In This New Paradigm, the First Check Is the Best Check

AI Is Unlocking the Institutionalization of First-Check Investing

AWS unlocked seed. AI is unlocking first-check. Before 2005, seed-stage investing barely existed as a formal category. Building an internet business required millions in upfront capital—hardware purchases, servers, commercial databases, data-center contracts, bandwidth, and staff—just to get online. Given those costs, the first true institutional round was generally the Series A, and most "seed" activity was sporadic angel funding rather than a repeatable strategy.

In 2006, Amazon Web Services changed the equation. Startups no longer needed to rack servers or sign colocation contracts; compute, storage, and networking could be consumed on demand. The cost to build fell by orders of magnitude, the number of credible teams exploded, and a new institutional category emerged: seed-stage investing. A professional seed strategy—small checks, fast cycles, dense networks—suddenly made sense.

During this period, legends were born:

• •First Round Capital (2004): $510K into Uber's seed → ~$2.5B at IPO
• •Lowercase Capital (2010): ~200x fund from Instagram, Twitter, Uber
• •Floodgate (2006): early in Lyft, Twitter, Chegg, & more
• •SV Angel (2009): early backer of Airbnb, Dropbox, & Pinterest
• •Felicis Ventures (2006): early in Shopify, Credit Karma, Adyen
• •True Ventures (2005): early in Wordpress, Fitbit, Peloton, Ring & more

The common thread was straightforward: with build costs collapsing, founders could do more with less; barriers to starting dropped; and an institutional approach to the earliest round—seed—became both possible and lucrative.

Today, AI is playing the AWS role for this era. Foundation models, open-source model families, and AI copilots compress product development and reduce the need for large teams at inception. Startups can rent world-class intelligence via API or adapt open models rapidly, turning months of engineering into weeks and pushing products to market with unprecedented speed. Distribution has improved in parallel: AI-native products often find pull faster, and workflows instrumented by AI generate clearer proof points earlier in the life cycle.

The result is unprecedented performance:

• •50%+ improvements in capital efficiency
• •~2× faster time to $1B+ valuations (≈4-year average)

Those dynamics have profound implications for venture returns. In the pre-AI pattern, many unicorns took eight to twelve years to reach IPO scale and diluted roughly two-thirds from first check to exit. In the emerging AI pattern, companies are reaching IPO-ready scale materially faster while diluting far less. Faster compounding and lower dilution improve DPI and IRR for the earliest investors—and they make it harder for downstream capital to "buy in" later at attractive prices. That is why so many seed and Series A firms are trying to move earlier. But "moving earlier" only works if you have a structural sourcing and access edge; otherwise, you end up competing for the same few signals everyone else can see.

This is precisely where first-check institutionalization becomes valuable. When build costs fall and time to scale compresses, the returns from investing first improve materially (i.e., first-check investors make more money because they incur significantly less dilution to exit and the time to exit compresses). As result, unlike in the current seed market, first-check investors generally don't need their portfolio companies to become worth $5B+ to generate outstanding returns; first-check investors—depending on their portfolio construction—can create meaningful returns from $500M+ exits, which are far more common than $5B+ outcomes. In today's secondary-enabled market, a disciplined first-check investor can prudently realize partial liquidity from top positions as they approach $500M+ valuations, de-risking the fund while preserving substantial upside. Given the rapid growth, capital efficiency, and active secondary markets in this cycle, selective sales can return ~0.5x–1.0x DPI early—providing downside protection—while maintaining core exposure to the outliers that drive overall performance.

021T is built for this moment. While many firms are trying to move earlier, we believe few have any durable advantage to execute this strategy well at scale. You cannot win first-check without a structural edge (e.g., brand, unique network access, etc), and ours is a programmatic model in the Harvard–MIT ecosystem. Cambridge is among the most productive startup nodes in the world, yet it is chronically under-served at inception—largely due to geography. Our strategy is to identify the strongest Cambridge founders early, help them converge on ambitious ideas, write the first check, and then bridge them to top-tier West Coast follow-on. In other words: institutionalize the first check where the talent density is highest and the competition at inception is thinnest.

AWS unlocked seed. AI is unlocking first-check. The firms that recognize this shift—and pair it with a true structural edge—will capture the next generation of outsized outcomes. 021T intends to be one of them.