Figma’s IPO, Executive Pay and the Price of AI Growth

In this August 2025 roundtable, Harry Stebbings, Jason Lemkin and Rory O’Driscoll are joined by HubSpot co-founder Brian Halligan. They examine Figma’s IPO, executive compensation, public-market liquidity and AI economics before discussing private financing and venture-fund strategies. Their figures, reports and forecasts are presented in their original time context.

Stebbings opens with Figma’s large first-day stock increase and asks whether the IPO was severely underpriced. The guests distinguish the offering price, the later trading price and the limited amount of stock available. They use differing approximate prices in the conversation; the episode does not establish an independent reconstruction of the order book.

Halligan recounts HubSpot’s own IPO: after a demanding roadshow, founders chose both the price and allocations from an oversubscribed book. He says bankers and the company could have different preferences between hedge funds and long-term institutional investors. This is his experience, not firsthand knowledge of Figma’s private deliberations.

Halligan says HubSpot chose $25 despite being told a major institution preferred $24. He stresses that founders ultimately decide and says he does not believe there was collusion. The example illustrates a negotiation between price and the desired investor base, not an allegation of wrongdoing.

Halligan believes Figma had limited shares for sale and strong demand. He says a company may accept a modest initial increase to attract investors it hopes will hold for years. O’Driscoll separates that familiar tradeoff from an unusually large first-day move, which raises a different question.

Lemkin recalls an employee-directed share program at an earlier company, saying the IPO increase created meaningful gains for employees who could afford to participate. He distinguishes that personal experience from the separate question of whether an offering was priced efficiently.

O’Driscoll challenges the claim that Figma could simply have sold all its IPO shares at the later trading price. He argues the actual order book may not have supported such a price before trading began. This is his analysis of market mechanics, not an inspection of Figma’s confidential bids.

O’Driscoll speculates that some long-term investors might sell after an unusually rapid rise exceeds their price targets. Halligan doubts many of the large long-only investors sold and says allocations can be a first step toward a longer holding. Neither establishes what specific Figma investors actually did.

Halligan says HubSpot’s pre-IPO meetings helped it secure institutional investors and that firms it did not meet took years to join. He regards a stable shareholder base as valuable, while acknowledging that timing and luck also influenced the outcome. His account reflects his role and continuing interest in HubSpot.

Asked whether a direct listing would have solved Figma’s price jump, O’Driscoll says direct listings can have different pricing mechanics but might still begin far below the subsequent peak. Halligan worries that bypassing a traditional process can make it harder to build desired institutional relationships. Both present opinions and counterfactuals.

The panel discusses how sentiment can change between pricing and trading, and why a sudden rush of demand may produce an outsized move. O’Driscoll compares this with private rounds following a prominent investor’s participation. Similar-looking price changes do not establish the motives of every buyer.

Halligan contrasts HubSpot’s IPO timing with Zendesk’s, suggesting that the investor base assembled at listing can matter years later. He explicitly says timing involved luck. The guests discuss activist pressure as one possible factor in a company’s trajectory without claiming a single cause for Zendesk’s sale.

Halligan describes the exhaustion of pricing night, the ceremony and the wait for an opening trade. At HubSpot, he recalls the excitement of seeing the company reach a billion-dollar market value. His story shows how operational decisions and the emotional weight of an IPO coincide.

Asked about dilution, Halligan says his personal wealth rose substantially and he was happy to reach the milestone. O’Driscoll observes that founders usually experience an IPO once while bankers run the process repeatedly. Their reflections describe incentives and information differences, not a complete account of any specific offering.

The discussion moves to Figma CEO compensation. Halligan favors more performance-linked awards and argues that conventional benchmarking against peer CEO pay may not motivate an already wealthy founder in the same way. The figures and package details discussed are their contemporary interpretations; potential awards are subject to their actual terms.

O’Driscoll explains the distinction between restricted stock units and performance stock units. He prefers linking awards to measurable operating outcomes, but notes that disclosing targets and changing circumstances can complicate multiyear plans. Stock-price milestones can reward performance, yet may also be affected by broader market movements.

The guests discuss whether a sharp post-IPO stock move could satisfy some price conditions faster than expected. Halligan notes that time-based vesting can still apply. The conversation is not a verification that all conditions for any particular award were met or that a headline award was already earned.

Halligan says HubSpot uses net new annual recurring revenue and an earnings floor in compensation. O’Driscoll likes the link to operating metrics, while both acknowledge that disclosure can have side effects. Halligan is describing his company’s approach, not a universal compensation prescription.

Lemkin says he has seen growth investors offer founders large contingent awards tied to high future valuations. O’Driscoll worries that market-price targets can become detached from operating progress and may later be renegotiated; Lemkin is less convinced. They are debating possible outcomes, not confirming future changes to any grant.

Halligan supports limited founder secondary sales, recalling that he sold some HubSpot shares before its IPO. He says the liquidity helped him feel more confident about remaining independent when an acquisition offer appeared. He also says exceptionally large early sales can raise different questions.

Asked whether Canva should prepare to list after Figma’s debut, Halligan says market windows are seasonal and he would get ready. Lemkin notes that founder preferences, existing liquidity and profitability complicate that choice. These are outside opinions; the guests do not announce Canva’s plans.

Halligan says public ownership was less intimidating than many founders assume: he found many public investors rational when management communicated a credible long-term plan. O’Driscoll agrees that ordinary venture investors can also be demanding, although activist public investors can create a different kind of pressure.

Halligan describes an IPO as exhausting but emotionally meaningful for the team. Lemkin distinguishes the experience of continuing to run a public company from receiving cash in an acquisition. A quoted market value is not the same as all founders immediately receiving that amount in cash.

Halligan says he sells the same number of HubSpot shares each month and receives additional equity over time. His shareholding and executive role are relevant to his comments on public ownership. He describes his own practice, not a trading plan appropriate for everyone.

O’Driscoll discloses Scale’s investment in HubSpot’s Series C after his earlier pass, and Halligan recalls raising during the 2009 downturn. They say many firms declined before Scale offered terms. Their shared investment history is relevant when they discuss HubSpot’s success.

The panel asks how HubSpot built a large business serving smaller companies. Halligan says he wanted to avoid traditional enterprise sales and believed the internet gave small businesses new ways to compete. He says customer acquisition and lifetime value were central to the model.

Asked whether the same playbook works in AI, Halligan cautions against copying old advice uncritically. Lemkin says small businesses may not have the time or staff to configure complex AI systems. He sees an opportunity for products that deliver useful results without a lengthy setup.

O’Driscoll suggests AI applications for small businesses may emerge after larger customers help define workflows, with much of the product preconfigured. This is his forecast. The panel distinguishes consumer-oriented tools from enterprise deployments and the underserved businesses between them.

Stebbings cites Meta’s quarterly growth, adjusted earnings and a decline in free cash flow. O’Driscoll says strong existing businesses give large platforms room to spend on AI, but warns against assuming that every AI investment is already earning a return. The metrics are different measures and refer to the reported quarter.

Halligan says he is more comfortable seeing AI products adopted by customers outside the technology industry than relying solely on technology companies buying from one another. Lemkin worries about a capital-expenditure cycle. O’Driscoll believes the long-term trend may be real even if financing gets ahead of near-term demand.

O’Driscoll estimates a wide gap between annual AI infrastructure spending and application revenue and sketches a possible period of retrenchment followed by growth. His figures mix broad categories and are estimates, not comparable audited totals. The timing and severity of any correction are speculative.

The guests consider leadership at Nvidia, Meta and Microsoft. Halligan praises Jensen Huang’s approach to the CEO role; Lemkin credits Satya Nadella with navigating a major AI partnership as a non-founder CEO. These are judgments about leadership, not verified rankings or guarantees about future performance.

O’Driscoll interprets Microsoft’s OpenAI partnership as a way to participate in AI without building every model internally. The panel discusses both the strategic benefits and the financial costs as they understood them then. Ownership and accounting details are not independently established by the discussion.

After Halligan leaves, Stebbings raises a rumored Cognition financing at a $15 billion valuation following its Windsurf deal. The guests cite approximate company revenue and discuss reported employee departures and offered terms. The financing, staffing details and transaction motivations are not independently verified here.

Lemkin speculates that the acquisition may have offered Cognition a broader product and distribution platform, while O’Driscoll considers the cultural implications of combining teams. Their descriptions of employees’ choices and the deal’s rationale are interpretations of reports, not findings about individual behavior.

O’Driscoll says complicated acquisition structures can make it difficult for employees outside the main negotiations to know their position. The guests acknowledge they do not know all the terms or comparative outcomes. They focus on uncertainty and the human effect of a fast-changing transaction.

Stebbings cites Ramp’s reported $500 million Series E at a $22 billion valuation and asks whether frequent large rounds reflect business needs or investor demand. O’Driscoll says both may matter and notes that payment and credit businesses can have different capital requirements from software-only companies.

Lemkin observes that a large absolute funding amount can represent a relatively small percentage of a highly valued company. The guests caution that a small transaction’s valuation signal and an existing investor’s paper markup are not identical to cash returned to a fund.

Stebbings asks whether a high-price round can become a ‘suicide round.’ O’Driscoll says the danger is especially acute if a company raises too little, fails to grow into the price and must return quickly for more capital. An investor can overpay even when the company itself benefits from funding.

O’Driscoll sketches how card balances might create funding needs and says he does not know Ramp’s exact financing arrangements. His interchange, float and capital estimates are illustrative assumptions, not a statement of Ramp’s actual balance sheet or borrowing requirements.

The conversation moves to CRV’s reported fund and narrower strategy. O’Driscoll says a firm may choose to concentrate on the stage where it is strongest, while limited partners can support a range of strategies. He does not claim to know the firm’s internal decision process.

Lemkin discusses the tradeoff between an opportunity fund and a main fund, including fees, carried interest and time. He says his own opportunity fund may add less personal carry than a much larger vehicle would. His estimates about his own returns and motives are not projections for CRV.

O’Driscoll says an early-stage manager may have fewer later-stage opportunities that can absorb substantial additional capital than expected. A manager can develop a separate growth strategy or raise a vehicle for a particular opportunity, but each path changes the work and the risks.

Stebbings asks whether Benchmark’s fund size and personnel changes point to a competitive problem. O’Driscoll declines to infer individuals’ motives and reframes the issue as specialist versus full-platform strategy. Lemkin says brand visibility and investment performance are related questions, not the same metric.

O’Driscoll says a specialist can be crowded out by a larger firm’s reach, while a broad platform can overextend and dilute returns across too many investments. Either model can work if executed well. The essential task is to understand the strategy and its specific risks.

The guests end by suggesting possible future interviewees and reflecting on the different perspective Halligan brought as a founder and CEO. Their discussion throughout is a collection of personal experiences and August 2025 assessments, not independently verified predictions or financial advice.