20VC: Anthropic Wipes Billions Off Markets | Citrini Research: The Ultimate Breakdown: Agents, "Ghost GDP", Consumer Spend etc. | Figma Earnings Beat & Four Public Stocks to Buy | Jack Altman Joins Benchmark

What It Covers

Harry Stebbings, Rory O'Driscoll, and Jason Lemkin analyze Anthropic's security release wiping $20B from cybersecurity stocks, Figma's 40% revenue growth quarter, the Citrini Research "Ghost GDP" macro thesis, OpenAI's $665B spending plan, and Jack Altman joining Benchmark — debating which public stocks to buy amid accelerating AI disruption.

Key Questions Answered

• •Valuation risk at perfection pricing: CrowdStrike traded at 16x revenues even after a post-Anthropic correction — still not cheap. When stocks price in zero tail risk, any narrative disruption triggers outsized selloffs regardless of business quality. Investors should prefer baskets of 20 B2B software stocks averaging 3x revenues and 8x EBITDA over individual high-multiple names, where idiosyncratic risk is harder to assess.
• •Momentum over value in current market: Five public stocks are up over the past twelve months: Palantir, Figma, MongoDB, Cloudflare, and Shopify. In a high-uncertainty AI environment, momentum has consistently outperformed value investing both in public markets and venture. Rather than bargain-hunting beaten-down names, follow price action as a proxy for which companies are executing through disruption.
• •Atlassian as the clearest value dislocation: Atlassian is down 74% over twelve months while simultaneously accelerating revenue growth from 20% to 23% at $6.3B ARR. No other large-cap software company combines that level of price decline with revenue acceleration. Increasing enterprise multi-year contracts add durability. For value-oriented investors, this represents the widest gap between price action and fundamental trajectory in the sector.
• •Ghost GDP concentrates wealth, shrinks consumer base: Jason Lemkin's team went from 12 people to 2 while maintaining 8-figure revenue — a real-world example of AI productivity gains not dispersing to workers. Fewer employed workers means fewer consumers buying goods and services. The macro risk is not GDP collapse but a structural softening of consumer spending concentrated in tech-heavy cities, mirroring Japan's 1990s productivity-without-distribution problem.
• •Agents require custom deployment — incumbents are losing the window: Every enterprise AI agent currently requires custom training, data cleansing, and forward-deployed technical staff. Existing B2B software companies lack the workforce to execute this at scale. Startups with hyper-niche focus are winning because they handle one vertical's agent end-to-end. Broad horizontal platforms like Monday.com or HubSpot face the hardest path because their 100-plus vertical use cases make standardized agent deployment nearly impossible.
• •PE-backed SaaS faces forced restructuring: Highly leveraged private equity-owned SaaS companies growing at single digits with debt at 6x EBITDA cannot grow their way out. Expect dramatic headcount cuts — potentially 50% reductions at some firms — and consolidation of 15-20 unicorns into Frankenstein roll-ups trading at 1-2x revenue. These merged entities will attempt IPOs around 2027, but represent distressed outcomes rather than genuine AI transformation stories.