Harry Stebbings. Rory ODriscoll and I have done 37 podcasts together this year (!). We didn’t miss a week. But man, the world changed in those 37 weeks.
My top learnings… and what’s changed:
#1. I Won’t Do a T3D2 Investment Again
I know this can be triggering, and I know it sounds unfair and wrong. But the rate at which the best AI + B2B startups can scale today is utterly unprecedented.
It’s not just the ones you know. Higgsfield is growing at rates I’ve never seen coming up on $200m run-rate in < 12 months. It’s not just Replit and Lovable and Cursor. It’s amazing there are 20+ others you’ve never even heard of getting to $100m ARR in 12-18 months now. Vertical B2B + AI leaders like @owner accelerating at scale.
You just have to find those ones.
#2. Agents Will Change Everything in B2B Software. For Real.
We’re just getting started. It’s not just startups. Leaders like Salesforce will be almost nothing like they were before, when most of their functions are run by agents.
These “AI Agents” in B2B really didn’t work for the first half of 2025. They didn’t. They got better with Claude 3.7, better with Claude 4, and I think, even better with Claude 4.5. @Replit got truly great only with Agent v3 just a few months ago.
Today? My agents talk to other agents, debate features, and run almost infinitely. And that’s just the change in the last 90 days.
The rate of change here is difficult to see, but it is epic. AI Agents really will work in 2026 in B2B in a way they were mostly experiments in 2025.
#3. You Just Have To Find Your AI Tailwinds
The difference between the leaders in B2B this year, and the ones that struggled? The ones that ended the year with stock prices up +40%, vs. -25%? They really tapped into AI Tailwinds.
Not just claiming they added AI to their products.
#4. There Is No B2B or Cloud Without AI
This sounds obvious today, but it wasn’t 100% clear when we started the pod. Cloud is “dead”. “SaaS” is dead. But B2B AI is attracting more budget than ever before. AI Cloud, AI SaaS, AI B2B, AI datacenter. Per Gartner, total enterprise software spending is accelerating faster than ever before. Go get that budget.
#5. Vibe Coding Turned Me Into a Builder. For Real.
I’m now “Replit Fluent.” 9 apps in production. 780,000+ uses already. 6 months ago I couldn’t finish my first project. 10 months ago this would have been impossible. Now I can will almost anything I want to build for the SaaStr community into production — the only limit is time. The learning curve in AI is real, but if you push through, it’s a superpower you never had before.
#6. Founders Should Stay
2,221 CEOs quit last year (up 24%). But — this is not the time or age of professional CEOs. Stay. Founders, stay. The rate of change in the Age of AI is just too fast. Your startup (or large company) will fail or fall far behind if founders aren’t at the helm. Not always, but almost always.
#7. It Only Counts as “AI” If You Deliver Insane Value
Co-pilots were the great failure story of 2025. They didn’t add enough value to justify the pricing. But Gamma, ElevenLabs, Palantir, Replit, Cursor all showed there is massive budget if the ROI is huge and often, near instant.
#9. “Seed Is For Suckers” (When Outcomes Are $100B+)
The math has changed: mega-funds like a]16z ($20B) and Founders Fund ($4.6B) can just write one big check into a winner. Why struggle for 8x over 20 years on a seed fund when you can stuff $1B into Databricks at the right price? Thrive’s strategy: “Buy the best property on every block — like Monopoly.”
#9. Wealth Creation Is Staggering. Staggering.
The numbers are almost hard to believe:
Nvidia: 80% of employees are now millionaires. 50% have a net worth exceeding $25 million. Wealth creation at a scale we’ve never seen in tech history.
Cursor: $29.3 billion valuation. $1B+ ARR. With roughly 300 employees. That’s $100M+ in enterprise value per employee.
Scale AI: Alexandr Wang became the world’s youngest self-made billionaire at 24. Meta paid $14.3 billion for a 49% stake — just to get him and a handful of top AI engineers. Wang’s net worth: $3.6 billion at age 28. This was all almost brand new in 2025, the massive price for talent.
The AI wealth creation machine is producing outcomes we’ve never seen before.
#10. Venture Capital Is Nothing Like It Was 18 Months Ago
The transformation is staggering:
- 2023: AI funding was $55.6 billion (27% of all VC)
- 2024: AI funding hit $100 billion+ (33% of all VC) — up 80% YoY
- 2025: AI funding reached $202.3 billion (50% of all VC) — up another 75%+ YoY
AI now captures half of all global venture funding. But more importantly, half of it is all going into very late stage growth deals. And that % isn’t going up.
Venture $$$ are way up, back to 2021 pace, but into more and more specific, ultra-high growth, often late stage opportunities.
And the pace of valuation acceleration is breathtaking:
- OpenAI: $29B in 2023 → $500B today, $800B soon
- Anthropic: $4B in 2023 → $350B today
- xAI: Didn’t exist in early 2023 → $230B+ today
- Lovable: $1.8B (July 2025) → $6.6B (Dec 2025)
- Harvey: $715M (Dec 2023) → $1.5B (July 2024) → $3B (Feb 2025) → $5B (June 2025) → $8B (Oct 2025)
Meanwhile, traditional B2B? Mega-rounds ($100M+) collapsed from 147 deals in 2021 to just 21 in the 12 months through mid-2025.
The money didn’t disappear. It just moved — all to AI.
If you’re not building AI-native, you’re competing for a rapidly shrinking pool of capital.
37 Podcasts, And One Moment In Time: AI in B2B Really Begins To Change Everything
37 podcasts. One year. And honestly? I feel like we captured a moment in time that will define the next decade of B2B software.
The rules changed. The playbooks changed. The math changed.
T3D2 is dead. Co-pilots failed. Founders need to stay. Agents are real. And the wealth creation is unlike anything we’ve ever seen.
2026 is going to be even wilder.
See you on the pod.
