David Tisch on early-stage venture in a volatile macro: 'the accordion never reaches early stage'

Summary

David Tisch, co-founder of Box Group, makes a structural argument for why early-stage venture is effectively immune to macro volatility — not because the environment is benign, but because the accordion effect takes 3 to 5 years to travel from late-stage markets down to seed. By the time any given shock reaches early stage, something else has already reset the macro. His illustration: the SVB crisis and rising rates were supposed to be the death of venture, then AI arrived and everyone ripped checks.

That decoupling is the core of his pitch to LPs. The day-to-day job — find great people, write a check, wait 5 to 15 years — doesn't change with tariffs or Liberation Day. The macro does matter for portfolio companies already out raising follow-on rounds or managing revenue retention, but it doesn't change the logic of a new seed investment.

Tariffs and the manufacturing thesis

On tariffs specifically, Tisch is skeptical that the current environment generates the kind of clear catalyst that COVID did for Airbnb. COVID had a legible endpoint — the world returns to normal — whereas the trade war has no predictable resolution. No one is building a factory overnight. The more plausible upside is narrower: American industrialization and re-shoring companies that were already placing that bet may find it easier to raise a Series B now, and that incremental funding could be enough to clear an awkward inflection point.

For a seed investor, timing thematic bets around tariffs is structurally incoherent anyway. A company funded today won't mature for 5 to 10 years, by which point the trade environment will have changed multiple times.

The patience problem in early-stage venture

LP fatigue is real but understandable. Tisch acknowledges that the expected liquidity window for early-stage funds has shifted meaningfully. What used to be a credible 7 to 10 years is now realistically 12 to 15. That's a substantial ask. A 15-year timeline means the young founders Box Group backs today were in pre-elementary school when the fund launched — a psychological mismatch between the investment cycle and anything an LP can intuitively reason about.

The M&A deficit

Tisch argues M&A has been the missing piece in the venture ecosystem for roughly a decade, and he expects that to change. His read on why it dried up: antitrust pressure pushed big tech to fight regulatory battles only on the largest deals, leaving the mid-size acquisition market cold. Simultaneously, founder valuation expectations decoupled from what acquirers would pay. Plaid agreed to sell, the government blocked it, and the signal to other potential acquirers was clear — don't bother.

The OpenAI-Windsurf negotiation (reported at roughly $3 billion) is the data point he finds most interesting. If a well-resourced lab concludes it's faster to buy than build in a category like AI coding tools, that logic could extend to Anthropic, xAI, Amazon, and eventually Fortune 500 companies that need to modernize quickly. The enabling condition is that the Mag 7 and large AI companies now carry valuations high enough to write acquisition checks that actually meet founder expectations — something that wasn't true for most of the past decade.

He's careful not to over-extrapolate. The Windsurf deal being strategic for OpenAI doesn't mean every well-traction AI wrapper is suddenly strategic to a wide range of buyers.

How AI is moving through the Box Group portfolio

Two portfolio examples anchor his view on AI and timing. Clay, which Box Group backed at the earliest stage, was built around a go-to-market product that needed AI to work — and simply waited for the underlying technology to catch up. Seven years in, it clicked. Casetext, a legal AI company, ran a similar playbook: the founder identified early how AI could differentiate the product, split off a team, built fast, and sold before the market got crowded. Both are what Tisch calls the real pattern — founders who can see a space early and stay solvent long enough for the world to arrive.

The flip side: three years into the AI wave, any portfolio company that hasn't figured out how to re-energize a stalled product is running late.

The early-stage market and YC's durability

On the proliferation of fellowships, free capital programs, and new seed vehicles, Tisch is measured. Most new entrants into early stage eventually scale up and drift later — that's how the economics of fund management work. The products offering free capital to founders are probably not free forever and are hard to scale. YC has compounded its reputation because founders who went through it recommend it to others; the NPS holds because the product actually delivers. New entrants have to earn that trust, and few do.