICOs collapsed fundraising and market access into a single step. Developers wrote the contracts, paid for audits and negotiated exchange listings. If exchanges refused to list the token, liquidity disappeared and many projects effectively died.
The article acknowledges that direct retail fundraising worked in some early cases. Ethereum raised $18 million in 2014 through that route and built infrastructure that still underpins the EVM ecosystem. But the model degraded over time. Retail buyers ended up with concepts and tokens, while the products and infrastructure needed to support token value remained thin.
The funding gap was quickly replaced by excess. In June 2017, ICO fundraising first overtook venture investment in crypto. By July, ICO fundraising was four times larger. By December, the market was openly saying ICOs would eliminate venture capital. The results did not support that story. More than half of ICO projects died within 120 days of token issuance, the article says, and academic research estimated that close to 80% were outright scams. The full ICO cycle raised $28 billion, but never developed a workable pricing framework for pre-revenue projects or a regulatory structure to constrain abuse.
SAFE notes with token warrants shifted incentives toward speed and listing
Crypto VC emerged inside that environment. Instead of going directly to the public, teams raised from a curated group of capital providers and promised future tokens. SAFE agreements with token warrants became a common structure.
At first, that solved a real problem on both sides. Founders got time to think through what should be tokenized. Investors gained a path to public-market liquidity for projects that would struggle to command traditional private-market valuations.
But the incentive structure changed quickly. If a seed fund bought into a project and the token went up 4x after listing, it could recover the full cost basis by selling just 25% of the position as it unlocked. Everything left over was profit. Founders, in turn, only had to prove that a token could be listed quickly and unlocked carefully enough to avoid concentrated selling pressure.
That pushed the entire system toward liquidity management rather than long-term business value. Portfolios started to optimize for listings, not durable product quality. Investors realized they did not need to wait for a project to truly work before exiting. Founders realized that was the behavior they needed to satisfy. Tokens stopped being tools to fund development and became tools to monetize a cap table.
Why so many token models failed: weak businesses and no legal claim on cash flow
The article says most token failures trace back to one or both of two problems. First, the business model was broken or never existed. Second, unlike equity, tokens generally do not carry legal claims on a company’s operating results.
The authors do not dismiss tokens entirely. Tokens can coordinate global stakeholders and help a network cold-start by subsidizing new behavior. In that sense, they compare them to venture-funded subsidies elsewhere in tech: Uber funding rides, delivery platforms funding shipping, all with the aim of creating a habit that can later be monetized.
The issue is conversion. Many crypto projects never turned user behavior into revenue. X-to-earn products lost users when token rewards disappeared because the product itself did not satisfy durable demand. Many DePIN projects repeated the same pattern, subsidizing supply while waiting for real demand that never arrived.
Citing Delphi’s State of Token report, the article says even a successful cold start is not enough. For token investing to work, token holders need some kind of claim on business outcomes. Without that, a discount to equity is not a market mistake. It may be entirely rational.
Friend.tech is one example used in the piece. The protocol generated tens of millions of dollars in fees, yet token holders were entitled to none of it because they had no legal claim on the income stream. That, the authors argue, explains why tokens often trade at large discounts to equity.
In 2026, the winners are finding ways to tie revenue back to the token
The article says the market has already begun to react. Some projects have moved back toward equity when token markets badly mispriced business value. Across Protocol is cited as an example. At the same time, the leading projects that broke out in 2026 all tied business revenue to the token in one form or another.
The piece uses a longer historical analogy to show how the premium has shifted over time. In the 1990s, “on the internet” was enough to command a premium. In the late 2000s, “mobile” worked. In the late 2010s, “onchain” did the same. Once the tide recedes, the question is no longer whether capital was abundant, but which businesses created real value with it.
Peter Pan, research partner at 1kx, is quoted directly: “Applications didn’t fail because they used token incentives. They failed because they weren’t good businesses. Crypto’s original sin is the lack of enough sustainable innovation to produce profitable projects or protocols with genuine product-market fit.”
That sets up the next section of the article: if cheap liquidity is over, what actually remains?
Three sectors have already reached sustainable product-market fit
The authors argue that three sectors have already crossed that threshold: stablecoins, prediction markets and perpetual exchanges.
Stablecoins
Stablecoin supply has surpassed $300 billion. Annual transfer volume has reached $46 trillion, and after excluding bot activity, real transaction volume is estimated at roughly $9 trillion. Stablecoin issuers combined have become the 17th-largest holders of U.S. Treasuries. Circle has gone public, and its stock jumped 167% on the first day of trading. Bank consortia have started issuing their own stablecoins as well.
The article says stablecoin issuance is now more naturally owned by growth equity, strategic acquirers and bank strategy teams than by classic seed investors. The venture firms that captured the largest upside were the ones that invested when stablecoins were still a concept. The current seed window has shifted upward into applications built on top of stablecoins. Ethena was seeded during that phase and became one of the major winners. Bridge was acquired by Stripe less than three years after it was founded.
Prediction markets
ICE has committed up to $2 billion to Polymarket. Robinhood has turned event contracts into its 11th business line to generate over $100 million in revenue. Susquehanna is building prediction-market operations as well. When the world’s largest exchange operator, a major retail brokerage and a large quantitative firm all enter within 12 months, the article says, the category has effectively been repriced.
Perpetual exchanges
Hyperliquid accounts for 44% of onchain perpetual volume. The team has just 11 people and generated more profit last year than most publicly listed exchanges, according to the article. Coinbase’s $2.9 billion acquisition of Deribit also established a clear valuation benchmark for the category.
All three sectors now generate billions of dollars in annual revenue. Four to six years ago, seed investments in them would have looked implausible. Polymarket was treated as a toy in 2020. Circle was overlooked for years as a payments company. When Hyperliquid launched, many assumed perpetuals were already a solved market.
This bear market looks different because revenue is visible
The article contrasts this cycle with the downturns of 2018 and 2022. In 2018, the hope was infrastructure: faster chains and cheaper block space. In 2022, it was institutional adoption. The authors say both narratives played out. This cycle differs because investors no longer need to rely as heavily on abstract future promises.
Blockchains, they argue, now support things traditional systems struggle to match: 24/7 dollar settlement without intermediary banks, and customer-verifiable assets inside custodial exchanges. Markets for many assets can gather global liquidity in one place.
That shift changes how revenue is earned. The leading applications cut token incentives from $2.8 billion to less than $10 million, while fee income kept rising. The article treats the arrival of traditional financial institutions as the clearest validation of category strength. Robinhood launched event contracts, tokenized equities and even its own chain because those businesses can become new nine-figure revenue lines for shareholders.
Still, perception lags reality. The article says half of the tokens listed on major exchanges last year fell more than 80%. Token prices and the business performance of the underlying entities have split apart.
Regulation is changing valuation, founder location and operating costs
A major part of the 2026 investment backdrop, in the authors’ view, is regulatory change. The GENIUS Act was signed into law in July 2025. The CLARITY Act passed the House with bipartisan support. The FDIC is drafting rules for banks that issue stablecoins.
The article says regulatory clarity affects crypto startups in three ways.
First, it lowers the risk premium embedded in deals. For a decade, U.S. crypto investments had to factor in the possibility that the government could shut down a company outright. The near-shutdown of several exchanges in 2023 remains an example in the text. Once policy turns more favorable, that non-diversifiable risk premium starts to disappear.
Second, it changes founder behavior and geography. Electric Capital data cited in the piece says the share of U.S.-based crypto developers has nearly halved over the last decade and is now below 20%. Building crypto companies in the U.S. often meant operating in a legal gray zone. A clearer framework reverses some of the adverse selection created during the ICO era.
Third, it changes costs. Compliance is described as the last durable moat in fintech. If a company can reduce the onboarding cost per compliant customer from $400 to $40, that advantage compounds every quarter. Compliant crypto infrastructure can compress that cost base materially.
The regulatory shift is not limited to the U.S. The article also lists several other jurisdictions: the U.K. is building a framework, with Coinbase and Robinhood already developing products for that market; Hong Kong’s stablecoin ordinance took effect on Aug. 1, 2025; South Korea introduced a Digital Asset Basic Act in the summer of the same year; and Japan continues to regulate stablecoins through its Payment Services Act while updating the rules over time.
Onchain revenue is diversifying, and tokens are no longer the only answer
The article then turns to revenue composition. Onchain protocols generated $11 billion in fees over the past year, and at least a dozen projects reached nine-figure annualized revenue.
Speculation still matters, but not all speculative revenue looks the same. Axiom sits at one end of the spectrum. Its income is tightly linked to meme coin trading. It generated $100 million in fees in four months and $700 million cumulatively, according to the article. But after the meme coin market broke down, quarterly revenue fell 86%.
At the other end are businesses such as Hyperliquid and Aave, which rely on trading and lending activity with stronger endurance across market cycles. The article says Hyperliquid, with an 11-person team, generated roughly $843 million in revenue last year. Its HIP-3 market is highlighted as an example of broader diversification: open interest totals $10 billion, with $4 billion tied to non-crypto assets, and daily non-crypto trading volume is about $3 billion.
Consumer software built on trading rails is another source of cash flow. Phantom wallet reached 17 million monthly active users in 2025 and generated around $325 million in revenue from swap fees. The authors say that looks more like an app-store model than an exchange. Phantom has not issued a token, and the article suggests it probably does not need one.
Institutional B2B businesses are also contributing. Tokenized asset providers and compliance infrastructure firms expanded the RWA market from $5.5 billion to $18.6 billion in a year and earned revenue through contracts rather than crypto price appreciation.
The business model furthest from pure speculation is stablecoin issuance. Tether generated more than $1 billion in profit in 2025 from its U.S. Treasury reserves alone, the article says. Circle turned a reserve-backed model into a public company. Where only one Circle emerged over the previous decade, the last two years have produced a much larger field of competitors.
The article also points to an important shift in token design: protocols paid token holders $96 million in one month through buybacks, a sign that economic rights are starting to flow back into token structures.
Valuation remains out of sync with where fees are actually generated
Even with healthier revenue streams, the sector’s valuation stack is still skewed. By the end of 2025, DeFi and financial applications produced 73% of onchain fees, while base-layer chains generated just 12%, according to the article. Yet chains still represented 91% of total protocol market capitalization, while applications made up only 6%. Chains traded at nearly 4,000x annual fees, applications at 17x. The authors say that mismatch continued into 2026.
They add that Uniswap, shturl.c and Polymarket each generated more monthly fees than Ethereum and Solana. In their view, the market is still pricing assets with the infrastructure logic of the previous cycle even though applications are the businesses producing the cash flow.
That is why the authors say they are concentrating on a handful of areas: blockchain-native financial applications, tokenized collateral and onchain credit. Their stated preference is depth over broad thematic coverage.
The article cites a Delphi portfolio made up of the top 10 cash-flow tokens, weighted by revenue. From January 2025 to May 2026, the basket returned 30.6%, while Bitcoin fell 17%, Ether fell 35% and Solana fell 58%. Their conclusion is that public markets are already rewarding real cash flow rather than narrative appeal, and private markets are likely to move the same way.
After stablecoins, the next asset migrations may center on Treasuries, equities and machine payments
The article argues that every asset that moves onchain creates an ecosystem of supporting companies. Stablecoins proved the template. Once dollars moved onchain, the market needed onramps, offramps, cards and treasury tools. Circle and Tether emerged from that stack.
The next two asset classes that may follow the same path are U.S. Treasuries and equities. BlackRock’s money market fund has already put Treasury yield onchain through Securitize. Tokenized equities reached $23 billion in monthly trading volume, with month-on-month growth close to 100%.
The push behind tokenized equities also comes from private markets themselves. Leading companies are staying private longer, leaving holders of private shares to trade among themselves off-market. Secondary trading in private fund interests reached $240 billion in 2025, up 48% year over year. Retail investors remain locked out before IPO. The article notes that Coinbase launched pre-IPO perpetuals in June using SpaceX as the reference asset, giving users synthetic exposure before an actual listing.
A third emerging market, the authors say, is machine payments. AI agents need stablecoins to pay. If software is going to transact autonomously, it needs a software-native settlement currency. Cloudflare and Coinbase jointly launched the x402 payment gateway standard so AI agents can call APIs and scrape webpages using stablecoin settlement instead of card rails.
Xavier Meegan, founder and CIO of Frachtis, is quoted in the article saying, “Agent-native trust is a new category that most AI investors are missing. As AI capabilities improve rapidly, security becomes critical. We can use a decade of blockchain research to secure multi-agent networks. Crypto-native investors are in a position to back one of the biggest markets of the next decade.”
The article then adds another layer: once an asset is tokenized, the full function stack around it also has to migrate onchain. Tokenized assets only unlock their value once they can be used as collateral. That is one reason the authors say onchain credit is a key area for future investment.
Banks, payments and FX are increasingly using crypto as a back-end rail
Banking is presented as one of the sectors furthest along in this transition. Traditional banks failed to keep up with users, which led to new banks. Those new banks then ran into the limits of the payment rails they rented, creating room for crypto-native banking models. Erebor, the article notes, secured a national bank charter this year. OpenFX raised $94 million to build stablecoin-based cross-border FX settlement.
The broader point is that crypto is no longer just a parallel financial system waiting for permission. It is becoming plumbing inside the existing financial system, and the next venture opportunity lies where the two layers meet.
The article offers examples where users barely notice the technology at all. Ramp’s enterprise customers can still settle payments to suppliers in São Paulo and Seattle over a weekend when traditional rails are offline, without needing to think about stablecoins. Over the past year, Visa used stablecoin-linked cards to process $3.7 billion of payments across 1.9 million cards in more than 200 countries, and plans to expand with Bridge to more than 100 countries. Excluding bot activity, the article says, $9 trillion already settles each year through structures like this.
The user may not see the crypto layer. The companies building the rails still collect the fees. Equities and Treasuries, the article says, are only at the start of the same curve. That changes the venture question from “Will this asset move onchain?” to “What support businesses will need to exist once it does?” Onchain equities need brokers, collateralized lending and market makers. Tokenized Treasuries need custody and distribution.
Exit routes are widening beyond token launches
The article argues that block-space supply is now in excess, while profits are flowing to applications. Hyperliquid generates more revenue than the chain it runs on. In 2018, it was just an exchange protocol and a hackathon demo. Today, value is clustering at the application layer, but applications often still lack capital.
As fintech and crypto merge, the authors expect a long tail of products aimed at users who do not know what a cross-chain bridge is. DeFi was almost the only financial product line in crypto for a decade. Now digital banking, brokerage, FX and credit are all being rebuilt on crypto rails.
That shift also changes who starts companies. The article cites Vinod Khosla’s long-held view that the companies that truly reshape an industry are often founded by outsiders. It then gives three examples from crypto. Hyperliquid founder Jeff Yan came from high-frequency trading at Hudson River Trading and wanted a better derivatives exchange. Circle founder Jeremy Allaire had already taken two internet software companies public before working on blockchain. Palmer Luckey, founder of Erebor, also built defense technology company Anduril.
The point is not biography for its own sake. The authors say these founders entered crypto because they needed better infrastructure for a business problem, not because they were searching for a utopian use case. Crypto is becoming infrastructure, more like Linux or databases. Infrastructure itself tends to commoditize. That is where venture has work to do.
It also means token issuance is no longer the only way out. The article says M&A reached a record $8.6 billion, with more than 130 venture-backed companies acquired, mainly by strategic buyers. Naver bought Dunamu for $10.3 billion. Coinbase acquired Deribit for $2.9 billion and later bought Echo. IPO windows have reopened. For the first time in crypto, the article argues, founders can realistically choose between acquisition, IPO and token-based cash return.
Brooke Pollack of Hutt Capital is quoted describing how her firm has adapted: at least three-quarters of its capital now goes to buying existing LP interests in funds, effectively making it a secondary fund with only a smaller allocation to primary investments.
Founders have already adapted; investors now need to catch up
The article closes by turning to investment discipline. Drawing on Fred Wilson’s writing after the internet bust, it argues that too much money funded too many people who were not well suited to venture investing. If capital had been more concentrated in the hands of experienced investors, losses would likely have been smaller.
Wilson also distinguished between thematic investing and thesis-driven investing. In a thematic model, investors identify a broad narrative and fill a portfolio with anything that seems to fit. In a thesis-driven model, they build a picture of how a specific market will evolve over five to 10 years and evaluate every investment against that framework.
The article also cites his comments on holding periods: if early-stage investments are held for seven years on average and a firm adds one or two new deals a year, each partner ends up managing seven to 14 companies at any one time. Early-stage investing is a service business, he said, and founders are the clients.
From that, the authors argue that crypto is moving toward three rules: break out of the industry echo chamber, build real vertical expertise and extend holding periods from a few years to something closer to a decade.
They say founders have already moved first. The era of treating the token itself as the product is ending. Teams are becoming more careful about whether they should issue a token at all and how it should be structured, rather than issuing one simply to create a quick exit for investors. If exits take five years or more, investors need to understand the business, and understanding the business starts with understanding the market it serves.
That longer cycle reduces the viability of scattered allocation across many sectors. Capital has to be more selective. The edge shifts toward specialization: knowing how a market makes money, who the customer is, what strong unit economics look like and which regulatory rules determine who can survive.
The article cites Cambridge Associates research covering 2001 to 2010. In U.S. venture and growth equity, vertically specialized firms produced a 2.2x gross multiple and 23.2% gross IRR, compared with 1.9x and 17.5% for generalist firms. The gap held across consumer, finance, healthcare and technology. Specialized firms were also more restrained when valuations overheated.
Large funds are broadening out, leaving room for smaller specialist seed firms
The article then asks a direct question. If the fundamentals are improving, why are so few investors willing to step in?
One answer is fund structure. Several major crypto-native firms have launched non-crypto funds or widened their mandates. By 2028, the article expects nearly half of distressed mid-sized crypto funds to liquidate or change strategy. Large multi-billion-dollar funds cannot efficiently deploy themselves through repeated $2 million seed checks, so they are moving toward areas such as AI that better match their scale. Mid-sized funds, meanwhile, are often leaving because past performance no longer supports new fundraising.
The seed opportunity remains open, but it now fits small specialist managers much better. The article says fewer than 20 firms are actively doing pre-seed and seed in a serious way. Deals that once closed in three weeks now require long diligence processes. Only eight new funds were created last quarter. Since funds usually deploy over three years after closing, seed competition in 2027 and 2028 may be very limited.
The authors present that as a cyclical reset. Opportunistic capital leaves. Generalists move on. The firms that remain are the ones LPs continue to back through multiple cycles and low-sentiment periods.
Patient capital, not consensus capital, is becoming the scarce asset
The piece ends where it began, with history. Collaborative Fund, it notes, wrote a decade ago that markets reward judgments that are non-consensus at the time but later prove right. The authors say crypto is exactly that kind of bet today. Very few people have enough depth to know which parts of the market are actually working.
Kinjal Shah, general partner at Blockchain Capital, is quoted as saying: “The best entry point is never in the middle of loud consensus. That belief has supported our conviction in digital assets from the start. This is a 10-year story, not a single bull-bear cycle. Capital is concentrating, several categories have already proven product-market fit, and users are at record highs. We believe digital assets will reshape financial markets and networks, and while others are turning away, we are choosing to keep increasing our exposure.”
The article compares the current phase to 2009, not 1999. Infrastructure is mature, regulatory frameworks are arriving and real use cases are visible, but sentiment remains poor. Historically, those have been the periods when some of the highest-DPI venture funds were formed.
The logic, in the authors’ telling, is that opportunists leave and stronger founders remain. LPs no longer need to be convinced that crypto matters in the abstract; large ETP allocations have already done that educational work. The question is which managers can navigate changing cycles, revise strategy and win access to the best founders through hard-earned market understanding.
The authors also suggest that when capital returns in force, it may not even be called crypto investing anymore. Just as few people now call themselves internet investors, crypto may dissolve into vertical categories such as brokerage, payments and banking. Amazon was once framed as an internet company. Netflix was too. Today they are classified very differently. Crypto may follow that path, with the companies built on these rails eventually described by function rather than by underlying architecture.
The final claim is a macro one. Verification, settlement and custody are trust services embedded in nearly every financial transaction, representing a global economic layer worth about $29 trillion, according to the article. AI is reducing the cost of intelligence. Stablecoins, the authors argue, are reducing the cost of trust. Once economic functions migrate to a more efficient onchain settlement layer, they are unlikely to move back.
That is why small teams with a few rounds of capital may now be able to build meaningful products around each newly tokenized function. Capital lasts longer than it used to. AI handles more of the software work. At a time when talent has also left the sector, judgment, distribution and patience become the scarce inputs.
The article describes that as the core value of venture capital in this cycle. Fred Wilson called it “slow capital” in 2009. By 2026, Will Mandis calls it “patient capital.” In a market crowded with low-quality generated content and an oversupply of financial engineering, the scarce advantage belongs to investors who can wait, who can discriminate and who know what is worth waiting for.
