邵艾伦 × 孙宇晨

核心不是押中单一赛道，而是把健康、工作与决策沉淀为可迁移数据，用多模型交叉分析并持续调参。机会集中在数据密集且处理不足的领域，AI基础设施、生物健康金融、全球支付与Agent结算可受益于更低的搜索、分析和交易成本，但波场规模、合规与采用仍待验证。

Doug Colkitt

AI’s progress from contest mathematics to Erdős problems, a claimed Navier–Stokes result, and 400 open problems has widened uncertainty around cryptographic hardness, without showing an imminent break.Doug sees signature forgery as an internet-wide threat, but favors Y2K-style migration over bunker mode today: years could permit hash-signature upgrades, while weeks might not, and formal verification may offer nearer-term security gains.

swyx × Brandon × Liam Fedus × Ekin Dogus Cubuk

Periodic Labs is building AI, simulations, and high-throughput physical labs as a closed loop, with timestamped failures, instrument data, and campaign lineage potentially differentiating its scientific intelligence.Automated characterization—not humanoid robotics—is the near-term bottleneck, while secure on-site deployments for semiconductor companies could evolve from Copilot-style assistance toward outcome pricing; whether throughput can compress discovery without better physical representations remains the key risk.

Raghu Raghuram × Matt Garman

AWS is underwriting durable AI demand with $220 billion of 2026 capex while reserving scarce accelerators for startups as well as frontier labs.Positive-ROI production workloads and single-digit maximum customer concentration temper demand risk, while AWS plans to buy 2 million NVIDIA GPUs over the next couple of years.Agentic workloads may strengthen AWS’s architectural edge, but power, memory, chips and enterprise evaluation remain constraints to monitor.

Fred Wilson × Rebecca Kaden × Michael Mignano × Molly O'Shea

USV’s $900 million fund targets 30-40 companies with $10 million-$30 million Series A checks as larger rounds and scaling costs reshape venture.Its AI thesis spans applications, infrastructure, physical AI, and energy, while usage retention, compute costs, inference subsidies, distribution, interoperability, trust, and leadership concentration remain key tests.

Logan Jastremski × Ben Fielding

Frontier AI’s scaling race may not repay its capital costs as distillation and open-weight releases erode model moats, shifting durable value toward products with large, paying, retained customer bases.Gensyn’s pivot from decentralized training to verifiable execution and forecasting reflects a changed bottleneck—from compute to trustworthy, time-stamped information—while Deli’s evidence markets could reward predictive information, subject to unresolved settlement, incentive, and product-demand risks.

José Macedo × Rohan Pujara

Pujara avoids consensus-heavy seed categories where inflated prices, copycats, and weak margins destroy early economics, underwriting distinctive founders instead.Valhalla’s $80 million–$100 million fund makes only two to four new investments annually and can concentrate 20%–30% behind a clear fund-returner.K2 Space’s more than $1 billion of signed business and $6.8 billion valuation illustrate the playbook, while the roughly 30-check diversification case remains unresolved.

Nathan Labenz × Thomas Sohmers × swyx × Evan Miyazono × Edward Hu

Frontier insiders increasingly treat powerful AI as operational reality, while a lab leader accepted a possible 10^27 FLOPs ceiling, making compute and synthetic-data governance live constraints.Memory costs are tightening: Positron cites a 4.5-times quote increase in one year and one week, yet over five-times capability gains per fixed memory.Agentic coding threatens generic CRUD SaaS while expanding bespoke software, but coordination, evaluation and reliability remain unresolved risks.

Thread Guy × Rhino

Rhino treats longevity—not a spectacular P&L—as the only durable proof of trading skill, and fades dominant narratives only when rhetoric and sentiment reach extremes.Shallow event-driven wicks can create tight invalidations against 20-30% upside, but with roughly 95% of liquid net worth in TradFi and 2026 roughly matching the Nasdaq, his next test is capturing opportunities without sacrificing survival.

李严冰 × Ion Stoica × Ed H. Chi × Bill Jia × 乔琳 × Li Fan × Dawn Song × 贾扬清 × 田渊栋 × 葛小川 × Connie Chan × 曹卿云 Qingyun Cao

基础模型领先或仅维持数月，开源的核心价值转向掌握升级、流量与专有数据控制权。Agent已在ExploitGym中发现未知漏洞并攻击基础设施，企业采购也从token用量转向可验证的商业ROI。云厂商接近1万亿美元的资本开支将推动算力从训练转向推理，电力、内存与软硬件协同成为关键瓶颈和机会。