How Deal Terms Vary by Stage and Region

· 13 min read
venture-capital startups finance

Venture deal terms are not one-size-fits-all. Where a company sits in its lifecycle and where it’s based both shape the terms that end up in signed documents, sometimes in ways that run counter to the conventional wisdom. At Aumni, we had a unique vantage point on this because our dataset wasn’t just tracking valuations and round sizes. We had the actual legal provisions from tens of thousands of executed financing agreements. That depth let us ask questions that most market reports can’t touch, and the answers were often surprising. This post pulls together two threads from that work: how stage dynamics create structural patterns (and paradoxes) in early-stage markets, and how geography influences deal terms in ways that matter for both founders and investors.

This post is for informational and educational purposes only. Nothing here is legal, tax, or investment advice. Consult qualified professionals before making any decisions based on this content.

The Early-Stage Concentration Paradox

There’s a paradox that shows up in early-stage venture markets from time to time, and it’s worth understanding because it fundamentally changes the math for both founders and investors. The paradox goes like this: seed deal sizes grow, but the rate at which seed companies graduate to Series A falls. More capital flows into fewer companies at the earliest stage, and fewer of those companies make it to the next one. If you’re building a portfolio or planning a fundraise without accounting for this dynamic, you’re working with the wrong assumptions.

In a healthy early-stage market, you’d expect deal sizes and graduation rates to move roughly together. Bigger rounds should mean better-capitalized companies, which should mean more of them hitting the milestones needed to raise a Series A. But that’s not what happens during a concentration cycle.

What actually happens is bifurcation. The top end of the seed market pulls away from the middle and bottom. A select group of companies (usually those with experienced founding teams, strong early traction, or positioning in whatever sector has captured investor enthusiasm) command outsized rounds. Meanwhile, the broad middle of seed-stage companies raises amounts that are flat or barely growing. And the graduation rate for the entire cohort drops, often dramatically.

This goes beyond the simple “good companies get funded” narrative. What we’re seeing is a structural shift in how capital and opportunity distribute at the earliest stage, with cascading effects on everything from portfolio construction to LP returns.

Why Bifurcation Happens at Seed

The seed stage is uniquely susceptible to this kind of bifurcation for a few reasons.

First, seed investing is where conviction matters most and data matters least. At Series A and beyond, investors can point to revenue growth, retention curves, unit economics. At seed, the decision is driven more by narrative, team pedigree, and market thesis. When investor sentiment shifts toward caution, the companies that still attract capital are the ones with the strongest narrative signals, and those companies often attract more capital than they need because multiple firms want in. The result is a handful of oversized rounds alongside a much quieter broader market.

Second, seed is where new thematic enthusiasm tends to concentrate first. When a new technology wave captures investor attention (as AI did starting around 2022-2023), the earliest dollars rush into seed-stage bets in that space. This pulls the top quartile of seed deals upward without lifting the rest of the market.

Third, the bar for graduating from seed to Series A is set by a different group of investors than the ones writing seed checks. Series A investors have their own deployment pace and selectivity criteria. When they tighten their standards (which tends to happen during the same cautious periods that drive seed bifurcation), the graduation rate drops even for well-funded seed companies. The two effects compound each other.

What the Paradox Means for Portfolio Construction

If you’re a seed-stage investor and graduation rates drop dramatically, the portfolio math changes completely.

At a healthy graduation rate, a portfolio of 20 seed investments gives you a solid group of companies advancing to Series A. That’s a reasonable number of shots on goal for generating fund-level returns. When that rate collapses to the low teens, that same portfolio gives you 2 or 3. The confidence interval around your returns widens enormously, and the outcome becomes much more dependent on the specific companies that graduate rather than the portfolio as a whole.

There are two responses to this. You can go wider, investing in more companies to maintain the same expected number of graduates. Or you can go deeper, concentrating your portfolio on the companies most likely to make it through the filter. Both approaches have tradeoffs. Going wider means smaller checks and less ownership. Going deeper means higher conviction requirements and higher concentration risk. Neither is obviously right, but doing nothing (keeping the same portfolio size and check size while graduation rates fall) is probably wrong.

For founders, the implication is that your seed round needs to do more work than it used to. If the historical path was “raise a seed, hit some milestones, raise an A in 18 months,” a low-graduation environment changes that calculus. Your seed capital may need to last 24 months or longer. The milestones required to unlock a Series A may be meaningfully higher than what the last generation of companies needed. And the timeline between rounds means you’re exposed to shifting market conditions for longer.

The 2022-2024 Cycle in Practice

The 2022-2024 period provides one of the clearest examples of this paradox in action. I was running the data science team during this stretch, and we watched the numbers shift in real time as the post-ZIRP correction reshaped early-stage activity.

The bifurcation was stark. The top end of the seed round distribution surpassed $7M – a number that would have been a respectable Series A not long before – while the median barely moved. At Series A and beyond, deal sizes remained below their 2021-2022 peaks. The seed market had effectively split into two separate markets. Meanwhile, the 2-year graduation rate from seed to Series A fell from around 41% to roughly 12%, a collapse from two in five companies advancing to barely one in ten. The new baseline for time between rounds settled around 24 months, compared to roughly 18 months during 2021.

What made this particularly painful for founders was the compounding effect. Fewer companies graduated, and the ones that didn’t graduate on time burned through their seed capital while waiting. Many landed in an awkward middle ground: too far along to re-raise at seed terms, but without the metrics to attract Series A interest. Some turned to bridge rounds, which kept the company alive but diluted the cap table in ways that made the eventual Series A less attractive. The paradox didn’t just filter companies out – it created a secondary set of problems for the ones stuck in between.

How to Spot the Pattern

This pattern isn’t unique to any single cycle. It tends to emerge whenever the market transitions from exuberance to caution, because that transition concentrates capital around consensus picks while raising the bar for everyone else. The signal to watch is the divergence between top-quartile and median seed deal sizes. When that gap widens, bifurcation is underway. For founders, that means paying more attention to graduation rates for your cohort than to headline round sizes. For investors, the variable that actually drives returns is the expected number of graduates per fund, not the number of investments.

Turning to Geography: Why Region Still Matters

Stage dynamics are only half the story. The regional analysis, done in partnership with the NVCA using data from over 100,000 venture transactions, looked at how deals get structured differently across geographies.

The conventional wisdom is that the market is national and deal terms are largely standardized. The NVCA model documents have created genuine baseline norms. But meaningful regional variation shows up when you dig into the data – driven by different investor bases, different legal communities, different norms around negotiation, and different competitive dynamics. And some of the findings were genuinely surprising.

Lead Investor Ownership and Round Dynamics

Across the dataset, lead investors at the median were taking a fairly consistent share of each round. But regional differences emerged when you looked at the tails of the distribution. In the most competitive markets (think San Francisco, New York), lead investors tended to take a somewhat smaller percentage of the round, reflecting the presence of larger syndicates and more co-investor participation. In less saturated markets, leads tended to take a bigger piece.

This makes intuitive sense. In competitive markets, founders have more leverage to assemble larger syndicates and spread ownership across multiple investors. In thinner markets, the lead typically needs to commit more capital to fill the round, which translates to a larger ownership stake. It also affects follow-on behavior: in competitive markets, multiple co-investors create a built-in syndicate for future rounds, smoothing the path to Series A. In thinner markets, a seed company’s ability to raise a follow-on often depends heavily on whether the single lead is willing to re-up.

Round sizes also varied by geography, though this tracked more closely with the stage and sector mix of each region’s ecosystem than with any inherent geographic premium. When controlling for sector and stage, the geographic premium shrank considerably – suggesting the “where” matters less than the “what” when it comes to how much a company can raise.

Liquidation Preferences and Protective Provisions

The data on liquidation preferences by region was one of the more interesting findings. The vast majority of deals across all geographies used standard 1x non-participating liquidation preferences, which is the industry norm and the default in the NVCA model documents.

But the frequency of atypical terms varied. In some regions, the prevalence of participating liquidation preferences (which give investors both their money back and a share of the remaining proceeds) was notably higher than in others. Similarly, the frequency of liquidation preference multiples greater than 1x, while rare everywhere, was not uniformly distributed.

These differences reflected investor culture and competitive dynamics. Where founders have multiple term sheets, investors are less likely to push for aggressive provisions because the founder can go elsewhere. Where competition is thinner, investors have more leverage. It’s the same force that drives bifurcation at seed, just expressed geographically: concentration of capital and attention creates a power law in terms as well as in deal sizes.

Anti-Dilution: Where Standardization Wins

Broad-based weighted average anti-dilution was the standard across essentially all geographies and all stages, which makes sense given that broad-based weighted average is the default provision in the NVCA model documents and the expected norm in nearly all venture financings. For the unfamiliar: anti-dilution provisions protect investors if a company raises a future round at a lower valuation (a “down round”). Broad-based weighted average is the moderate version, adjusting the investor’s conversion price based on a formula that accounts for the size of the down round relative to the company’s overall capitalization. It’s widely considered the fair middle ground.

Full-ratchet anti-dilution, the more aggressive variant that gives investors full price protection in a down round (effectively repricing their shares to the new lower price regardless of round size), remained extremely rare. The data showed it appearing in a small single-digit percentage of deals. Where it did show up, it tended to be in later-stage rounds and in situations where the company’s negotiating position was weakest. This makes sense. Full ratchet is a punitive provision from the founder’s perspective, and any founder with alternatives will resist it.

Regional variation in anti-dilution was minimal, which suggests that this is one area where the standardization efforts of the NVCA and the broad adoption of model documents have been genuinely effective. Unlike liquidation preferences, where there’s more room for negotiation, anti-dilution has settled into a consensus norm that doesn’t vary much from market to market. I found this interesting because it shows that standardization can work when there’s a clear “correct” answer that both sides broadly agree on. The challenge with other provisions is that there’s more legitimate disagreement about what’s fair, which leaves more room for negotiation and, therefore, more room for regional norms to diverge.

The Enhanced Model Term Sheet

This analysis tied directly into the Enhanced Model Term Sheet that we developed in partnership with the NVCA. The idea was straightforward: if you’re drafting a term sheet, wouldn’t it be useful to know how the terms you’re proposing compare to what the market actually looks like?

The Enhanced Model Term Sheet embedded market benchmarks directly into the NVCA’s standard term sheet template. Hyperlinked deal terms let users click through to see the market data for that specific provision, broken down by stage, time period, and geography. It drew on anonymized data from more than 100,000 transactions representing over 40,000 investors with combined assets under management exceeding $1 trillion.

The goal wasn’t to tell anyone what their terms should be. It was to give both sides of the table better information about what’s normal. In my experience, a lot of friction in term sheet negotiations comes from asymmetric information. One party thinks a provision is standard and the other thinks it’s unusual. Having data to ground those conversations makes the negotiation more efficient and the outcomes more fair.

What All of This Means for Founders and Investors

Pulling the stage and regional threads together, the main takeaway is that context matters more than people assume. A founder raising a seed round in a hot sector in San Francisco during a period of high investor enthusiasm is operating in a fundamentally different market than a founder raising seed in a mid-tier metro during a period of caution. The deal size, the terms, the likelihood of graduating to Series A, and the timeline for doing so are all different. Treating “the venture market” as a single uniform thing obscures the variation that actually determines outcomes.

For founders, the practical advice is to understand your specific context before you’re in the middle of a negotiation. What does the graduation rate look like for companies at your stage, in your sector, in your geography? What are normal terms in your market? If a term sheet arrives with a participating liquidation preference and that’s standard where you are, it’s a different signal than if it’s rare. Same provision, different meaning depending on context. The Enhanced Model Term Sheet was designed to answer exactly these questions.

For investors, the variation is both a risk and an opportunity. Portfolio construction can’t be static – it has to adapt to the current graduation rate environment. And regional differences mean that what constitutes “market” varies depending on where you’re deploying capital. A fund that invests nationally but uses a single set of term expectations will be miscalibrated for at least some of the markets it operates in. The founders and investors who navigated the 2022-2024 cycle most effectively were the ones who used data to benchmark their specific situation rather than relying on rules of thumb derived from aggregate numbers.

Further Reading

Some of the data and analysis in this post originally appeared on the Aumni blog and here.