Ep. 035 - Tech DD’s, Performance Projections, Supply Chain, Investment Thesis (Consulting)
SemiAnalysis Consulting is turning bespoke AI-infrastructure diligence into scalable products, including a token-supply simulator linking chip deployment, workload allocation, throughput, and token pricing.NeoCloud underwriting centers on contractual survivability: 99.9% availability expectations, sub-90% performance, and mismatched 15–20-year leases versus five-year GPU contracts can threaten financing and trigger remedies, while capacity timelines determine whether clients build or lease.
183: 与Henry的「AI季报26Q3」:Muse引爆个人助理、Astra进入机器人、OpenAI收入猛增
个人助理正从聊天入口转向主动执行的操作层:Dots月费100/200/500美元,Muse靠免费版本和社交平台分发。GPT-6 Astra在RoboDojo中平均成功率超22%,领先GPT-5.5与π0.5,但算力短缺和安全事故仍是风险。
AI, Infrastructure, and the Next Investment Cycle
David GeorgeSarah WangAlex ImmermanSantiago Rodriguez
AI’s rally is being driven by earnings rather than multiple expansion, with the S&P 500 below 20 times earnings while growing about 15%.Capex is projected to rise from $416 billion in 2025 to about $780 billion in 2026 and exceed $1 trillion annually from 2027; enterprise adoption remains concentrated in pilots, with inference-cost declines and workflow deployment as catalysts.
E253|谁在给大模型出题、卖题、判卷?聊聊AI数据行业的野蛮生长
AI数据估值急涨:AfterQuery今年4月A轮估值3亿美元,9月新一轮已推至32亿美元;Meta入股Scale AI给出超290亿美元估值。行业从人工标注转向rubric与RL environment,垂直采购和研发能力成为关键护城河,但rubric是否见顶、Benchmark权威、数据造假及数商能否持续杠杆化仍待验证。
The $10 Trillion Token Economy — Alex Atallah, OpenRouter & Anjney Midha, AMP
Stripe–OpenRouter reframes the token economy as a security market, with token flows projected at roughly $5 trillion in five years and $10 trillion over ten.OpenRouter’s moat spans over 10 million developers and roughly 10 trillion tokens daily, while frontier labs struggle to turn checkpoints into revenue.Fraud could drive discrete-task pricing and bring-your-own-inference, while the deal accelerates OpenRouter’s roadmap.
Meta's Muse tops charts; Menlo's AI bubble alarm; Factory at $5B
Meta's Muse, called the first real ChatGPT competitor, pairs autonomous agents with a free consumer LLM and added roughly $100B to Meta's market cap.Jev targets 20% of LLM API spend at one-fifth the price, potentially turning $20B into $4B of revenue.OpenAI's projected $278B net burn and roughly $700B of CapEx keep capital intensity central to the AI trade.
Gabe Stengel - Building Investing Superintelligence - [Invest Like the Best, EP.492]
Patrick O'ShaughnessyGabe Stengel
Rogo’s model-era progression—from o1 Pro’s reliable search to Opus 4.5 handling junior finance work—supports a two-year race to redesign investment firms, not wait for better models.Its wedge is private-market dealmaking, where data rooms, DDQs, CRM and LP reporting still lack infrastructure, while high-value ideas can justify extreme token spend.The watchpoints are compaction, standardization, and whether incumbents move aggressively enough.
Building One of the Fastest Growing CPG Companies in History | Peter Rahal of David Protein & RXBAR
David has reached two years with Rahal citing this year's revenue as “over 400 million” and “300 million” across bars, frozen food, ready-to-drink beverages, and candy.Epogee vertical integration secured a patented modified triglyceride after David represented about 90% of sales and supply, but contracts, litigation, antitrust claims, and 300–400% growth make supply execution the key variable as the company expands toward Nestlé-scale ambition.
Frontier Labs Want to Slow Down, OpenAI Delays Its 2026 IPO, Anthropic Flags 5 Bioweapon Cases
Peter DiamandisSalim IsmailDave BlundinDr. Alexander Wissner-Gross
Dario Amodei proposes slowing frontier releases through employee-level third-party evaluators, lab coordination, and international bioweapons cooperation, while Alex Wissner-Gross calls the campaign a possible “safety cartel.”Anthropic reports biological-weapons cases and says it can no longer confidently assure models remain below the dangerous threshold, strengthening the case for mutual API testing while leaving cartel-like coordination unresolved.
88.在新加坡,与祥峰郑俊聪的访谈:硬科技,下一代 Long China 与华人创新的信心史
祥峰中国四年不投DeepSeek、智谱、Kimi等大模型,转向AI for science与物理AI,新基金至少50%投具身智能。宇树以成本降一个梯级打开应用场景,量产一致性成为壁垒;中国供应链按周迭代与香港、内地上市,重塑Long China的创新与退出逻辑。









