Deep Dive
Zombie Chains and the Cost of Bad Picks
The video opens with a hard lesson on avoiding losers. InvestAnswers walks through a graveyard of tokens that peaked in 2018 and never recovered: XRP fell from $4 to $1, Cardano is a "complete zombie chain," Litecoin crashed from $360 to $45, and Stellar slid from 86 cents to 16 cents. The core principle is that 99.8% of crypto will go to zero, and 62% of the top 100 tokens are dead within five years. Grayscale just withdrew ETF applications for three of these zombies β Cardano, Polkadot, and Hedera β which the host calls a bizarre three-year delay. The takeaway is ruthless: tokenomics matter, usage metrics matter, hope is not a strategy. If a chain hasn't found product-market fit by now, it's unlikely to catch up. Michael Saylor selling Bitcoin to prop up MicroStrategy's stock is also flagged as a poor move.
Bitcoin Floor Holds Despite Summer Doldrums
Bitcoin has been hovering at $64K since February β roughly six months of stability despite turmoil (COL card issues, ETF outflows, general chaos). The host argues this is a bullish signal because it suggests a real floor. Summer volumes are light, which explains why August is only up 1.5% so far. Last week's ETF inflows hit $900 million, the strongest week since April 2026, which suggests institutional confidence is returning. The power law decay oscillator β a four-year cycle chart β indicates the bottom is in, with less than 5% odds of dropping to $50K or $40K. Every $1 billion into Bitcoin historically drives price up 3%, so if flows continue, upside is locked in. All buyers from the last 90 days are at breakeven, meaning new money isn't bleeding. Alt season index hit 55, the strongest altcoin bounce since October 2025, and Solana, Tron, and others are outperforming Bitcoin for the week. Matt Hougan from Bitwise expects Bitcoin north of $100K by year-end, a forecast based on institutional and sovereign flows he's tracking.
Solana Dominates On-Chain but Market Doesn't Price It
This is the section where data collides with valuation. Solana does 65% of all on-chain activity across crypto but holds only 2% of total market cap β a massive mispricing. For 26 consecutive months, Solana has led in app revenue: 43% of all application revenue goes to Solana apps versus 15% for Ethereum and 10% each for Hyperliquid and Binance. Yet Ethereum trades at 40x higher valuation despite lower transaction volume and higher fees. The host has been saying this for five years, and Vibu's usage-based valuation model quantifies the disconnect: if L1s were priced by usage share, Solana would jump from $75 to $590, Avalanche from $6 to $11, and Ethereum would crater from $1,874 to $43. Cardano would fall from 18 cents to $0.001, confirming it's worthless on metrics. Binance would drop from $692 to a lower price, and XRP would fall from $1 to 9 cents. The industry simply doesn't know how to value utility.
AI Is Not a Bubble β Trillions Are Flowing In
The host shifts tone hard here: AI is the most important technology since electricity, and the evidence is in the capital. Yesterday, Larry Fink (BlackRock), Apollo, Blackstone, KKR, and Nvidia all met and publicly urged raising trillions for AI infrastructure β specifically to compete globally (coded language for beating China). This isn't retail enthusiasm; it's the planet's biggest capital allocators committing. Stock fear and greed is at 64 (very greedy), but the weekly performance is mixed: Nvidia +3%, Microsoft +4.5%, Meta +4%, Tesla +2.4%, but Amazon -2% and Google -6%. This is typical August chop. Nvidia is now building its own frontier model (Neumotron, 1 trillion parameters) to run AI agents from its AI factory, which will price pressure competitors like OpenAI and Anthropic. Riot Blockchain just got a huge Anthropic investment for compute, and the stock popped 25% β miners are pivoting to AI. The stock market bull run, averaging 5.6 years historically but now 8-12 years since 1987, is only three years in. If the pattern holds, 2-5 more years of upside remain.
Memory Supercycle Is the New Gold
Here's where infrastructure gets concrete. The GB300 GPU costs $4 million; the next-gen Vera Rubin costs nearly $8 million. But the critical chart is memory's share of total cost: it's jumped from 10% to 25%, a 435% increase in memory expense. A kilogram of DRAM now costs $94,000 β this is the commodity that matters. Demand is infinite because data centers are being built at scale, especially in red states (Southeast, Southwest) where electricity is cheap and regulation minimal. Tesla and SpaceX are allegedly committing to buying 30% of Vera Rubin supply. The host notes that blue-state data center buildouts are minimal due to cost and regulation, while AI talent is still 49.3% concentrated in California, with only 3.7% in Texas and 9.1% in Washington β but that will shift as infrastructure moves. SpaceX fell to $130.40 last week but had a mind-blowing earnings call; retail turned net sellers for the first time since the June IPO, but unlocks will continue through October. Palantir is growing revenue at 39.9% CAGR and is defying Michael Berry's prediction it'll hit $1 β at $170s it's frothy but justified by growth.