The Filter Upstream Changed. You'll Feel It in 18 Months.
Seed investors are the intake valve of the venture stack. What they select today is what shows up in your growth portfolio, your M&A pipeline, and your competitive landscape in two to three years. The filter just changed, and many growth-stage investors and operating companies haven’t noticed yet.
An early-stage investor is like a high school teacher. They are not competing with the university teacher — they’re preparing students to get to that stage. The high school curriculum just got rewritten. The university hasn’t updated its admissions criteria.
The misconception: growth equity doesn’t care about early-stage
They should. The ZIRP era proved it. Loose seed money created a flood of companies with inflated valuations. Growth funds inherited that overhang. “We’re still digesting that, actually”, says Tina Cheng, Partner at Cherubic Ventures, which has backed seed-stage founders for over a decade.
The same transmission mechanism is operating now, just with different content. What’s being selected for at seed has fundamentally shifted, and every stage downstream will feel the consequences.
What’s changing at the intake
Software lost its premium. Seed investors are deprioritizing pure software companies. When the cost of writing software approaches zero, pure software becomes a commodity. Too easy to replicate, too hard to defend. The price of software-only businesses will compress over time because the barrier that once protected them no longer exists.
For the growth investor, the implication is: the next generation of companies reaching Series B and C will look different from the last one. Fewer SaaS plays. More companies where software is a layer on top of something real: infrastructure, supply chain, biotech, energy, regulatory moats. The software companies worth a lot today all share a pattern — a real-world component underneath. Amazon is a warehousing and logistics machine. Uber is a driver network and regulatory moat. The code is the interface, not the advantage.
“Software sitting closer to physical work, that’s what we look for”, Tina shares.
The benchmarks inflated past recognition. A million dollars in ARR used to be the Series A bar. Now it barely registers. The criteria that used to be a Series B is now a Series A. “It’s the same fraction of the company, but you need to write bigger checks.” When Cursor and Lovable can hit $50M ARR in a year with tiny teams, the old milestones are decorative. Companies that would have been Series A candidates two years ago float between rounds — too mature for seed, too small for what Series A now demands.
For growth-stage operators, this means the companies arriving at your door already survived a harder filter. They’re leaner, faster, and built on different assumptions. Your evaluation frameworks need to match.
Distribution replaced technology as the selection criterion. Cal AI reached $30M ARR through influencer marketing and paid acquisition. No VC money. The product wasn’t technically remarkable. The go-to-market was everything.
When everyone is building, reaching buyers becomes the constraint. Seed investors now pattern-match on distribution capability as much as technical differentiation. The companies graduating upstream will have distribution baked in from day one. If your portfolio companies still treat go-to-market as a post-product function, they’ll face competitors who never made that separation.
What to do instead
Update your pattern-matching to what’s being selected upstream. If you’re evaluating growth-stage companies or acquisition targets using last cycle’s criteria — SaaS metrics, pure software moats, traditional ARR benchmarks — you’re reading an expired map. The companies coming through the pipeline were selected for different traits: physical-world integration, distribution-first thinking, capital efficiency at AI-era speed.
Stress-test your portfolio’s moats from the bottom up. For every software company in your portfolio, ask: what happens when a seed-stage company, built with near-zero software costs, enters this market? If the answer depends on your codebase, that’s not a moat anymore. The durable advantages are data flywheels, regulatory positions, customer integration depth, and domain networks — the things that can’t be vibe-coded in an afternoon.
Build upstream intelligence as a practice. Watch what seed funds are selecting against. What categories are they avoiding? That’s your 2-year forecast for what the competitive landscape looks like.
The diagnostic
Do you know what seed investors are selecting for right now? If not, you’re evaluating companies against yesterday’s criteria — and the ones that pass your bar may be the ones that couldn’t pass theirs.