Deep Dive
Bitcoin bounces off bear market floor amid whale stacking
Bitcoin bounced off its 200-day exponential moving average this week and is now holding above it, a level considered structural bear market support. The fascinating part is this support is holding strong without treasury purchases or ETF inflows—meaning institutional demand hasn't returned yet. Using the power law floor model developed by Joe Consorti, each market cycle establishes higher lows. The 2022 cycle bottomed at $16K, the model projects $60K as the 2026 floor, $188K by 2030, and $582K by 2035. That's a 10x in a decade, though these are floor prices, not peaks—Bitcoin could spike to $1 million and fall back to $580K. More intriguingly, long-term holders stacked 371,000 Bitcoin in just 30 days despite retail, treasury, and ETF demand being absent. Axel from CryptoQuant documented this, bringing long-term holder supply to 16.34 million BTC, leaving only 3.5 million available for new buyers. If big money returns, supply is critically tight.
China's AI models overtake US leaders while OpenAI faces structural collapse
China's DeepSeek Kimi-3 now ranks first on Frontend Code Arena, dethroning Claude Fable 5, GPT-5, and every other Western model by a significant margin. David Sachs underscored the moment: America's politicians and bureaucrats are banning data centers and pre-approving frontier models while China moves fast. The irony is China was forced into efficiency—lacking compute resources, they built exceptionally effective models with trillions of parameters anyway. Now nearly 60% of AI models used internally by US companies are Chinese, because capital flows to what's cheapest and fastest regardless of origin. Meanwhile, OpenAI and Anthropic have a pricing problem. Comparing raw pricing on AI compute, Anthropic charges $56 per 1 million tokens, OpenAI charges $26, but Chinese models run at 50 cents. Meta, Google, and XAI cluster between $1 and $2. At these margins, neither Anthropic nor OpenAI can survive profitably—both are trillion-dollar valuations with no path to profitability as cheaper alternatives proliferate. OpenAI is a dead man walking, per the host's assessment. Anthropic followed suit.
Solana monopolizes tokenized assets while Ethereum struggles with scale
Solana is processing 96% of all tokenized equities spot trading volume, with $3.5 billion handled in June 2026 alone while every other blockchain fought over $141 million combined. This makes Solana a quasi-monopoly in real-world asset tokenization, an arena the host believes could eventually dwarf stablecoins. Tom Lee's recent prediction that ETH will hit $250,000—a 50x from $1,800 that would value it at $30 trillion, larger than all of the Mag 7 combined—assumes ETH is the foundation layer for agentic AI. But Ethereum only processes 15 transactions per second while Solana handles 100x more volume and has 10x more users. If AI agents choose the fastest, cheapest chain with best finality (which the host argues is the product in an AGI era), Solana wins. Even if Tom Lee is only 10% correct and ETH reaches $25K instead of $250K, Solana would theoretically be valued 10x higher given its superior throughput and adoption. The technology gap makes this a matter of time.
Tesla FSD scaling, SpaceX Pentagon deal, and the startup turnaround
Tesla's Full Self-Driving fleet has logged 12 billion miles—equivalent to driving to the moon and back 51,000 times. The fleet is adding 30 to 35 million miles daily with fully autonomous vehicles, representing real-world AI scaling at commercial scale. Cyber cabs and robo taxis are imminent, with reports suggesting deployment within weeks. When the fleet awakens, self-driven miles will explode and Tesla transitions from being a pick-and-shovel play to being the application of baby AGI itself. Meanwhile SpaceX news sent the stock from $122 to $130 intraday: the Pentagon is rumored to be building its own AI powered by SpaceX compute. If true, SpaceX would add a fourth major customer beyond Reflection AI, Google, and Anthropic—who each pay $28-29 billion yearly for compute. Adding the Pentagon at another $10-20 billion annually would make SpaceX a top-tier hyperscaler. Starship hit a snag when 60 seconds before launch on July 19, two Raptor engines needed replacement, but the team successfully fueled and drained both booster and ship without incident, showing operational maturity. Launch rescheduled for July 20. Top VC Mike Moritz noted that Tesla and SpaceX are unprecedented—no precedent exists for valuing them, and nobody can replicate their vertical integration.
Stripe's PayPal bid and the blockchain future of payments
Stripe is reportedly bidding $53 billion to acquire PayPal, whose market cap sits around $50 billion after plunging from north of $320 per share to $50—zombie levels on the ATR. The irony is stark: the original internet payment company founded a quarter century ago is now acquisition target for a pre-IPO fintech startup. If Stripe wins, they'd lock up global payments. Stripe is also pivoting toward blockchain infrastructure, betting that all future transactions eventually settle on distributed ledgers. For PayPal shareholders, the $53 billion bid is bottom fishing—the board would never accept it. But the signal matters: legacy payment rails are being displaced.