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