Jake Saper on Zoom, AI and the Hard Decision of When to Sell

Emergence Capital general partner Jake Saper joins Harry Stebbings to discuss Zoom, fund construction, AI and the difficult decision of when to sell. Saper speaks as an investor in many of the companies he describes, and the results he cites are his firm’s account.

Before Saper joined Emergence, the firm had developed an investment thesis that video conferencing could improve on Webex. His interview assignment involved evaluating an early competitor, so he arrived at Zoom with useful knowledge of the category.

Zoom had around $2 million to $3 million in revenue and was growing quickly. Saper says founder Eric Yuan’s technical experience and the reliability of the product impressed the team, including a partner who used it to call family in Argentina.

The proposed investment was unusually large for Emergence: a $20 million check from a $250 million fund at a roughly $200 million post-money valuation. The concentration and price mattered because the business was still at an early stage.

While building the financial model, Saper and a colleague found that Yuan had calculated churn too pessimistically, treating some upgrades and pauses as lost customers. They shared the discovery before final negotiations, which Saper says helped establish trust.

Saper says Zoom was profitable and did not have to raise, making the round competitive. Emergence committed to helping build an enterprise sales motion alongside product-led growth; he presents that support as a reason the founder chose the firm.

Experienced founders and younger founders can both succeed, Saper says. An investor may help improve the odds through hiring, go-to-market advice or support, even when a founder is capable of building without depending on that investor.

A prepared investment thesis can help a firm recognize an opportunity quickly, but Saper rejects the idea that every great venture return comes from a fully formed thesis. He describes learning from an initial Chorus investment and applying that experience to later voice-AI opportunities.

Saper recalls that Emergence’s Chorus investment returned more than three times capital and probably more than five, but says he does not remember the exact figure. He distinguishes a deal multiple from the amount it contributes to a whole fund.

He says the third fund has distributed roughly 16 times paid-in capital, with Zoom accounting for more than ten fund multiples and Salesloft also contributing materially. These are his retrospective fund and deal figures; DPI measures cash distributed relative to paid-in capital, not merely a private valuation.

Saper argues that both winner-take-all and markets with several substantial competitors can produce investment returns. He cites Emergence’s investment in Doximity as a network-effect example, and views the website-building market as potentially able to support multiple large players.

He expects some established website providers to reinvent themselves and others to consolidate or be acquired by private equity. He cautions that private-equity purchases are not automatically a rescue for venture investors who bought at much higher prices.

Saper describes Guru, where he has served on the board, as a company that added AI-enabled search to knowledge management, restructured and returned to growth. His description is a portfolio insider’s account, and the future value of that position remains uncertain.

The firm assesses whether a company can become an outlier rather than assuming every fast-growing business will do so. The founder, market and evidence of customer demand all matter, and their relative importance changes with the stage and opportunity.

Saper says some attractive deals remain non-consensus, while many are competitive. A distinctive insight or an early relationship can help an investor, but price and access alone do not establish that an investment will return a fund.

Emergence uses a “what you have to believe” framework: identify three to five assumptions that would need to hold for a particular investment to return the fund. Those can include dilution, market size, competition, team quality and the capital a company will need.

During diligence the team collects evidence both for and against each assumption. The resulting analysis is meant to expose what must be true, not to turn a highly uncertain future into a guaranteed calculation.

Saper describes Emergence as a focused B2B software investor. Partners make relatively few investments on average, allowing the firm to involve the whole partnership in research and, he says, provide continued support afterward.

He says none of his own investments had gone to zero so far but readily acknowledges that some still could. He considers whether the record reflects insufficient risk-taking; it is a time-specific portfolio statement, not a promise of downside protection.

According to Saper, nine in ten of the firm’s early-stage investments raised a subsequent round, one in five raised at a valuation above $1 billion and one in ten went public. These are firm-reported progression metrics, not cash returns or a causal proof of the firm’s contribution.

When a deal becomes a priority, every partner makes reference calls and some visit the company. The goal is to assemble firsthand evidence, compare different impressions and reach a collective decision rather than have one sponsor defend the case alone.

The process requires coordination, but the firm can conduct many calls in parallel. Partners share detailed notes and nightly summaries, and Saper says trust is necessary when one person asks the entire team to prioritize a deal.

For Regal, Saper recalls meeting a founder outdoors near Denver airport during the pandemic after a competitive financing process. The in-person meeting and subsequent calls with partners helped establish mutual understanding before Emergence invested.

He admits that a high-touch process will not appeal to every founder. Stebbings worries it could cause the firm to miss a major winner; Saper says some founders nevertheless choose that model because they want sustained partner involvement.

Saper says AI demand has enabled some companies to grow far faster than the earlier SaaS benchmarks. He cites portfolio companies Bolt and Together, including his reported growth figures, while noting that rapid revenue does not by itself establish durability.

He suggests pairing rapid growth with strong net dollar retention, such as 120%, as one way to evaluate whether existing customers expand over time. He stresses that many young AI businesses do not yet have enough years of retention data to establish a durable pattern.

The conversation explores where AI value may accrue across models, applications and services. Saper’s investment perspective favors understanding the customer’s problem, distribution and the economics of delivering a product rather than assuming the fastest-growing layer captures all the value.

AI-enabled services introduce a pricing and margin question. A company may win customers by offering a cheaper service, but if automation does not work as expected, the cost of delivering it can undermine gross margins.

Saper says he has become less convinced that incumbents will capture most AI value. Focused startups can move quickly on a narrow problem, though Stebbings notes that large companies are also shipping new capabilities rapidly; the balance remains uncertain.

He cites Veeva, an Emergence investment, as an example of starting with a narrow pharmaceutical CRM market and expanding into more products for the same customers. He believes deep relevance to a buyer can create opportunities to broaden the relationship.

On software capturing labor spending, Saper says customers he has spoken with often grow without adding staff rather than immediately laying people off. That is an observation from particular conversations, not a measurement of AI’s overall employment impact.

He considers Salesforce’s distribution and position in core workflows meaningful advantages. He also speculates that AI could help move legacy applications from mainframes to the cloud, a view about potential disruption rather than an established financial outcome.

Saper predicts a possible period of AI disillusionment as experiments fail to deliver lasting value. He also warns that an autonomous agent could take a harmful action and provoke a backlash, arguing that products need appropriate guardrails; this is a risk scenario, not a reported incident.

Emergence develops partners internally, Saper says, and seeks to avoid leaving founders with an investor who departs soon after a financing. He describes the risk of “orphaned” investments when the original sponsor leaves and the replacement has less commitment.

He argues that a multistage investor’s decision not to follow its small seed check can signal concern to the market. As a focused early-stage investor, he has a competitive interest in this argument, and acknowledges that financing structures differ.

Saper says Emergence deployed slightly less than $2 billion and returned a little over $8 billion in cash across its history. He notes that private holdings are additional; the cash figure and unliquidated portfolio value should not be added as though both were already distributed.

Emergence sold Salesforce shortly after its IPO, a decision Saper says forfeited substantial later upside. He was not at the firm for that sale and uses it as a reminder that the exit decision can matter as much as the initial investment.

For public holdings, the team reviews positions after quarterly earnings and considers whether to hold, sell or occasionally buy more. Saper says board membership offers context but also limits trading to legally permitted windows; possessing information does not permit trading on material nonpublic information.

He cites Doximity as a case where the firm bought more at IPO and reports an approximately threefold return from that purchase. This is his description of a particular investment, not a fund-wide result.

Saper describes an internal comparison of recent public holdings: selling at lockup expiry would have returned about $2 billion less than the firm’s actual sales, while selling every holding at its individual peak would have returned about $2 billion more. The analysis explicitly excluded Salesforce and the peak scenario is hindsight, not an executable strategy.

The firm’s follow-on checks can add substantially to returns, but Saper acknowledges that some later investments went badly. He calls fast growth without a coherent customer use case or sustainable margins “mirage” product-market fit.

For difficult reserve decisions, Emergence may bring the founder back to the full partnership, allowing others to challenge the deal sponsor’s attachment. Saper recalls a bridge investment in Intacct that helped the company survive and later sell, while acknowledging it is a selected example.

He guesses IPO activity may return the following year, while citing macroeconomic and political uncertainty. That is a prediction from the time of the interview, not a known schedule.

In the quick-fire round, Saper chooses Microsoft as a hypothetical ten-year holding and identifies several stage-focused firms he respects. These are personal preferences rather than investment recommendations.

He discusses Ironclad, a company in which he later invested after missing an earlier round. Saper explicitly acknowledges his bias when arguing for further upside; the valuation scenarios discussed are not realized returns.

His most disappointing exit was Comfy, whose bookings grew but whose deployments were hampered by different incentives for buyers and implementers. He says the company was sold to Siemens and the investment returned a small profit.

Asked to choose among AI model companies at hypothetical valuations, he favors Anthropic conditionally and expresses reservations about the others. These are personal judgments on uncertain future performance, not evidence that any valuation is correct.

Saper recalls winning an investment in Assembled after a mock board meeting with its founders. The example illustrates how the firm tries to demonstrate board support before a financing, alongside the competitive work needed to win a deal.

He says Emergence’s founders gave up their carry when retiring, creating a path for younger partners to become equals. He credits that structure with helping the firm retain people, while describing his own experience rather than a universal model.

Saper closes with optimism that better AI tools could reduce repetitive work and leave more room for creativity and human connection. His hopes about future capabilities and work are forecasts, not assured outcomes.