---
title: "Who Wins the GTM AI Race: Startups vs Incumbents?"
author: "Unique Research"
sourcePublication: "Unique Research (FFCAP English library)"
originalPublishedAt: "2024-09-19T16:30:00+08:00"
canonical: "https://ffcap.cn/en/research/20240919a05w1h00"
source: "https://ffcap.cn/en/research/20240919a05w1h00"
language: "en"
cnSourceUrl: "https://view.inews.qq.com/a/20240919A05W1H00"
---

# Who Wins the GTM AI Race: Startups vs Incumbents?

_Unique Research · Insight · September 19, 2024 · Framing: Mark Roberge (Co-founder & MD, Stage 2 Capital) at SaaStr Annual 2024_

We are at the threshold of the most consequential technology shift of our lifetimes. For the first time in twenty years, B2B software has a blank canvas. At SaaStr Annual 2024, Stage 2 Capital’s Mark Roberge staged the question: in the GTM AI race, who wins — startups or incumbents?

## Key takeaways

1.  AI is a platform-scale break — a rare chance to reimagine B2B software, not bolt features onto old workflows.
2.  Founders must stress-test fitness for an AI-native world 5–10 years out, and reshape product/solution around that future.
3.  We are in a hype cycle (think internet ~1998). Do not trade long-term relevance for short-term narrative.
4.  Christensen’s Innovator’s Dilemma still binds large firms — understand it to find white space.
5.  Build an AI strategy that survives the hype, not a copy of today’s patterns.
6.  Think big, start small — ambitious vision, manageable first wedges.
7.  Bezos (books) and Jobs (music) modeled focused starts inside larger visions.
8.  Startups may hold edges in agility, top AI talent access, and ship speed; their hard problem is data.
9.  Time is a startup advantage while large buyers’ security teams freeze incumbent GTM AI adoption.
10.  Re-imagine product and commercialization; separate real disruption from froth; sequence the roadmap deliberately.

## Will your company still matter in an AI-native world?

Founder job #1: picture the market you serve once AI is ambient in 5–10 years — then reverse-engineer today’s product. Ignore that and you risk irrelevance. The internet parallel is sharp: 1997 as year one, today’s heat like 1998 — Goldman/McKinsey talent flooding into “put the brochure online” ideas, with Snowflake and Uber still invisible. Early leaders (AOL, Netscape) did not always finish. High valuations on incremental AI are a warning. Gluing AI into existing workflows is not disruption.

## An AI strategy past the hype cycle

Aim for multi-decade innovation, not a one-season AI slide. Christensen’s dilemma: scaled firms optimize for large, profitable products and ignore early disruptors that look small or cannibalizing. When you re-underwrite your startup, keep that clear — and stay sober through the hype.

## Design large, move in small steps

A ten-year vision the market cannot buy yet bleeds customers; staying only in today’s business invites disruption. Resolution: big destination, small wedges. Ship an MVP to ~$10M; test and expand for a few years; then open a new market/product toward ~$50M. Today’s idea may be obsolete — hold a vision beyond current demand, start small, wait for the market to ripen.

**Bezos was not trying to sell books.** He ranked ~50 categories and started with books — high SKU count, margin, low online-risk for cautious buyers — then spent years hardening warehouses and logistics competitors could not match. **Jobs** took a similar path via music (item #2 on Bezos’s mental list). Many of the world’s most valuable companies were born on tech-revolution waves. That is the opening.

## Startups vs incumbents

To beat incumbents, startups need high-quality data — collection becomes a core business muscle. Enter niches where large firms have not yet amassed data. On talent: top engineers often prefer startups’ lower bureaucracy and freedom to ship frontier work over max cash at giants. Who wins — data-rich, AI-talent-thin incumbents, or talent-rich, data-thin startups? Many bet on startups via diminishing returns at scale, plus creative data-acquisition paths when you are not the incumbent.

Widen the aperture: we are in hype; the market is not ready for the full vision. Find a wedge where incumbents lack data, and build the moat around data you can uniquely get.

## Time as startup advantage

Stage 2 Capital’s experiments: enterprises are curious about emerging GTM AI but security teams inexperienced with it force a wait-and-see. That freezes the carriers — huge budgets, headcount, and engineering depth that turn like aircraft carriers. Startups move like F-15s.

**If you can:** redefine the company vision for post-AI; separate hype from real disruption; hold the ambitious vision; accept the market is not ready; sequence the product roadmap; and target the market with precision — you hold an edge.

— END —

---

Editorial English edition published on UniqueCapital: https://ffcap.cn/en/research/20240919a05w1h00
On-site reading page: https://ffcap.cn/en/research/20240919a05w1h00