Bankless used its latest podcast episode to make a broader point about where the next wave of crypto repricing may come from: not necessarily new blockchains, but applications that already have revenue, users, and a mechanism that sends part of that value back to the token.
The episode featured Austin Barack, founder and managing partner at Relayer Capital, in a discussion centered on Venice, Hyperliquid, Pump.fun, and ether.fi. The conversation, later compiled in Chinese by Peggy at BlockBeats and published by MarsBit, asked how application tokens should be valued and whether crypto is moving from an infrastructure cycle into an application cycle.
From narrative chasing to a search for growth and value
The main framework in the episode was that token selection is changing. In earlier cycles, the market often priced projects around new chains, new protocols, token incentives, and long-dated narratives. After a long compression in valuations, the gap has widened between projects with genuine product-market fit and fast-growing revenue, and tokens that still lack meaningful usage.
Barack said one of the few ideas that can work across cycles is to find the intersection of growth and value: projects growing fast enough that the market has not fully priced them in. He drew a distinction between that and traditional value investing. In his telling, crypto is not attractive because it offers slow-growing mature businesses at low multiples. What makes the asset class unusual is that broad market weakness can leave valuations suppressed even while underlying business growth expands severalfold.
Barack said Relayer Capital was founded about two and a half years ago and originally split its attention roughly evenly between early-stage investing and liquid markets. Now, around 95% of the firm’s focus has shifted to liquid tokens, with Crypto x AI and 24/7 trading plus tokenized assets as two primary themes. He tied that shift not only to the possibility of a stronger market, but also to the fact that a prolonged bear market has already filtered out many narrative-driven assets.
Buybacks and burns are creating a new valuation language
A major thread in the podcast was that what matters is not just how much a protocol earns, but whether that revenue reaches the token through a stable and transparent buyback or burn mechanism. For years, crypto projects could produce revenue without any clear link between business performance and tokenholder outcomes. That is starting to change in some applications.
Barack argued that programmatic buybacks and burns let investors borrow part of the equity toolkit and look at buybacks relative to token market value as a rough earnings-yield style metric. He also stressed that this is not the same thing as a price-to-earnings ratio in public equities. Shares usually come with legal claims on residual earnings and assets. Tokens often do not. Teams can change buyback ratios, and new businesses can be housed under equity entities rather than under token-linked structures.
The episode’s TL;DR made the same point directly: buyback multiples cannot simply be treated as equity P/E ratios because tokens typically lack clearly defined residual claims, and the split of value between equity and token remains a core risk.
Venice and VVV: revenue growth, token burns, and a scenario model
Venice was one of Barack’s central case studies. He described it as an AI application built around privacy and censorship resistance, where users can access different frontier and open-source models from one platform. Current revenue mainly comes from paid subscriptions and purchases of extra compute credits.
According to the discussion, Venice has created two programmatic burn paths for VVV. When users first buy different subscription tiers, the platform burns a corresponding amount of VVV. When users buy additional credits, about 5% of the purchase amount is used to burn the token.
Barack estimated that as of August 2026, Venice was running at an annualized revenue rate of about $107 million, with annualized token burns of roughly $8.3 million. He projected that by 2027, revenue could reach $336 million and token burns could rise to $70 million. At a 50x buyback multiple, that model implied a token valuation of around $3.5 billion. Based on his projected circulating supply at that point, he put VVV at about $43.9, versus a price of around $16 when the episode aired.
He did not present that framework as a firm forecast. The episode said the model rests on optimistic assumptions. Of the projected $70 million in burn value, about $29 million was tied to Minds, a product that had not formally launched yet and accounted for more than 40% of the total.
Minds is described as a product that would let advanced users and developers combine different models, prompts, and tools to build structured AI applications for mainstream users and earn a share of usage-based revenue. The form factor was described as close to an AI app store.
Host David Hoffman pushed back on the model at that point. He said buying credits is an extension of an existing service, while Minds is an unproven new business line, and the two should not carry the same risk weighting. Barack agreed with that criticism and called his model slightly above a base case. Using a scale where 0 is extremely bearish, 5 is base case, and 10 is fully optimistic, he said his assumptions sit around 6.
The model also assumes Venice could eventually include renewals in its burn framework and raise the burn ratio on credit revenue from 5% to 10% in 2027. The episode said those measures have not been formally committed, so the $43.9 figure is better understood as a scenario valuation built on multiple business and mechanism assumptions, not an unconditional price target.
Why return cash to a token so early?
The Venice case opened a more basic question: why would a fast-growing startup spend cash on token buybacks or burns instead of reinvesting it into expansion?
The comparison in the episode was straightforward. In traditional markets, high-growth companies usually direct most of their capital to hiring, research, development, and customer acquisition. Large-scale buybacks rarely come early. Venice, by contrast, has been returning part of its business revenue to VVV from an earlier stage, which means less capital available for reinvestment.
Barack said this makes more sense in crypto’s dual equity-token structure. Tokens can help a project gather attention, bootstrap networks, and create new product features. But without clear legal constraints, the market does not automatically believe all company value will eventually accrue to the token. In that setting, a programmatic burn is not just a distribution choice. It is also a trust-building mechanism.
He described Venice’s current setup as gradual. Early burns involved some discretion, then the platform added burns tied to first-time subscriptions, then brought 5% of credit purchases into the burn model. In his view, that preserves a large share of capital for growth while still giving the token an identifiable value-return path.
The write-up also noted that Venice had previously raised $65 million. Barack’s interpretation was that outside financing provided expansion capital, allowing more operating cash flow to be directed toward the token. He added that investors with token warrants can reduce some of the misalignment between equity holders and token holders.
Even then, the balance remains fragile. If growth slows, inference costs rise, or competition intensifies, the company may need to retain more cash. If burn ratios stay too low, the token may fail to capture enough of the business upside. The episode said VVV should be assessed not only through burn totals, but also through revenue growth, gross margins, operating expenses, and whether the team continues to honor the value-return framework it has outlined.
Pump.fun: a low multiple tied to doubts about revenue durability
Pump.fun was used as a contrast case. Its revenue is more directly tied to crypto trading cycles. Barack said that based on market data at the time of the episode, Pump.fun was trading at about 5x buyback value, while Hyperliquid and Lighter were at roughly 30x to 40x. In his view, that gap says more about market bias toward different revenue types than about a simple hierarchy of business quality.
Pump.fun’s core business comes from meme coin issuance and trading. Many investors see that revenue as heavily dependent on short-lived speculative enthusiasm and less durable than revenue from perpetuals venues. The episode acknowledged that this concern has precedent, since crypto has seen products whose revenue surged in one cycle and then fell by more than 90%.
Barack’s counterargument was that Pump.fun’s performance over more than two years shows stronger resilience than the market initially expected. A single meme coin may fade quickly, but demand for highly volatile, high-variance speculative products may persist over longer periods. He compared Pump.fun to casinos, lotteries, prediction markets, and very short-dated options. The purpose of that comparison was not to say the products are identical, but to explain a type of user demand that can persist even when expected returns are negative for participants as a group.
On that basis, he said Pump.fun’s buyback multiple could move from roughly 5x to 10x. If business size stayed the same, multiple expansion alone could imply roughly 2x upside. If on-chain trading activity and meme coin activity recover together, revenue could also increase.
There is another side to that case. The episode said Pump.fun had at one point used all of its revenue for buybacks, but later changed that policy so that 50% of revenue over the next 12 months would go to buybacks and the rest would be used for business development. Whether that ratio continues after the 12-month period remains undecided.
That matters because the market can track the authenticity of Pump.fun’s revenue on-chain, but cannot assume a permanent claim on a fixed share of it. Any valuation of PUMP has to discount that institutional uncertainty rather than treating total platform profit as if it fully belongs to token holders.
Hyperliquid: higher multiples and stronger reflexivity
Hyperliquid received a higher valuation multiple in the discussion. One reason is that its crypto perpetuals business already produces substantial revenue. Another is that the HIP-3 market is expanding the set of tradable contracts to include stocks, commodities, indices, and contracts tied to private companies.
Barack said Hyperliquid shows what 24/7 blockchain-based trading, instant settlement, and global price discovery could look like at scale. He went further and said that in the future, some private assets might first form price signals on-chain, with traditional financial institutions later using those signals as references for issuance pricing.
The episode did not present that as settled. According to Barack, newer real-world asset markets on Hyperliquid have contributed meaningful volume, but because those markets are still in expansion mode, revenue growth has not matched that volume contribution yet. The highest-margin business remains crypto trading.
That is why Hyperliquid was framed as more reflexive than a typical application token. When crypto capital flows back into the market, HYPE may benefit not only from a broader rerating in tokens, but also from higher trading volume, larger fee generation, and more buybacks. If market activity fades, the same mechanism can work in reverse.
ether.fi: the business changed, the market category did not
Barack described ether.fi as a different kind of mispricing. The issue there is not simply multiple compression. It is that the business mix has changed while the market still classifies the project under an older label.
ether.fi entered the market as a liquid restaking protocol. At the peak of the 2024 restaking narrative, its fully diluted valuation briefly reached about $8 billion. As expectations around the restaking segment faded, ether.fi’s valuation fell as well, and it continued to be treated more like an asset in the same family as Lido and other staking protocols.
Barack said that framing no longer matches the company’s revenue mix. Based on data he cited in the episode, more than 65% of the business now comes from Neo Bank products, including credit card transaction revenue and lending income generated when users borrow against assets in their accounts. Yield and staking now account for roughly 35%.
As the platform adds tokenized stocks and more on-chain assets, ether.fi is moving from a digital neobank toward a more comprehensive on-chain brokerage model. Users can hold and trade different assets, borrow against those assets, and use a credit card for everyday spending.
One advantage of that model, as framed in the discussion, is that ether.fi does not need to build every part of the financial stack from scratch. On lending, for example, it can plug into existing DeFi protocols such as Aave and participate in revenue sharing. The richer Ethereum becomes in stablecoins, lending markets, and tokenized assets, the broader the service set ether.fi can offer to users.
Barack said ether.fi’s credit card daily transaction volume had increased from about $300,000 a year earlier to $3 million to $4 million, a gain of more than 10x. He also said only about 4% of current revenue comes from lending interest, compared with roughly 60% to 70% for the traditional digital bank Nubank. In his view, that points to room for further expansion in ether.fi’s revenue structure.
Using a potential $30 million in buybacks over the next 12 months and a 30x multiple, Barack estimated that ETHFI could be worth more than $1, about twice the price when the episode aired. The write-up added an important caveat: the $30 million figure is above another model he cited that projected $21 million, and his case assumes faster future growth. That makes the result another optimistic scenario rather than a firm target.
The broader point, the episode said, is about whether the market’s category system has fallen behind the business itself. If most of ether.fi’s revenue now comes from payments, lending, and brokerage-style activity, valuing it with a liquid restaking framework may miss what the project has become. If that newer business line fails to sustain growth, then the reclassification thesis can also break down.
Fundamentals can reduce dependence on the market, not eliminate it
The discussion closed by drawing a line around how far fundamentals can take application tokens. Real revenue can provide a valuation floor. It does not remove crypto cyclicality.
Barack described the relationship as partly coupled and partly decoupled. Venice, Pump.fun, Hyperliquid, and ether.fi can build a valuation basis from their own user growth, revenue, and buybacks. Even if Bitcoin moves sideways, their tokens can still rerate if the business keeps expanding. But they remain crypto assets. When capital rotates from equities, AI, and other markets back into tokens, projects with visible fundamentals may be among the first to attract professional allocations. For Pump.fun and Hyperliquid, higher trading activity could also add revenue on top of a broader market move, creating positive feedback between token price and business performance.
Venice sits a bit differently because its direct connection to crypto trading cycles is weaker. Its more relevant outside variable is AI usage. If multi-model access, privacy-focused AI, and generative applications keep growing, Venice may draw demand from outside purely crypto-native use cases. If user growth or paid conversion falls short, the token does not automatically earn its way to the valuation embedded in the model just because crypto prices rise.
Barack framed the whole shift in longer-cycle terms as well. Citing data mentioned in the episode, he said execution-layer infrastructure accounted for more than 95% of industry revenue for most of crypto’s history. Now application revenue has risen to about two-thirds of the total. He expects that share to keep moving toward applications and eventually exceed 90%.
That is still a forecast, not a settled fact. The variables that need to be tested are whether revenue continues shifting from chains and execution layers to user-facing applications, whether application cash flow can be transmitted to tokens in a durable way, and whether buyback mechanisms can remain intact through business growth, market downturns, and regulatory change.
If those conditions hold, the repricing focus in crypto may shift again. The market may spend less time asking which infrastructure project has the grandest story, and more time asking which applications can turn real usage into recurring revenue and return a credible share of that value to the token.