Ed Sim on Seed Valuations, Inception Investing and Reserving for Growth

Boldstart founder Ed Sim describes three sizes of very early financing: a small discovery round, a classic round of roughly $3 million to $5 million and a larger round often raised by an experienced founder. He speaks with Harry Stebbings about why these rounds are becoming more competitive and what early investors should reserve for later.

Sim calls Boldstart an 'inception' investor: it works with founders around the idea and company-formation stage, helping test the concept and recruit initial employees. He says rounds can range from hundreds of thousands to $10 million, though he generally prefers a $3 million to $4 million structure. This describes his firm's strategy and commercial interest.

He dislikes rigid labels such as pre-seed and seed because they can imply an extra financing and more dilution. He cites a PitchBook analysis in which the median company raising pre-seed was 1.2 years old and one raising seed was 2.7 years old, using data over a long period. Those figures are his cited analysis, not a universal definition of either stage.

Sim says large multistage funds had accumulated capital and, after growth investing slowed, were writing substantial early checks to access potential outliers. He points to early investors in Instacart as an illustration of the importance of entry price. His explanation is a market interpretation from 2023.

Stebbings argues too much early cash can harm most startups; Sim agrees that exceptional allocators are rare. He worries high valuations make it harder to recruit employees who expect meaningful equity upside and set a demanding bar for the next round. These are risks in their view, not proof that every large seed round fails.

When a founder proposes raising a range, Sim sometimes asks whether a smaller amount would work. He wants to see whether the team can operate under constraints and avoid spending simply because cash is available. Stebbings points out that founders may rationally value the longer runway of a larger offer.

Sim nevertheless says he will participate in a larger round for an exceptional founder. He describes backing Ian Swanson's AI security company in a $10 million round with a co-lead, after years of knowing him and iterating on the idea. The plan was to combine stages and fund a longer period of development; Sim says the company later raised another round and signed customers.

Sim's 'discovery' round is, in his framework, under $2 million for a team exploring an early market. His preferred classic round is around $3 million to $5 million, while a larger $6 million to $10 million round often backs an experienced founder pursuing a substantial market. These categories are his working vocabulary, not industry standards.

Stebbings worries that a large seed valuation leaves little room for an ordinary follow-on if sales take longer than expected. Sim agrees that extra cash can create pressure as well as comfort. The pair argue that runway only lasts if a founder controls spending rather than hiring and expanding to match the balance sheet.

Sim gives Superhuman founder Rahul Vohra as an example of unusually disciplined use of capital: he says Vohra put a later $10 million raise in a separate account while continuing to operate on earlier funds. Sim also recognizes that such caution may have slowed the company. Boldstart was an investor, so the account comes from an interested participant.

Stebbings questions whether traditional seed-fund sizes can support larger checks and reserves. Sim argues an inception firm must be able to lead small rounds, co-lead larger ones for exceptional founders, and still have capital to support companies later. He favors a concentrated portfolio and meaningful ownership rather than a large number of tiny option checks.

In an illustrative model, Sim describes 20 to 25 companies, an initial check around $3 million to $3.5 million and a target of 15% ownership. Follow-on checks may be larger than the first, and technical enterprise products may need bridge rounds before product-market fit. These are his fund-design assumptions, not guaranteed outcomes.

He recalls supporting Snyk through three rounds before its Series A. When progress was slower than hoped, a new technical approach helped the team sign enterprise customers, he says. Sim argues that a concentrated investor close to the company can see signals outsiders miss and must reserve enough money to support a promising founder; Boldstart has a direct investment interest in Snyk.

For this strategy, Sim suggests a fund around $150 million to $250 million rather than a very small fund or a billion-dollar vehicle. He says Boldstart began with a much smaller fund in 2010, but believes market conditions and the need for reserves have changed. His proposed size is tied to his specific ownership and concentration model.

Stebbings raises a contrary view that higher entry prices and lower public-market multiples could make enterprise SaaS unattractive for investors. Sim says disciplined entry prices and ownership can still work with smaller exits than the largest recent outcomes. He believes cloud migration and security needs continue to create demand, while acknowledging the importance of dilution.

Sim emphasizes capital efficiency and says he avoids funding companies whose spending would largely flow to compute providers. He also describes an opportunity fund designed to support the best companies later and maintain ownership. That structure advances Boldstart's own investment strategy and carries its own risks.

On board seats, Sim distinguishes a very small round, where he says a formal board may not be needed, from a $3 million to $4 million priced round. He often joins the board but wants a cadence appropriate to the founder, not an automatic weekly meeting. Early on, he sees the board partly as preparation for later governance.

He wants founders to ask for help when it is useful and says the relationship must be based on trust. Stebbings notes that other investors favor more frequent contact. Sim does not claim one cadence works for every company or founder.

Sim believes AI can be transformative but says making money by chasing every AI-labeled company at high prices is hard. He compares enthusiasm around the sector with earlier internet cycles. As an investor, he asks first what problem is solved, what is distinctive and whether customers will pay.

He estimates that more than half of Boldstart's companies in certain funds then had AI-related offerings, with more expected to add them. Stebbings distinguishes an AI-first company from an existing product enhanced with AI. The percentages are Sim's portfolio snapshot and forecast at the time, not sector-wide data.

Both speakers see an advantage for some established private companies with customer data and the flexibility to move quickly. Sim cites Snyk's earlier acquisition of machine-learning capabilities and its effort to find security issues in AI-generated code. His confidence in that example reflects Boldstart's investment in the company.

Stebbings predicts a difficult acquisition environment; Sim expects more deals as investors reconsider overvalued holdings. He thinks some investors may accept a return of capital and redeploy it. Discussing Atlassian's acquisition of Loom, he says a deal can be attractive to founders even when a later investor's entry valuation was higher; his assessment is not a complete breakdown of each investor's return.

Sim rejects the claim that growth investing is dead, while acknowledging a sharp decline in valuations. He says strong infrastructure companies with rapid revenue growth could still raise at substantial valuations. His examples and figures are observations from the 2023 market, not a current financing benchmark.

For founders preparing the next round, Sim prefers a fair price and a partner who believes in the business over the highest possible valuation. A very high price can make future performance, hiring and exit options harder. He adds that the right financing depends on circumstances; some compute-intensive founders genuinely need more capital.

Looking back, Sim says excess cash can create a cycle of costly hiring and expectations that become painful when growth misses a target. He also learned that repeated preemptive rounds can arrive before a company has produced enough new evidence to justify increasing an investment.

Stebbings questions automatic follow-on decisions based on early traction, noting that rapid initial adoption is not always durable. Sim agrees that difficult technical products may take longer to show results. He looks for whether customers pay, whether the product is built for the long term and what defensibility exists.

Sim describes Kustomer as a realized investment win and says competing with an incumbent required more features than the founders first expected. Even a new customer-centered architecture had to meet basic requirements such as knowledge-base tools before buyers would switch. He says the company needed a bridge between rounds, illustrating the cost and time of enterprise adoption.

He identifies Snyk as a potential major win that had not yet exited at the time. Its new category took years to reach early revenue milestones, he says, but product enthusiasm and a growing pipeline offered encouraging signals. Sim contrasts that with his mistake of leaning too heavily into rapid follow-on rounds without enough new evidence.

In quick-fire questions, Sim avoids a precise prediction for how much AI investment will fail because he considers the category hard to define. He expects AI features to spread across enterprise software but believes substantial capital chasing language-model businesses could be lost. Both are forward-looking opinions.

He finds the early-stage market intensely competitive and thinks some very large inception rounds reflect irrationality. For limited partners, he emphasizes entry price and ownership, and expects new firms to emerge alongside established brands. These views align with Boldstart's own positioning.

Asked what might reopen IPO markets, Sim points to interest-rate expectations and cites historical forward valuation multiples for high-growth companies. He speculates about a future range of 10 to 14 times forward revenue for the strongest businesses; his figures and forecast are historical opinions, not current valuation guidance.

Sim wants Boldstart to remain focused on inception investing rather than expand into a large asset manager. He plans to add operating partners to help early founders, while retaining a concentrated approach. He says the firm slowed new deals when it saw excessive market activity and later increased activity as familiar founders began new companies.