While AI will significantly boost productivity and reduce costs for many cognitive and physical tasks, the claim that wealth scarcity will collapse and labor costs will approach zero overlooks both history and current evidence. Scarcity does not vanish — it transforms. Frontier model training and large-scale deployment remain extremely energy- and compute-intensive, and humanoid robotics face major scalability, maintenance, and safety challenges far beyond “pennies per token.” Multiple studies (including from Brookings and BIS) link AI investment and adoption to rising income inequality, with gains disproportionately accruing to capital owners and high-skilled workers rather than broadly eliminating distribution issues. The idea of Universal High Income delivered purely through deflation is optimistic; IP protection, network effects, and potential monopolistic pricing will sustain premiums on advanced and customized goods/services. Energy and compute are important scarce resources, but they are not the only decisive ones — talent ecosystems, data quality, institutional frameworks, and geopolitical supply chains matter equally. Europe’s risk-based regulatory approach (such as the AI Act) aims to build trustworthy AI that could become a global standard and long-term competitive advantage, rather than a fatal disadvantage. US, Korean, Taiwanese, and Japanese strengths are real, yet each carries its own risks. AI is a powerful tool, but economic outcomes have always depended on policy, institutions, and choices — not automatic abundance. Distribution questions remain relevant. Open to data-driven discussion.
I've been looking closely at $SPCX (Space Exploration Technologies, SpaceX) and the recent price action deserves a closer look. Since the Nasdaq debut in mid-June the stock opened around $150, briefly pushed into the low-to-mid $220s, but has since given back a lot of ground — down roughly 35% from those post-IPO highs and currently sitting near $145–150. That's a sharp correction in a very short window. A few concrete factors stand out. At the IPO the valuation was aggressive — reports put it above 90x trailing revenue, with some brief peaks approaching 140x. 2025 revenue came in around $18.7 billion (mostly Starlink), yet the company posted a ~$4.9 billion net loss for the year. Q1 2026 showed another heavy deficit of roughly $4.3 billion, driven by continued heavy spending on Starship development and other infrastructure. When a company is still deeply unprofitable and trading at those multiples, any hint of slower growth or execution risk gets punished quickly. The near-term catalyst everyone is watching is the first major post-IPO share unlock window after the August 6 earnings. With such a large float coming online, selling pressure is a real near-term headwind even if the underlying business is progressing. Broader growth-stock and semis sentiment has also been choppy, which hasn't helped. That said, the core infrastructure story hasn't changed overnight. Starlink continues to scale subscribers and revenue, while Starship remains the biggest lever for eventually collapsing launch costs and increasing flight cadence — both classic supply constraints in the space sector. If those programs hit their milestones, the long-term optionality in orbital access and global connectivity is still substantial. Analyst targets remain wide (average around $239, with some as high as $800 and lows in the $115 range), reflecting exactly this tension between near-term losses/dilution and multi-year potential. Right now the market is clearly pricing in near-term pain and skepticism around the path to profitability. The question is whether the upcoming earnings and unlock dynamics will provide enough visibility to stabilize sentiment, or whether further downside tests the conviction in the Starship/Starlink thesis. I'm watching the August earnings and the cadence of Starship test flights most closely over the next 4–6 weeks.
The market thinks we're in an AI boom. I think we're in a liquidity drain. Big Tech is spending hundreds of billions. $SPCX is raising tens of billions. Rates remain elevated. Everyone is fighting for the same pool of capital. Eventually, someone gets starved. That's usually when the real correction begins.
$SPX breaks below 7,250 and I think the market finally gets the correction most investors have been ignoring. Not because the economy suddenly collapses. But because valuations have become dependent on perfection: • AI spending keeps accelerating • Earnings keep surprising higher • Liquidity remains abundant • Geopolitical risks stay contained Markets don't need bad news to fall.They only need expectations to become too high.A deeper pullback wouldn't scare me. It would create the best opportunity in months to accumulate high-quality companies at more reasonable valuations. Corrections are where future returns are born.
Vision AI priced in. Physical AI is not.We already saw the “eyes” of AI get re-rated —computer vision, multimodal models, inference at scale. But the next phase is structurally different. Physical AI is not a software upgrade. It’s a capital cycle shift.We are moving from: 1. Perception systems (Vision AI) 2. Action systems (robotics, autonomy, industrial execution) The stack is forming: • Vision → Cognition → Action (VLA models) • Simulation layers (Omniverse / Cosmos) • Edge compute (Jetson / robotics silicon) • Industrial deployment (FANUC / ABB / KUKA ecosystem)This is not SaaS margin expansion. This is CAPEX re-acceleration across global industry. What matters now: Energy + compute constraints Robotics unit economics Supply chain bottlenecks (3nm, HBM, LPDDR5X) Markets are still pricing this like a narrative. But the leading indicators suggest a structural cycle. AI moving from screens → factories → physical world.The gap between Vision AI and Physical AI is where the next multi-year alpha will be formed.
