Tushar Jain 2.6 17 ideas

"Whoever owns the customer relationship... and whoever owns the liquidity [wins]." AI agents and users will seek the path of least resistance. The value in crypto infrastructure won't accrue to the "pipes" (blockchains) but to the regulated endpoints that own the customer (Robinhood) and the deep liquidity pools (Coinbase). They are "natural monopolies." LONG Regulated Crypto/Fintech Infrastructure. Fee compression; decentralized exchanges (DEXs) like Hyperliquid stealing market share.

"Markets hate uncertainty and with the two cap table structure right now markets perceive some uncertainty on value capture." Helium is undervalued because of the complex split between the company (Nova Labs) and the token (HNT). As the industry consolidates these structures (folding labs into protocols to align value), the uncertainty discount will vanish, repricing the asset higher to match its utility. LONG HNT as a restructuring play. Regulatory "hangover" persists; failure to successfully merge/align the entities.

"AI is not making any more land." In a future where AI creates an abundance of digital goods and labor (deflation), capital will flee to assets that cannot be printed or automated. Physical land becomes the ultimate store of value and scarcity. LONG Real Estate Investment Trusts (REITs). Commercial real estate collapse due to remote work/AI displacing office jobs.

"Applications capture 60% of the value but have a minority of the market cap... that to me looks like an opportunity." As professional investors enter the space, they will value assets based on discounted cash flows (DCF). DeFi protocols (Aave, Maker) generate sustainable revenue ($10s of millions), whereas Layer 1 blockchains are priced on speculative monetary premiums. Capital will rotate into these cash-flowing "equities of crypto." LONG DeFi "Blue Chips" with proven revenue models. Regulatory hurdles remain; "fee switches" (returning capital to token holders) could face legal challenges.

"I don't think stable coin transfers deliver almost any value to a blockchain... they have to be essentially free." The market overvalues L1 blockchains based on the metric of "transfer volume." If transaction fees race to zero (to support AI agents and payments), L1s lose their primary revenue driver. The "fat protocol" thesis is dead; value accrues to the app, not the base layer. AVOID Layer 1s relative to Applications (The short leg of the pair trade). L1s could develop new monetization layers or maintain a monetary premium as "digital gold/oil."

"I think nuclear has a ton of opportunity... not just because of demand side because of AI but also I think the innovations to build safe nuclear reactors will be faster." AI drives a dual-tailwind for nuclear: 1) Massive energy demand from data centers, and 2) AI accelerates the material science required to build safer, cheaper reactors. This creates a structural supercycle for uranium and nuclear infrastructure. LONG Nuclear producers and miners. Regulatory gridlock; public perception of nuclear safety; long lead times for reactor construction.

"I think we're going to see ASICs everywhere because chip design has been one of the most human capital constrained things out there." Chip design software (EDA) has been limited by the number of human engineers. AI removes this bottleneck, allowing companies to design custom chips (ASICs) for every specific workload. This leads to an explosion in volume for the software providers that power chip design. LONG EDA Software monopolies (Cadence/Synopsys). AI models eventually writing their own chip design code without legacy EDA tools (disintermediation).

Solana is implementing "Application Controlled Execution" (ACE) and "AlpenGlow" (consensus upgrade), while maintaining credible neutrality. ACE allows apps to monetize transaction ordering (e.g., capturing MEV for the protocol rather than leaking it), creating a new revenue stream. Combined with neutrality, this makes SOL the superior substrate for "Internet Capital Markets" over corporate chains. LONG. Technical upgrades (Fire Dancer, ACE) combined with the "Internet Capital Markets" thesis position it to capture institutional issuance. Inflation rate remains high; network stability issues (though improving).

There is a cultural shift toward "Entertainment Finance"—people trading for fun (like sports betting) rather than just profit. This is a "megatrend" driven by human nature and the zeitgeist. Assets that cater to this (memecoins, betting markets) are servicing a high-growth consumer demand sector. LONG. Do not dismiss these assets as "useless"; they are consumer products. Regulatory crackdowns on gambling-adjacent crypto products.