Podcast Summary: "What VCs Are Looking For in AI Startups Today"
The AI Podcast | Host: Jaden Schaefer | March 3, 2026
Episode Overview
In this episode, host Jaden Schaefer unpacks the evolving dynamics of venture capital interest in AI startups as of 2026. He outlines how investors' expectations have shifted beyond simply having “AI” in the pitch, and details the nuanced factors that VCs now look for when considering investments. Drawing from recent industry interviews—including insights from leading VC partners—Schaefer provides a roadmap for founders and professionals trying to succeed in today's competitive AI landscape.
Key Discussion Points and Insights
1. The Changing VC Landscape for AI Startups
- AI Hype and Selective Funding: Investors have poured billions into AI, but the spread of “AI-wrapped” SaaS products means VCs are now more discerning.
- Shift in Criteria: Simply adding an AI feature or labeling a product as “AI-powered” is no longer enough to attract funding.
2. What VCs WANT: Depth and Real Workflow Ownership
- AI Native Infrastructure: Jaden summarizes Aaron Holiday’s view (Managing Partner, 645 Ventures):
- “The categories still getting the most interest are AI native infrastructure, so that’s vertical SaaS built on proprietary data, systems of action that actually complete tasks, and platforms embedded deeply into mission critical workflows.” (03:10)
- Execution, Not Just Assistance: The key is products that “actually complete something." AI must automate and execute, not just suggest or assist.
- "AI that actually completes something..." (04:05)
- Proprietary Data Moats:
- VCs want companies with unique data and real integration into customers’ critical tasks.
- Products must have data or capabilities that ChatGPT, Anthropic, or other generative models cannot easily replicate.
- Abdul Abdirhan (F Prime): "Vertical software without any proprietary data moats is no longer super compelling." (06:45)
- Real Domain Expertise: Expertise and control over workflow are highly valued.
3. What VCs DON'T Want Anymore
- Discouraged Models:
- Thin workflow layers, generic horizontal tools, surface-level analytics, and basic automations:
- "If your differentiation mostly lives in UI and automation, that’s no longer enough. The barrier to entry has dropped..." (Igor Ryabinsky, AltalR Capital, 08:10)
- Thin workflow layers, generic horizontal tools, surface-level analytics, and basic automations:
- Low Barriers to Replication: Products easily copied or replicated by existing APIs or large LLMs are a “red flag.”
- Example Pitfalls:
- Productivity tools, basic CRMs, simple wrappers on top of existing APIs: all considered vulnerable.
4. The Rise of Agents and Devaluing Workflow Stickiness
- Agents Are Doing the Work: As autonomous agents advance, traditional SaaS moats (like getting humans to use your workflow tool) are eroding.
- "If agents are doing the work, a lot of the traditional workflow stickiness is becoming a lot less relevant..." (10:00)
- Direct Execution: Integrations and platforms like Zapier, Bubble, and Make are losing relevance as AI agents increasingly perform tasks directly with minimal user involvement.
- Shift in Pricing Models: Per-seat SaaS pricing models are giving way to usage- or consumption-based models (e.g., tokens or API usage).
5. Case Study: The Calai Acquisition
- Counterpoint: Can “Thin” Apps Succeed?
- Calai, perceived as a thin wrapper around ChatGPT for calorie tracking, reached 15 million downloads and over $30M recurring revenue.
- Their success stemmed from excellent growth hacking on social channels, leading to acquisition by MyFitnessPal.
- "They had 15 million downloads, they had over $30 million in annualized revenue. And really, a lot of their unlock was that they really hacked the growth, hacking on social media TikTok and making shorts." (12:45)
- VCs’ Focus Remains on Defensible, Deep Tech: Even though quick and thin solutions can win with strong distribution, VCs favor more robust, defensible business models with deep integrations and proprietary data.
Notable Quotes & Memorable Moments
-
On the AI SaaS Trend:
“Every company that is raising money now is no longer just a SaaS, it’s like an AI company. And so... while every company has kind of added AI to their... pitch deck, I think it’s becoming a lot more selective on who's actually getting money. Right? You can't just put AI on your pitch deck and get money.” (02:10, Jaden Schaefer) -
On True Product Value:
"What they want tools to be able to do is have some sort of custom data set, some sort of, you know, deep integration into something that's super, super critical. And it's not something that just like a ChatGPT or Anthropic can, can replicate easily." (07:05, Jaden Schaefer) -
On Defensible Moats:
"Vertical software without any proprietary data moats is no longer super compelling." (06:45, Abdul Abdirhan) -
On the Commoditization of Workflow Tools:
“If your differentiation mostly lives in UI and automation, that's no longer enough. The barrier to entry is dropped, which makes building a real moat a lot harder.” (08:10, Igor Ryabinsky) -
On Pricing Models:
"These kind of set in stone pricing, per seat or per subscription... are looking a lot weaker compared to the consumption based approaches..." (09:35, Jaden Schaefer) -
On AI Agents Changing SaaS:
"If agents are doing the work, then a lot of the traditional kind of workflow stickiness is becoming a lot less relevant. In the past, getting humans to operate inside of your software was a pretty powerful moat. Now if an agent can perform the task directly, then owning the human interface doesn't actually matter that much in my opinion." (10:00, Jaden Schaefer) -
On the Calai Example:
"They had 15 million downloads, they had over $30 million in annualized revenue. And... their unlock was that they really hacked the growth, hacking on social media TikTok and making shorts." (12:45, Jaden Schaefer)
Important Timestamps
- [01:00] — Introduction: The evolving AI startup funding landscape
- [03:10] — Top VC priorities for AI startups (Aaron Holiday, 645 Ventures)
- [06:45] — Importance of proprietary data and defensible moats (Abdul Abdirhan, F Prime)
- [08:10] — Red flags for VCs: UI/automation-heavy, shallow products (Igor Ryabinsky, AltalR Capital)
- [09:35] — Shift in SaaS pricing models: Consumption vs. seat-based
- [10:00] — Agents vs. workflow: The fading value of workflow stickiness
- [12:45] — The Calai acquisition: Can “thin” AI apps still win?
- [14:00+] — Recap & Closing thoughts
Takeaways for Founders and Professionals
- Focus on Depth: Build around critical workflows, proprietary data, and tasks AI can autonomously complete—not just augment.
- Build Defensible Moats: Ensure your product can’t easily be copied by LLMs or API wrappers; lean on domain expertise and unique datasets.
- Embrace Flexible Business Models: Consider usage-based pricing over traditional SaaS structures.
- Distribution Still Wins: Even thinner products can break out, but are less likely to attract VC backing unless growth is exceptional.
- Stay Agile: Speed, adaptability, and focus now outperform massive codebases or entrenched workflows.
This episode offers a concise, practical primer for anyone watching or building in the AI startup space, summarizing where smart money is moving—and why.
