Private companies don’t need to issue shares for traders to speculate on their valuations. That’s what makes pre-IPO perpetuals different from tokenized equities. Take a private company like Anthropic. A perpetual market can offer price exposure to its estimated valuation without the exchange holding Anthropic shares or having access to its shareholder register. Here’s how the structure works. … α/ Reference Valuation The contract needs a benchmark. Depending on the venue, that might come from private-market transactions, external valuation estimates or a market-derived index. Because there’s no continuously traded public stock, the quality of this reference is critical. β/ Mark Price The venue calculates a mark price used to determine unrealized profits, losses and liquidation risk. This may differ from the latest traded contract price to reduce manipulation from isolated trades. γ/ Funding Payments Perpetual contracts have no fixed expiration. Periodic funding payments between longs and shorts help manage differences between the contract price and its reference. The exact funding mechanism depends on the exchange. δ/ Collateral and Margin Traders deposit collateral, often stablecoins, to open positions. The exchange records their exposure and margin requirements, but no private shares change hands. A $10,000 long position with $2,000 collateral represents 5x exposure to the contract’s price movement, not $10,000 of equity ownership. ε/ Liquidation If the mark price moves against the trader and their margin falls below the required maintenance level, the exchange can liquidate the position. This happens entirely within the derivatives system. ζ/ Financial Settlement When a trader closes a position, the venue calculates the difference between entry and exit prices. Profits and losses are credited or debited through the venue’s settlement system. The trader receives no company shares. > No ownership transfers. > No voting rights. > No shareholder registration. And unless the contract explicitly provides otherwise, an eventual IPO doesn’t automatically convert the perpetual into public equity. … The entire structure separates price exposure from legal ownership. A private company’s valuation becomes something traders can take long or short exposure to, while the actual shares remain outside the derivatives market. The critical question is how reliably the venue can price that exposure when the underlying equity has no continuous public market.

Pre-TGE projects are easy to get bullish on when the technology sounds impressive. But I think the better way to research them is to separate what’s already shipped from what’s still developing. @primus_labs is one of the privacy infrastructure projects on my research list, so here’s my scorecard across seven areas. Not a price prediction. Just products, technology and execution. … α/ Working Applications — 9/10 Primus has moved beyond cryptographic research. > Confidential Vault on BNB Chain > Proof-of-Reserves applications > zkTLS SDKs and verification contracts The Vault reportedly attracted $1M+ in encrypted deposits on day one. β/ Ecosystem Integrations — 9/10 Unitas, Brevis, Perena and AllScale cover different financial workflows. > Unitas: Reserves and confidential finance > Perena: Onchain reserve verification > AllScale: Encrypted payroll testing I like that these aren’t four versions of the same product. γ/ Developer Activity — 8.5/10 Primus’s GitHub shows active development. > zkTLS SDK and extension updates > FHE and MPC tooling > Reserve-verification repositories Multiple components are progressing simultaneously. δ/ Cryptographic Architecture | 9/10 zkTLS authenticates external information, while FHE enables encrypted computation. AutoHoG reportedly processes encrypted 64-bit transfers in ~90ms. The combination extends beyond private payments. ε/ Business Model — 8.5/10 Primus documents an institutional Proof-of-Reserves service offering: > Configurable verification > Public dashboards > Enterprise deployment > Subscription pricing There’s already a commercial product structure, even without disclosed revenue. ζ/ Security & Verification — 8/10 Public code, cryptographic documentation and verifiable computation provide a foundation for evaluating the architecture. Independent audits and production security evidence remain important. η/ Adoption Potential — 8.5/10 Confidential finance, reserve verification, payroll and authenticated financial data give Primus several potential customer categories. The next milestone is recurring usage across these applications. … Overall Research Score: 8.6/10 IMO, Primus has assembled something beyond a standalone privacy feature. There are working applications, developer infrastructure, financial integrations and a defined enterprise offering. The strongest part is how these components connect. > zkTLS establishes authenticity. > FHE protects computation. > Financial applications turn both into products. That’s why Primus remains one of my higher-conviction pre-TGE infrastructure projects. The next phase is converting that technical foundation into sustained commercial adoption.

I’ve always found the separation between Solana and EVM development a bit unnecessary. You have Solana offering fast, cheap execution, while a huge portion of crypto’s applications are still built with Solidity. Developers end up choosing between the two, even when their applications could benefit from both. That’s why I like what @RomeProtocol is building. Rome runs an EVM directly inside Solana as a single program, rather than treating the two environments as separate chains that need bridging. The setup allows: 1. Solidity contracts to call Solana programs in one atomic transaction. 2. EVM and Solana users to keep their respective wallets. 3. Both sides to interact with the same contract state and liquidity. And this is already live on Solana mainnet, with a Halborn audit and public code. You can also trace EVM transactions back to their Solana instructions on the Via explorer. For me, the appeal is pretty simple. Developers shouldn’t have to abandon familiar tooling just to build applications that tap into Solana. Rome gives them a way to work across both environments without splitting the underlying state.

Four crypto treasury companies, four different approaches to growing shareholder exposure. The October 26 disclosures from @BitMNR, @Strategy, @defidevcorp and Strive show how differently these companies are allocating capital. α/ BitMine (BMNR) > ETH holdings: 6.05M > Weekly accumulation: +10,452 $ETH > Weekly buybacks: 2.04M shares > 2026 buybacks: $399.6M BitMine is accumulating $ETH while repurchasing common shares. It also has 5.23M $ETH staked, generating additional $ETH rewards. β/ Strategy (MSTR) > $BTC holdings: 848,517 > Weekly purchases: 0 $BTC > STRC repurchases: $204.9M > USD reserve: $4.83B Instead of buying more $BTC, Strategy is retiring preferred capital to reduce outstanding senior claims and future dividend obligations. γ/ DeFi Development Corp. (DFDV) > $SOL holdings: 2.60M > Weekly increase: +5,346 $SOL > CHAD dividend: 13% → 14% on Nov. 2 DFDV combines $SOL accumulation and staking with preferred financing, avoiding immediate common dilution at the cost of recurring dividends. δ/ Strive (ASST) > $BTC holdings: 43,000 > Weekly accumulation: +6,182 $BTC > Diluted shares: 126.79M → 130.04M > Gross $BTC per diluted share: +13.9% Strive increased $BTC holdings by 16.8% while its diluted share count grew only 2.6%. That’s measurable gross per-share accretion, although preferred claims still need to be considered. … The distinction is in how each company improves its capital structure. BitMine reduces common shares. Strategy retires preferred capital. DFDV uses preferred financing to expand its SOL treasury. Strive grows )BTC holdings faster than its diluted share count. Each approach has different financing costs and implications for shareholders. A larger treasury doesn’t automatically mean greater economic exposure per common share.

Over $44M in 30-day revenue for @pumpfun. Solana launchpad competition isn’t even close. Latest DeFiLlama numbers: > @Pumpfun: $44.2M > @LaunchOnSF: $9.22M > @Bonkfun: $2.21M > @GraphiteProto: $1.49M > @Raydium LaunchLab: $846K Pump is doing nearly 4.8x its closest competitor and 20x Bonkfun. Launching tokens is one business. Keeping the trading activity and capturing the fees afterward is where pump has built its advantage.

FHE is becoming a competition over how cheaply encrypted computation can run at scale. @primus_labs recently reported some promising AutoHoG benchmarks for encrypted execution. The numbers: > Transfer latency: ~90ms > Throughput: 210+ transfers/second > Hardware: 8× RTX 5090 GPUs > Workload: Encrypted 64-bit transfers Zama is taking a different approach with FHEVM, providing encrypted Solidity types, coprocessors and access controls for building confidential smart contracts. Primus is working further down the computation stack, where execution speed and hardware efficiency determine how expensive encrypted operations become. For stablecoins, payments and other financial applications, those costs can add up quickly. A confidential application might have excellent developer infrastructure, but if processing every encrypted transaction requires expensive hardware resources, scaling becomes an economic problem. That’s where Primus has an opportunity. If AutoHoG’s reported performance translates into lower execution costs in production, efficient encrypted computation could serve applications across multiple chains. The benchmarks aren’t directly comparable with Zama’s broader FHEVM architecture, but both are addressing different requirements of confidential finance. Zama is building the tools for encrypted applications. Primus is working on making the computation underneath them cheaper. As adoption grows, the economics of running those applications could determine how far confidential finance can scale.

One of crypto’s biggest inefficiencies is having liquidity and applications that can’t interact without additional infrastructure. Ethereum has established DeFi protocols. Solana has its own liquidity and native applications. But developers shouldn’t have to rebuild everything just to access both. That’s the approach @RomeProtocol is taking. Rome runs a real EVM inside Solana, allowing Ethereum-compatible applications and Solana programs to share liquidity and state. > Solidity contracts can run using familiar Ethereum tooling. > Solana-native programs can interact with EVM applications. > Both environments can access the same underlying market state. The compelling part isn’t simply supporting two developer ecosystems. It’s reducing the separation between applications that were originally built for different execution environments. One state, different ways to access it.

Privacy tokens are getting repriced a lot faster than the underlying usage. Over the past 90 days: > $ZEC: +140% > $XMR: +67% > @Zcash shielded holdings: +13% > Shielded supply: 26.04% → 29.15% Pretty significant divergence. $ZEC has rallied more than 10x the percentage growth in its shielded holdings, although the two metrics aren’t directly comparable. Doesn’t mean the rally is unjustified. Just means the market is pricing in considerably more than what’s currently visible in the adoption data. There are two distinctions being overlooked here. 1. Shielded holdings aren’t the same as private transactions. You can move $ZEC into a shielded pool and leave it there indefinitely. The balance increases, but recurring payment activity doesn’t. Dollar-denominated balances can be even more misleading. If $ZEC doubles while the number of coins held privately stays constant, the reported USD value doubles without additional deposits. @Monero has a different measurement problem. $XMR hides transaction amounts and counterparties by default, so observing more transactions doesn’t necessarily establish how many new users are participating. 2. Privacy adoption doesn’t always translate into native-token demand. Starknet introduces another variable because its privacy infrastructure can support assets beyond $STRK. Someone using confidential stablecoin transfers doesn’t necessarily need substantial exposure to the native token. This is where the investment thesis gets complicated. The market could be pricing in genuine growth in privacy demand, anticipating future adoption, or simply responding to momentum and positioning. All three can happen simultaneously. And the $ZEC data does show more supply moving into shielded holdings, so there’s measurable growth underneath the rally. But 144% price appreciation against 13% growth in shielded holdings leaves a substantial gap between the token’s performance and that particular adoption metric. That gap can narrow through stronger usage, a price correction, or some combination of both. My Take: Wouldn’t short privacy simply because prices have outrun adoption metrics. Momentum can remain strong even when fundamentals lag. But I wouldn’t confuse a vertical price chart with proof of mass adoption either. The trade can work long before the investment thesis is proven. Knowing which one you’re buying is the important part.

One of the biggest lending markets isn’t necessarily where the lending activity is most profitable. @eulerfinance is a pretty good example of that right now. The protocol has around $291M in TVL, but the distribution of capital looks nothing like the distribution of fees. Here’s the breakdown: Monad > $208M TVL > $482.8K in 30-day fees > 71.4% of total TVL Ethereum > $52.7M TVL > $488.9K in 30-day fees Avalanche > $4.48M TVL > $635.1K in 30-day fees > Just 1.5% of TVL, but roughly 37% of fees Avalanche is generating more lending fees than Monad despite holding around 46x less capital. That’s a pretty significant difference, and it tells you something about how lending protocols actually work. Deposits alone don’t generate lending income. Borrowers do. A market with hundreds of millions in collateral can generate relatively little interest if borrowing demand is weak. Meanwhile, a much smaller market with expensive or heavily utilized debt can produce considerably more fees. Euler’s earlier market-level research helps explain the difference. Monad had accumulated substantial collateral without corresponding borrowing activity, while some of the smaller deployments were generating much higher interest from concentrated lending positions. There’s another distinction worth making here: gross lending fees aren’t the same as revenue retained by Euler. Interest can flow to lenders and curators, while the protocol itself captures only a portion, depending on its fee configuration. So I wouldn’t interpret Avalanche’s fee lead as proof that it’s the strongest lending market. A small number of expensive borrowing positions can inflate that figure, and those rates may not persist. But the broader comparison is revealing. Monad holds most of Euler’s capital. Avalanche generates the largest share of its reported lending fees. In DeFi lending, where the deposits sit and where the economic activity happens can be two very different things.

A protocol can spend millions buying its own token and still leave holders worse off. Because buying tokens is only half the equation. The other half is how much supply enters circulation, whether purchases are burned, and what investors paid for the token. A tracker using DefiLlama data estimates $1.51B in buybacks across 23 programs over the year through October 15, 2026. The latest 30 days accounted for $155M. Three protocols made up roughly 72%: > @HyperliquidX: $71.4M (46.1%) > @Pumpfun: $30.4M (19.6%) > @Uniswap: $9.76M (6.3%) > Other programs: $43.44M (28%) But who actually benefits? … α/ Spending Isn’t Value Creation A $100M buyback against a $1B market cap represents 10% of valuation. A $150M buyback against $10B is only 1.5%. The bigger spender isn’t necessarily offering better economics. And token treatment matters. Purchased tokens might be burned permanently, held in a treasury, or redistributed later. Those aren’t equivalent outcomes. … β/ Three Programs, Three Different Risks Hyperliquid: Trading fees finance $HYPE purchases through its Assistance Fund. A September SEC-filed document described 99% of protocol fees going to the fund and 46.7M $HYPE acquired and permanently removed by August 23. But future emissions and contributor allocations complicate the supply picture. An unlock isn’t automatically a sale, yet token destruction doesn’t guarantee shrinking circulating supply either. Pumpfun: Its dashboard targets 50% of revenue for $PUMP buybacks and burns. The tracker estimates a 17.8% annualized buyback yield. That’s not a 17.8% investor return. If speculative trading cools, fee revenue and buybacks can cool with it. Its dashboard also flags reporting limitations after custom-pairs changes. Uniswap: $UNI’s UNIfication framework connects eligible fees to token burns through contracts. The proposal also included a 100M $UNI treasury burn and a proposed 20M $UNI annual growth budget. Burning treasury tokens isn’t necessarily the same as buying circulating tokens. And trading volume isn’t automatically revenue captured by $UNI. … γ/ What Net Absorption Reveals Imagine three hypothetical tokens: > A: $100M bought; $150M of newly liquid supply. Net: -$50M. > B: $80M bought; $20M newly liquid. Net: +$60M. > C: $50M bought; no new liquid supply. Net: +$50M. Relative to market caps of $1B, $2B and $500M respectively, those simplified net absorption rates are -5%, +3% and +10%. C has the smallest buyback, but the strongest relative absorption. These are illustrations, not actual protocol measurements. They assume purchases permanently remove liquid tokens and all new supply becomes liquid. Real unlocks aren’t automatic sales, and treasury-held tokens can return to circulation. … δ/ Buyback Yield Isn’t Earnings Yield A governance token doesn’t automatically confer shareholder rights or a claim on protocol profits. And fee-funded purchases can be cyclical. In bull markets, trading rises, revenue grows and buybacks accelerate. In downturns, buying power can contract just as holders want to sell. Annualizing a strong month doesn’t make its revenue durable. … ε/ My Take Buybacks can connect protocol success to token value. But dollars spent are a poor standalone leaderboard. The better comparison combines sustainable fee revenue, permanent burns, actual circulating-supply changes and valuation. The winner isn’t necessarily the protocol buying the most tokens. It’s the one creating the most durable value per dollar investors pay.