Figma’s IPO Filing, Melio’s Sale and the Pressure on Software Investors

Harry Stebbings and investors Rory O’Driscoll and Jason Lemkin discuss Figma's IPO filing, Melio's sale, venture returns and the competitive pressures facing software companies in the AI market.

Stebbings cites Figma's filing: $821 million in revenue, 46% year-over-year growth, approximately $1.5 billion in cash and no debt. The speakers' figures reflect their reading at the time and should be interpreted using the filing's underlying periods and definitions.

The panel highlights positive free cash flow and a roughly 40% quarterly free-cash-flow margin. Adding growth and margin yields a rough 'rule of 80' calculation, not a single audited financial metric.

Lemkin considers a potential $20 billion to $30 billion valuation, noting that even $20 billion would be a high revenue multiple. This is a pre-IPO estimate, not the eventual offering price.

The guests compare an independent public listing with Adobe's abandoned acquisition. A cash sale would have offered different timing and certainty, while independence lets the team continue building its own company.

They argue the proposed acquisition could have expanded Adobe's reach into software design and development. Their assessment of strategic value and what Adobe should have done is opinion.

O’Driscoll notes that Figma serves developers as well as designers, according to figures the panel attributes to the filing. The expansion beyond one job title is central to his view of the product's value.

Lemkin says acquisitions can depend on an internal champion who may leave. Stebbings draws a parallel with portfolio companies whose venture partner moves on, potentially weakening their advocate inside the fund.

O’Driscoll says a board member who has left the investing firm may still be valuable, but the company also needs an advocate who can participate in decisions about allocating that firm's capital.

The panel debates whether the original deal partner or another partner should decide follow-on reserves. Stebbings values historical knowledge; Lemkin worries that deal sponsors may keep funding weak companies because of personal incentives.

O’Driscoll suggests combining a partner's recommendation with a robust group process. A finite reserve pool must balance supporting likely winners with protecting value in struggling investments.

Stebbings says the companies he initially expected to return his first fund were not the ones he now expects to do so. He questions whether seed investors can reliably identify their eventual winners early enough to allocate reserves.

O’Driscoll thinks the picture can be different after product-market fit, when more operating history is available. He describes his firm's historical analysis as a probabilistic guide, not a guarantee.

The guests discuss pay-to-play financings. O’Driscoll says distressed rounds rarely produce the best outcomes, although he cites FedEx as an example where a later down round worked well for investors.

Lemkin recalls a company that reached an IPO after a pay-to-play and a bankruptcy, leaving earlier investors with losses. The story illustrates why an eventual public listing does not by itself establish success for every investor.

O’Driscoll says the opportunity cost of supporting a struggling company includes time as well as cash. Personal commitment to a founder can matter even when a purely financial calculation would favor stopping.

He describes a founder who survived for years with a promise that the fund would cover shutdown costs if necessary. The eventual return is uncertain, but he values the founder's persistence.

Stebbings estimates that Index could receive around $3.5 billion across Figma and Scale-related outcomes, while noting holding and lockup constraints. The figure is an estimate, not cash already distributed to limited partners.

O’Driscoll describes a market with fewer but larger winners as companies remain private longer. Successful early investments can return substantial capital, but identifying and accessing those winners remains difficult.

Stebbings asks whether future exit sizes could justify larger funds. O’Driscoll warns against extrapolating a step change in company size indefinitely into the future.

The investors consider large AI funding needs and the case for larger funds. O’Driscoll says some capital-intensive opportunities can justify large checks, but investors may also pay prices that already anticipate exceptional outcomes.

Stebbings introduces Xero's proposed $2.5 billion acquisition of Melio as a meaningful exit. Lemkin finds it discouraging relative to the growth and revenue numbers he cites from Xero's presentation.

Lemkin cites $153 million in revenue and 127% growth and asks why a company at that trajectory would sell. O’Driscoll questions whether the trailing growth rate is sustainable; neither has established the company's future growth.

O’Driscoll discloses that his firm invested in Bill.com, a company in the bill-payment space. He sees strategic sense in Xero's deal in a crowded market, while acknowledging he had not read the same release.

The panel compares the proposed sale with Melio's earlier private valuations and discusses liquidation preferences. Their simplified estimates do not establish actual distributions to any particular investor or founder.

They note that investors with different entry prices and preferences can have different incentives when evaluating a sale. A preference that limits one investor's downside does not mean all shareholders receive the same return.

Lemkin distinguishes a private buyer's stock from cash or publicly traded consideration: a private valuation may be difficult to realize. The panel debates how those differences affect an acquisition's appeal.

Lemkin mentions another revenue figure from March and questions the sale again. The panel does not reconcile its different figures or establish a consistent ARR period, so the figures should not be treated as interchangeable.

The discussion turns to founder secondary sales. Lemkin says large investors sometimes encourage founders to take liquidity and argues existing investors may also want an opportunity to sell.

O’Driscoll notes that co-sale rights can provide such an opportunity, while the others point out that rights and waivers vary by agreement. Their exchange describes investor experience, not a universal legal rule.

Stebbings asks whether very large compensation could reduce AI researchers' focus. O’Driscoll argues wealth can reveal different motivations and should not automatically be assumed to demotivate everyone.

The panel considers whether spending on AI infrastructure will earn an adequate return. They agree that very large capital expenditure creates risk, while the timing and extent of any overinvestment remain unknown.

Stebbings asks whether limited partners are more persuaded by Menlo's realized history, such as Chime, or the future potential of its Anthropic investment. The guests say both track record and forward-looking opportunity can matter.

Lemkin says limited partners he speaks with understand how large outcomes must be to return large funds. O’Driscoll adds that past access to strong deals can help support a manager's case for future access, though it does not guarantee it.

Stebbings asks whether Couchbase's roughly $1.5 billion acquisition signals a broad return of private-equity buyers. O’Driscoll believes it may be a targeted bet rather than evidence that every similar software company will find a buyer.

The panel cites roughly $215 million in revenue and modest growth for Couchbase. O’Driscoll notes that a particular product or AI thesis could matter to a buyer; he does not claim to know its final investment results.

Lemkin says two of his portfolio companies received mediocre acquisition offers but had not heard from private-equity firms. He presents these as a small anecdotal sample, not a market-wide survey.

Stebbings asks why more software businesses are not consolidated like consumer subscription products. O’Driscoll points to roll-up models, but says deals require sellers and buyers to agree on a price.

The guests compare possible acquisition multiples. Lemkin says some buyers target around twice revenue, while sellers may hope for more; these are examples from the discussion, not valuations for every company.

Stebbings says Grammarly might assemble a productivity suite through acquisitions and discloses that he owns Grammarly stock. His optimism about its future is an investment view.

Lemkin is skeptical that many such roll-ups will occur merely because they make sense on paper. O’Driscoll agrees that a gap between buyers' and sellers' prices can prevent deals.

Lemkin argues that established software companies need to show growth from AI, not just add a feature. His June 2025 deadline is a strongly held investment judgment, not an objective test of every company's future.

The panel debates whether AI budgets expand the overall software market or move spending between vendors. Lemkin cites a set of company growth metrics but acknowledges the comparison does not prove causation.

Stebbings asks whether technology can capture spending previously allocated to labor. He also discloses an investment in a company he compares with Superhuman; his claim about its projected revenue is a forecast.

Lemkin cites Clio's acquisition of vLex as an example of an older company becoming relevant to AI. He views the transaction as evidence that incumbents can adapt, not proof that every older business has done so.

The conversation turns to a reported $30 billion Oracle–OpenAI deal. The guests use the reported amount to discuss Oracle's infrastructure strategy, without establishing the complete contractual terms or timing of payments.

O’Driscoll says Oracle shifted resources toward AI infrastructure. He cautions that the economics depend on sustained demand and the continuation of the investment boom, both uncertain.

Stebbings mentions reports about Surge AI fundraising and discloses an investment in Mercor. The panel speculates about opportunities for data providers following Meta's investment in Scale AI.

The guests express sharply negative opinions about Scale's prospects and speculate that other suppliers could gain business. Their comments do not verify customer departures, revenue transfers or the future of Scale.

They discuss reported Surge revenue and its low-profile approach. O’Driscoll offers a hypothetical explanation for its growth while making clear he is assuming, rather than reporting, its customer mix.

The panel turns to CEO transitions, including a move from LaunchDarkly to Asana. The investors discuss the pressures of building companies over long periods rather than claiming to know an individual's private motives.

O’Driscoll says a journey expected to last four to six years can stretch to more than a decade. Exhaustion and changing personal circumstances may lead a founder to appoint a president, change roles or seek liquidity.

The guests debate winner-take-most dynamics: extremely large outcomes can raise the cost of missing a leading deal and encourage investors to pay more. They recognize the resulting risk to fund returns.

In a quick-fire round, the guests predict whether Elon Musk will form a political party and discuss a founder's possible future paper wealth. Their responses are speculative, based on assumptions rather than known future events.

The panel closes after distinguishing paper wealth from realized cash and thanking the guests. Their quick-fire predictions reflect views at the time of the recording.