Video Title: Crypto's next winners won't be blockchains
Video Author: Bankless
Compilation: Peggy, BlockBeats
Editor's Note: Against the backdrop of a prolonged valuation contraction and capital reflow in the crypto market, discussions on token investment are shifting from 'what's the next hot narrative' to 'which projects have already established verifiable business models.' However, as metrics like revenue, buybacks, and burns are becoming a new valuation language, a more crucial question emerges: Can tokens be priced like stocks based on cash flows, and do their holders truly own the value created by the protocol's growth?
This episode of the Bankless podcast invited Austin Barack, Founder and Managing Partner of Relayer Capital, to discuss valuation methods for application tokens, using projects like Venice, Hyperliquid, Pump.fun, and ether.fi as examples. The conversation also explored the potential path for the crypto market's migration from an infrastructure cycle to an application cycle.
In this dialogue, Austin's approach was not simply to find tokens with the highest revenue or the strongest buyback programs. Instead, he deconstructed token investment into a set of more fundamental structural questions: Does the product have genuine demand? Can revenue grow sustainably? Can the business value be reliably transmitted to the token? And is the market still pricing businesses that have changed using old classifications?
First, the screening logic for crypto assets is shifting from merely chasing growth to seeking the intersection of 'growth and value.' In the past, the industry typically relied on new Layer 1s, new protocols, and token incentives to create growth expectations, with valuations largely reflecting forward-looking narratives. A prolonged bear market has compressed this premium, gradually differentiating a few projects that have found product-market fit and rapid revenue growth from a multitude of tokens lacking real utility. This means a downturn doesn't just bring price discounts; it also provides investors a window to identify real businesses. The projects truly worth attention need to possess both growth speed and reasonable valuation, not just one or the other.
Second, token value is beginning to shift from abstract 'utility' towards observable value flow-back. Venice uses a portion of revenue from new subscriptions and credit purchases to burn VVV; Hyperliquid uses most of its platform revenue to buy back HYPE; Pump.fun and ether.fi have also established their own buyback mechanisms. In the past, there was often a lack of a clear link between protocol revenue and token performance; project growth didn't necessarily translate into holder gains. Now, programmatic buybacks and burns are establishing valuation anchors for tokens, similar to discounted cash flow models. However, this stock-like framework still has its limits: buyback ratios can be adjusted, and the rights relationship between equity entities and tokens is not yet fully institutionalized. What investors truly need to assess is not just revenue size, but also the sustainability and credibility of the value flow-back mechanism.
Third, the quality of revenue is more important than the revenue itself. The market long gave Pump.fun a relatively low valuation, partly because investors doubted the sustainability of meme coin trading demand and struggled to understand user demographics different from their own. As platform revenue has demonstrated resilience for over two consecutive years, this perception is changing. Similarly, Venice's valuation depends not only on current subscription revenue but also on its ability to evolve from a multi-model gateway into an AI application platform connecting developers and ordinary users. This means valuation cannot mechanically apply buyback multiples; it also requires judging whether revenue comes from temporary incentives and hype or from a recurring user behavior pattern.
Fourth, the market's old classifications for projects might be creating new pricing anomalies. ether.fi was once viewed as a liquid restaking protocol, but currently over 60% of its business already comes from Neo Bank products like credit cards and lending, and it is further expanding into a comprehensive on-chain brokerage platform. If the market still prices it according to the restaking sector, it might overlook the changes in its revenue structure. More importantly, ether.fi can directly leverage lending, stablecoin, and tokenized asset infrastructure on Ethereum to expand its product offerings with relatively low organizational and capital investment. This suggests that the true value left over from the infrastructure cycle may not continue to concentrate on the underlying protocols, but could be captured by the applications best at packaging these capabilities to serve users directly.
Fifth, while application revenue can provide a valuation floor, it cannot completely detach tokens from crypto cycles. Projects with buyback mechanisms can establish a relatively independent pricing foundation based on business growth, but they still belong to the token asset class and are affected by market capital flows, BTC and ETH trends, and changes in on-chain activity. The difference is that during a market upswing, trading-focused applications like Hyperliquid and Pump.fun may benefit from a dual push of capital inflow *and* business expansion; when the market weakens, real revenue becomes an important buffer distinguishing them from purely narrative-driven assets.
If this conversation is compressed into one judgment, it would be this: The core of the next crypto asset revaluation may no longer be about who has the grander infrastructure narrative, but rather about who can convert real usage into sustainable revenue and credibly return a portion of it to token holders. In this sense, this article discusses not just whether a few tokens are undervalued, but whether the crypto market can evolve further from a narrative-driven financing system into an application economy based on products, cash flow, and value distribution.
The following is the original content (edited for readability):
TL;DR
· The core opportunity in the crypto market is shifting from underlying infrastructure to the application layer. The essence is that revenue and users are starting to replace block space narratives as the new source of value.
· Whether application tokens can be revalued depends not on how much money a protocol earns, but on whether that revenue can be transmitted to the token through stable, transparent buyback or burn mechanisms.
· Venice combines AI application growth with a token burn logic, but its $43.9 target price relies on optimistic assumptions like Minds launch and increased burn ratios, and should not be viewed as a definitive valuation.
· Pump.fun's low valuation primarily reflects market skepticism about the sustainability of meme coin revenue, but over two years of revenue resilience suggests that high-volatility speculative demand might be a type of long-term consumer behavior.
· Hyperliquid has stronger cyclical reflexivity than typical applications: capital inflows can both lift HYPE valuation and simultaneously boost trading volume, fees, and buyback scale.
· ether.fi is still priced as a restaking protocol, but its main revenue has shifted to payments and lending. The market's old classifications for projects may not have caught up with the change in business structure.
· Buyback multiples cannot be directly equated to stock P/E ratios because tokens typically lack clear residual earnings rights. Value allocation between equity and tokens remains a core risk.
· Fundamentals can reduce a quality token's correlation with the broader market but cannot eliminate crypto cycles. Truly sustainable valuation still depends on revenue quality, value flow-back, and mechanism continuity.
Main Points
Investment methods in the crypto market are never fixed.
Strategies that worked in 2017 may not apply in 2021; sectors that were popular in 2021 or 2024 could lose their appeal in the next cycle. In Austin Barack's view, one approach that can be reused across cycles is to find the intersection of growth and value: projects growing fast enough, but whose valuations do not yet fully reflect that growth.
This is not traditional value investing. Investors enter the crypto market not to find a mature company growing 10% annually with a P/E of 4x. The real allure of crypto assets lies in their intense capital cycles potentially creating a combination less common in traditional markets: businesses growing several-fold, while valuations are suppressed by the overall market downturn.
Barack founded Relayer Capital about two and a half years ago, with a strategy covering both early-stage investments and liquid markets. At inception, the fund's attention was roughly split between the two, but now about 95% of its focus is on liquid tokens, primarily concentrated on two themes: Crypto×AI and all-weather trading & asset tokenization.
The reason isn't just that the crypto market might be re-entering an up cycle. Barack believes that the prolonged bear market has helped the market complete a screening process: as most tokens lost their narrative premium, a handful of projects that truly found product-market fit, have rapidly growing revenue, and still have relatively reasonable valuations began to stand out.
From Chasing Narratives to Calculating Buybacks: Tokens Begin to Have a New Valuation Language
For a long time, the valuation of crypto projects relied heavily on forward-looking assumptions like market size, network effects, and token utility. Even if protocols generated revenue, there was often a lack of a clear link between that revenue and the token.
Now, some applications have begun directly converting business revenue into token buy-side pressure or supply contraction through programmatic buybacks or burns. This allows investors to borrow some methods from stock valuation, approximating a token's 'earnings yield' by observing the ratio of buyback amount to token market cap.
However, this method cannot be directly equated to a P/E ratio.
Stocks typically represent legal rights to a company's residual earnings and assets, whereas token holders may not have equivalent rights. Project teams can change buyback ratios and may place new businesses under the equity entity. Therefore, buyback multiples have strong explanatory power only when the value flow-back rules are relatively transparent and business revenue is sustainable.
Venice is a key case discussed by Barack. It's an AI application emphasizing privacy and non-censorship, where users can call different frontier and open-source models on the same platform. Current revenue primarily comes from paid subscriptions and additional compute credit purchases.
Venice has also established two types of programmatic burn mechanisms for VVV: when users first purchase subscriptions of different tiers, the platform burns a corresponding amount of VVV; when users buy extra credits, about 5% of the purchase amount is used to burn tokens.
According to Barack's estimates, as of August 2026, Venice's annualized revenue run rate is about $107 million, corresponding to an annualized token burn of about $8.3 million. He projects that by 2027, revenue could increase to $336 million, with the burn amount potentially rising to $70 million. Applying a 50x buyback multiple, his model suggests a token valuation of about $3.5 billion; factoring in the projected circulating supply at that time, the target price for VVV is about $43.9, while the price at the time of recording was around $16.
This model carries explicit optimistic assumptions and is not a definitive prediction of future revenue.
Of the projected $70 million burn amount, about $29 million comes from the Minds product, which has not officially launched yet, accounting for over 40%. Minds plans to allow advanced users and developers to combine different models, prompts, and tools to create structured AI applications for regular users, earning revenue share based on usage, resembling an AI app store.
Barack believes Minds is not a completely new product detached from Venice's existing business, as it still revolves around existing models, users, and usage scenarios. However, host David Hoffman pointed out that credit purchases are merely an extension of existing services, while Minds is an unproven new business line, and the risks of the two are not equivalent.
Barack acknowledged this critique and described his model as 'slightly above base case': on a scale where 0 is extremely pessimistic, 5 is base case, and 10 is wildly optimistic, he placed these predictions around a 6.
The model also assumes Venice may incorporate subscription renewals into the burn mechanism in the future and increase the burn ratio for credit revenue from 5% to 10% by 2027. These measures are not yet confirmed commitments. Therefore, the $43.9 figure is better understood as a scenario-based valuation built on multiple business and mechanism assumptions, not an unconditional price target.
Venice's Core Dilemma: Why Would a Startup Repurchase Tokens So Early?
The Venice case also reveals a core contradiction facing application tokens: Should a high-growth startup reinvest cash into product expansion or return it to token holders?
In traditional markets, companies in high-growth phases typically allocate most funds to R&D, hiring, and customer acquisition, rarely engaging in large-scale stock buybacks early on. Yet Venice has been using a portion of its revenue to buy back and burn VVV from the early stages of business development, somewhat sacrificing capital available for reinvestment.
Barack believes this practice is related to the dual equity-token structure in crypto markets. Tokens can help projects quickly gather attention, bootstrap networks, and design new product features. However, in the absence of clear legal constraints, the market cannot naturally believe that all value created by the company will ultimately accrue to the token.
Therefore, programmatic burns are not just a value distribution method but also a mechanism to build trust. Teams need to demonstrate through action that business growth can be transmitted to VVV, not just remain within the equity entity.
Venice currently employs a gradual approach: early burns had some discretionary nature, followed by adding first-time subscription burns, and then allocating 5% of credit purchase revenue to burns. Barack sees this arrangement as providing value flow-back for the token while retaining most funds for growth.
Venice's previous $65 million funding round also somewhat alleviates the conflict between buybacks and reinvestment. Barack's understanding is that external funding provided capital for expansion, allowing the company to allocate more operating cash flow to the token; relevant investors also hold token subscription rights, which helps align interests between equity investors and token holders.
However, this balance remains fragile. If business growth slows, inference costs rise, or competition intensifies, the company may need to retain more cash. Conversely, if the burn ratio remains too low for too long, the token may struggle to fully share in business growth. Therefore, judging VVV's value requires not only observing the total burn amount but also simultaneously tracking revenue growth rate, gross margin, operating expenses, and whether the company consistently executes its value flow-back commitments.
Pump.fun and Hyperliquid: Both Have Revenue, So Why Does the Market Assign Different Multiples?
Compared to Venice, Pump.fun and Hyperliquid's revenue is more directly linked to crypto trading cycles.
Barack stated that based on market data at the time of recording, Pump.fun was valued at roughly 5x its buyback amount, while Hyperliquid and Lighter had corresponding multiples around 30-40x. In his view, this gap reflects market bias towards different types of revenue.
Pump.fun's core business comes from meme coin issuance and trading. Many investors believe such activity relies on short-term speculative hype, making its revenue less sustainable than that of perpetual futures platforms. Similar concerns aren't unfounded: the crypto industry has seen products whose revenue rapidly rose within one cycle and then declined by over 90%.
But Barack argues that Pump.fun's performance over the past two-plus years shows its revenue is more resilient than the market initially expected. While individual meme coin hype may fade quickly, user demand for high-volatility, high-variance speculative products might be long-lasting.
He analogizes Pump.fun to casinos, lotteries, prediction markets, and ultra-short-term options. The point here isn't to equate meme coin trading with these products completely, but to explain a demand mechanism: even if participants face negative expected returns overall, some users will continue to participate due to high volatility and small-probability high payouts.
Based on this judgment, Barack believes Pump.fun's buyback multiple could re-rate from around 5x towards 10x. If business scale remains unchanged, this multiple expansion itself could correspond to roughly a doubling in price; if on-chain trading and meme coin activity revive simultaneously, revenue could grow further.
However, Pump.fun's risk also stems from the relationship between equity and tokens. The project initially allocated 100% of revenue to buybacks but later adjusted to using 50% of revenue for buybacks over the next 12 months, with the rest invested in business development. Whether this ratio continues after 12 months remains to be decided.
This means the authenticity of Pump.fun's revenue can be observed via on-chain data, but there's no permanent guarantee of how much revenue the token can sustainably receive. When valuing PUMP, investors need to apply a discount for this institutional uncertainty, not directly treat all platform profits as token holder earnings.
Hyperliquid, on the other hand, commands a higher valuation multiple. On one hand, its crypto perpetuals business already generates substantial revenue; on the other, its HIP-3 market is expanding trading to stocks, commodities, indices, and contracts related to private companies.
Barack believes Hyperliquid showcases the potential of blockchain for all-weather trading, instant settlement, and global price discovery. In the future, some yet-to-be-listed assets might even form price signals on-chain first, which traditional financial institutions could then reference for issuance pricing.
But this view still awaits market validation. According to Barack, Hyperliquid's recently added real-world asset markets have contributed significant trading volume, but because they are still in an expansion phase, they haven't yet brought proportional revenue growth. Currently, the higher-margin business remains primarily crypto asset trading.
Therefore, Hyperliquid exhibits stronger cyclical reflexivity than typical applications: when crypto capital flows back, HYPE may benefit not only from a general token valuation recovery but also from simultaneous growth in platform trading volume, fees, and buyback scale; if market activity declines, this mechanism can also work in reverse.
ether.fi Has Changed, But the Market's Classification Hasn't Caught Up
ether.fi represents another valuation dislocation: the project's main business has changed, but the market still prices it using old labels.
ether.fi initially entered the market with liquid restaking services. At the peak of the restaking narrative in 2024, its fully diluted valuation once reached around $8 billion. As market expectations for the restaking sector cooled, ether.fi's valuation declined and continued to be viewed as an asset similar to staking protocols like Lido.
Barack believes this classification no longer accurately reflects ether.fi's current revenue structure. According to data he provided during the episode, over 65% of business now comes from Neo Bank products, including credit card transaction revenue and lending revenue generated from users borrowing against account assets; yield and staking services now account for only about 35%.
As the platform adds tokenized stocks and more on-chain assets, ether.fi is evolving further from a digital neo-bank into a comprehensive on-chain brokerage platform. Users can hold and trade different assets, borrow against them, and use a credit card for daily spending.
The advantage of this model is that ether.fi doesn't need to build all financial infrastructure from scratch. For lending, for example, the platform can call upon existing DeFi protocols like Aave and earn revenue share. The richer the stablecoin, lending market, and tokenized asset infrastructure on Ethereum becomes, the broader the service range ether.fi can offer users.
According to data provided by Barack, ether.fi's credit card daily transaction volume has grown from around $300,000 a year ago to $3-4 million, an increase of over 10x; currently, only about 4% of its revenue comes from lending interest, whereas traditional digital bank Nubank derives about 60-70% of its revenue from this segment. In his view, this indicates ether.fi still has significant room to expand its revenue structure.
Based on a potential forward 12-month buyback amount of about $30 million and a 30x valuation multiple, Barack calculates that ETHFI's price could exceed $1, roughly double the price at the time of recording. However, the $30 million figure is higher than the $21 million predicted by another model he referenced, and he assumes ether.fi's future growth might accelerate, so this result also falls into a relatively optimistic scenario.
The truly noteworthy aspect of this case is not the specific target price, but whether market classification is lagging. If most of ether.fi's revenue already comes from payments, lending, and brokerage, continuing to use a liquid restaking protocol valuation framework might not reflect its current business; conversely, if new business growth fails to sustain, the so-called 'reclassification' may also not hold.
Fundamentals Can Reduce Correlation, But Cannot Eliminate Crypto Cycles
Having real revenue doesn't mean application tokens can completely decouple from Bitcoin and crypto market cycles.
Barack describes this relationship as 'partially coupled, partially decoupled.' On one hand, Venice, Pump.fun, Hyperliquid, and ether.fi can establish a relatively independent valuation foundation based on their own user growth, revenue, and buybacks. Even if Bitcoin moves sideways, tokens could still be revalued as long as the business continues to expand.
On the other hand, they still belong to the crypto asset class. When capital flows back from stocks, AI, and other markets into tokens, these projects with fundamental support might be among the first to enter professional investors' portfolios. Pump.fun and Hyperliquid would additionally benefit from increased trading activity, creating positive feedback between asset price and business fundamentals.
Venice's direct link to crypto trading cycles is relatively weaker; its main external variable is AI usage. If multi-model access, privacy AI, and generative applications continue to grow, Venice may have a demand source different from purely crypto applications; if user growth or monetization falls short of expectations, its token won't automatically achieve the model's valuation simply because the crypto market rises.
Barack ultimately places this change within a longer industry cycle. Citing data, he notes that for most of crypto history, execution layer infrastructure contributed over 95% of industry revenue; today, application revenue share has risen to about two-thirds. He expects this proportion to continue tilting towards applications, eventually exceeding 90%.
This prediction is not yet a fact, but it points to the core variables needing verification in the next phase: Will revenue persistently migrate from Layer 1s and execution layers to user-facing applications? Can the cash flow generated by applications be stably transmitted to tokens? And can buyback mechanisms maintain continuity through business growth, market downturns, and regulatory changes?
If these conditions hold, the primary objects of valuation in the crypto market may further shift from 'infrastructure providing block space' to 'applications selling financial and digital services using blockchain.' At that point, what the market seeks will no longer be just the next high-performance Layer 1, but rather which products truly connect the crypto world with external demand, and which tokens can sustainably share in that growth.
