The Twenty Minute VCPredictions for 2026: Top Buy & Biggest Short | Why Salesforce Could Win & NVIDIA’s Challenges
At a glance
WHAT IT’S REALLY ABOUT
Venture leaders recap 2025, debate AI, and forecast 2026 winners
- The panel crowns Anthropic (and CEO Dario Amodei) as the defining force behind 2025’s “vibe coding” and agentic-product inflection, while also praising operational leaders like Gwynne Shotwell (SpaceX) and execution-driven founders like Vlad Tenev (Robinhood).
- They argue fund performance should be judged by liquidity and exits, naming Index for major wins (e.g., Figma, Wiz) while also highlighting “aesthetic”/brand-driven momentum from Neo, plus standout stage-specific performance from Hummingbird, Benchmark, Thrive, and Founders Fund.
- Breakout-company discussion centers on AI-native products with sharp distribution and clear ROI, such as OpenEvidence’s doctor-focused research assistant and ElevenLabs’ rapid ARR growth enabled by dramatic cost/time compression for end users.
- For 2026 public-market calls, Rory favors buying Google and shorting NVIDIA on valuation and CapEx-cycle asymmetry, while Jason rejects near-term NVIDIA shorts and frames Salesforce as a contrarian “AI attach-rate” rebound if Agentforce delivers measurable, paid outcomes.
- They predict a wave of late-2026 IPOs led by SpaceX and Anthropic (with Jason adding Canva and Databricks), and close by warning that any unemployment uptick—AI-caused or not—could trigger a severe political and social backlash against AI given executives’ public “job loss” rhetoric.
IDEAS WORTH REMEMBERING
5 ideasClaude’s capability jump is framed as the hidden enabler of the 2025 agentic-app boom.
They claim tools like Cursor/Replit/Lovable only “really worked” after Claude 3.5/3.7+, shifting AI in products from demos to dependable workflows and changing how software gets built and shipped.
In venture, exits and liquidity are still the “coin of the realm,” despite new narratives.
Rory pushes objective scoring by realized/announced outcomes (e.g., Index with Figma/Wiz and exposure to Revolut), while also acknowledging multiple strategies can work if executed well.
AI breakout winners pair a narrow, urgent user job-to-be-done with monetization that naturally follows usage.
OpenEvidence succeeded by focusing on doctors (high-intent queries) and matching advertiser demand (pharma), while ElevenLabs exemplifies user-led adoption driven by massive cost/time savings.
“AI-influenced revenue” is not the metric—net-new paid attach and pricing power are.
They criticize performative AI accounting (explicitly calling out Adobe) and argue winners will be those who can charge more for AI via clear value delivery, citing Notion’s successful paid AI upsell as a rare example.
Traditional SaaS laggards may outperform if they convert AI into a paid, high-ACV agent motion.
Jason’s Salesforce thesis is that customers will pay meaningful incremental dollars for GTM agents that remove headcount or unlock neglected leads—if deployment becomes fast and reliable—creating upside with less valuation risk than high-flyers.
WORDS WORTH SAVING
5 quotesIf unemployment ratches up even two or three points, regardless of the reason, could just be the business cycle, I think you'll, you'll see a tech lash that makes what we're dealing with now an understatement.
— Rory O’Driscoll
Society will be terrified of AI.
— Jason Lemkin
If the robots are gonna take over society, I want to be sure that I own the robots.
— Rory O’Driscoll
There is no ceiling on venture, which changes all the math and calculations.
— Jason Lemkin
This is not software. Like, you can't launch Base44. This-- There's no Base44 of GPUs. It just doesn't exist.
— Jason Lemkin
High quality AI-generated summary created from speaker-labeled transcript.
