SaaStr 818: "Anthropic, Cursor, Fal & Bessemer: The Realities of Scaling AI"

The Official SaaStr Podcast
Release Date: September 5, 2025

Episode Overview

This SaaStr podcast episode features a panel led by Talia Goldberg (Partner at Bessemer Venture Partners) with leaders from Anthropic (Kelly Loftus), Cursor (Jacob), and Fal (Gorkem). The discussion explores how the emergence of AI-first businesses is upending traditional SaaS metrics and go-to-market (GTM) strategies, focusing on the operational and financial realities of building and scaling AI companies. The panelists share candid stories about margin pressures, unique sales structures, hiring tactics, productivity measurement, and maintaining collaborative dynamics in a competitive AI landscape.

Key Discussion Points & Insights

1. AI Disrupts Traditional SaaS Metrics

• Gross Margins Are Lower in AI:

  • Traditional SaaS scaled with near-zero marginal costs per added customer. In contrast, AI companies face significant costs to serve each user because of GPU workloads, making gross margins much lower.
    • “No one in AI really has 80%, 90% gross margins because the way we sell software changed.” — Gorkem, Fal (04:23)
  • Tradeoff: Lower margins, but much faster growth rates than historic SaaS.

• Old VC Benchmarking Struggles

  • Metrics like CAC (Customer Acquisition Cost), LTV (Lifetime Value), and classic “good/better/best” frameworks don’t map cleanly to AI company realities anymore.
    • “uvcs… the metrics are totally broken… the businesses look different today.” — Talia Goldberg (03:21)

2. The Rapid Scaling of AI Companies

• Growth Rates Are Exponential

  • “Companies are breaking out going from 0 to 50 million like really, really fast.” — Gorkem, Fal (04:23)
  • Unlike traditional SaaS (triple, triple, double, double), the AI space sees unprecedented velocity.

• Smaller, Leaner Teams

  • “Seems like there's so much demand for AI that people don't need these massive sales teams and they can get things done with much leaner, leaner teams.” — Gorkem, Fal (04:23)

3. Operational Realities: Margins, Pricing, and Sales Structures

• Cost Structure and Margins

  • Models are both getting cheaper and more expensive: older, smaller models are cheaper to run, but customers are demanding ever-larger, state-of-the-art models which spike inference costs.
    • “The margins even got actually lower because it’s so expensive to run these models. People demand them.” — Gorkem, Fal (09:09)

• Dynamic Pricing Evolution

  • AI startups are experimenting with usage-based and value-based pricing, moving away from flat per-seat pricing.
    • “It’s really hard to predict… you look at the value being delivered to the customer… and the total value being created by this technology, which is large and quickly growing.” — Jacob, Cursor (06:54)
  • For highly skilled engineers, even expensive tools can have massive ROI if they increase productivity.

• No Quotas in Sales Teams

  • None of the panelist companies currently have classic sales quotas (variable compensation based on hitting a number). Instead, they rely on shadow targets and collective goals.
    • “When I joined, we did not have the concept of quotas… it’s hard to pick a number on what you want to actually measure each rep by.” — Kelly, Anthropic (11:48)
    • “No quotas at Anthropic today.” — Kelly, Anthropic (12:45)
    • “We decided, okay, this is useless. We are not doing quotas. It’s impossible.” — Gorkem, Fal (13:07)

4. AI in Internal Workflow & Hiring Tactics

• AI Drives Internal Productivity

  • Tools built around AI are used for inbound lead qualification and automating tasks like onboarding through AI-powered Slackbots (Claude).
    • “Employees can go in, ask questions, Claude is on the back end… and then retrieve and answer the employee’s question. It’s been extremely useful for productivity.” — Kelly, Anthropic (19:10)

• Hiring via Research Grants

  • Fal hires directly from recipients of open research grants — applicants propose projects related to efficient AI, get compute resources, and some are later hired full-time.
    • “We hired maybe four people through that research grants program and has been really useful for us.” — Gorkem, Fal (16:29)
    • “That’s a genius tactic.” — Talia Goldberg (17:29)

• Technical Sales Teams

  • Cursor’s sales org is filled with technical people, often building tools to accelerate sales themselves with Cursor.
    • “Many people in our sales Org are also building tools to help a sales Org using Cursor and I think that's a trend that will continue.” — Jacob, Cursor (15:42)

5. Navigating Competition vs Collaboration in the AI Ecosystem

• Symbiotic Relationships
  • Cursor is a major Anthropic customer, with mutual collaboration and feedback loops—despite potential overlaps in products.
    • “We want to partner with companies like Cursor to drive the models forward and push the capabilities of what is actually possible with these models.” — Kelly, Anthropic (22:39)
    • “Whenever the models get better, we're very happy because it means Cursor becomes more valuable to our users.” — Jacob, Cursor (23:36)

6. Metric ‘North Stars’ and What Matters Most

• What Do Leaders Track?
  • Fal cares about attracting “big logos,” diversified revenue sources, and retention/churn.
    • “We want to make sure it's coming from at least 30-35 different companies rather than being concentrated at the top.” — Gorkem, Fal (27:04)
  • Cursor cares most about building products they personally want to use, measuring real usage and user delight alongside revenue.
    • “The thing we care about most of all is whether we personally want to use it or use a new feature in our like day to day life.” — Jacob, Cursor (28:22)
  • Quality and frequency of real user interactions (such as number of auto-complete tabs pressed) matter even more than abstract financial targets.

7. Measuring Productivity & Impact

• Challenging to Quantify Output
  • AI tools are enabling 2x or greater productivity gains in software engineering, but measuring the exact impact is nuanced.
    • “To pick a single acceleration factor is difficult because it depends where you were previously… we very much care about making something that we want to use.” — Jacob, Cursor (24:54)

Notable Quotes and Memorable Moments

• On Gross Margins and GTM:
  • “No one in AI really has 80%, 90% gross margins because the way we sell software changed. …everyone has less margins than traditional SaaS but everyone is growing like crazy.” — Gorkem, Fal (04:23)
• On Pricing Evolution:
  • “You look at the value that's being delivered to the customer… how can we increase that value and how can we be part of that supply chain that takes the electricity and converts it into something that is useful to people.” — Jacob, Cursor (06:54)
• On Predicting Sales:
  • “We decided, okay, this is useless. We are not doing quotas. It’s impossible. To predict one thing now we are experimenting. Maybe we can do shorter term quotas…” — Gorkem, Fal (13:07)
• On Collaboration Amid Overlap:
  • “We want to partner with companies like Cursor to drive the models forward and push the capabilities of what is actually possible with these models.” — Kelly, Anthropic (22:39)
• On Product & User Metrics:
  • “The thing we care about most of all is whether we personally want to use it or use a new feature in our like day to day life.” — Jacob, Cursor (28:22)

Timestamps for Important Segments

| Timestamp | Segment | |-----------|------------------------------------------------------------------------------------------| | 02:05 | Panel intros: Guests briefly share backgrounds | | 03:21 | Broken SaaS metrics & what actually matters in AI businesses | | 04:23 | Why AI companies have lower gross margins and very fast growth | | 05:49 | Pricing and margin experimentation; the “cost to serve” dynamic | | 09:09 | Lower margins as a result of customers wanting bigger, costlier models | | 11:14 | Rapid scaling of go-to-market/sales teams at Anthropic | | 12:45 | The no quotas approach on sales teams in AI | | 16:14 | How AI is used to automate go-to-market/sales—examples | | 19:10 | Use of Claude AI for internal onboarding at Anthropic | | 20:27 | Key strategic decisions and the importance of doubling down on technical advantage | | 22:39 | Navigating collaboration vs competition between infrastructure and application companies | | 24:54 | How to think about productivity gains from AI tools | | 27:04 | What metrics do these founders actually care about |

Panelists

• Talia Goldberg (Host, Bessemer Venture Partners)
• Kelly Loftus (Head of Startup Sales, Anthropic)
• Jacob (Cursor, ex-Tab9/OpenAI/Supermaven)
• Gorkem (CTO & Co-Founder, Fal)

Tone & Style

The conversation is candid, fast-paced, technical, and honest, with panelists frequently challenging conventional wisdom and sharing “what’s really happening” inside fast-growing AI companies. There’s a strong spirit of experimentation, humility, and appetite for innovation—all delivered with a practical, operator-first mindset.

Conclusion

This episode provides a rare, inside look at the new rules of scaling AI-first businesses. Founders and investors will find it particularly rich with insights about evolving financial metrics, sales strategies, team building, product focus, and collaborative competition in today's AI landscape. The panelists stress that agility, focus on core user needs, and a willingness to break from SaaS tradition are the new foundations for AI company success.