Chime’s IPO, Scale AI’s Deal and the Return of Public-Market Liquidity

Harry Stebbings welcomes investors Rory O’Driscoll and Jason Lemkin, plus Garrett from Handshake, to discuss Meta's investment in Scale AI, public listings and the financing market.

The panel begins with the reported $14.8 billion Meta–Scale transaction. Garrett says Handshake has seen a surge in demand as model labs consider diversifying their suppliers; he represents a company that could benefit from that shift.

Garrett says demand for Handshake's services tripled in the week around the announcement. This is his company-specific account, not independently verified market data.

O’Driscoll describes Scale's business supplying human expertise for AI training and evaluation, moving from simple data labeling toward more specialized work. The revenue figures he cites are estimates in the discussion.

The panel describes Meta's reported minority investment, the distribution of proceeds and the movement of some Scale executives to Meta. Its account of precise proceeds and deal mechanics is based on participants' understanding, not a complete transaction disclosure.

The investors suggest other model providers might hesitate to share sensitive work with a supplier linked to a competitor. This is a competitive concern and forecast, not proof that all customers have left.

Garrett says future training may require domain experts, audio, tool use and evaluations of agent trajectories. He sees quality, volume and turnaround speed as central requirements for customers.

Asked whether Scale will still have more than $100 million in revenue a year later, Garrett says yes. He also says substantial spending is looking for alternative providers; both statements are his prediction and market observation.

Stebbings interprets Meta's investment as a way to secure talent and signal an AI commitment. O’Driscoll agrees this could be more important than the near-term financial return, but they do not know Meta's internal rationale.

Lemkin compares the cost of recruiting an exceptional leader with Salesforce's earlier Quip acquisition. The guests stress how much larger technology-company balance sheets and the scale of these bets have become.

Asked about a hypothetical $4 billion offer for Handshake, Garrett says he would not accept it. The other guests caution that a hypothetical answer may differ from a real decision with actual terms.

Garrett describes Handshake's opportunity as an extension of its existing employment network. Lemkin sees a broader lesson in staying in business long enough to take advantage of a new market opening.

After Garrett leaves, Stebbings asks whether proceeds from Scale, Chime and Circle will encourage limited partners to reinvest. O’Driscoll expects more liquidity than in recent years, while noting that the venture market is very large.

Lemkin distinguishes cash distributions from public shares still subject to lockups and later sales. He says the limited partners he knows look at money actually returned rather than treating all paper gains as immediately available.

The panel considers the unusual structure of the Meta–Scale transaction and potential regulatory implications. O’Driscoll's comments about legal review and motives are his analysis, not a finding by regulators.

Lemkin predicts a severe decline for Scale; O’Driscoll agrees its customer relationships could be difficult, while noting that the company still has talented employees. These are opinions about the future, not established outcomes.

Stebbings discloses that he is an investor in Mercor and asks how spending might be distributed among providers. O’Driscoll thinks some labs could bring work in-house, while acknowledging that he could be wrong.

Stebbings cites estimated returns for early Scale investors. The panel's figures describe reported or estimated outcomes and do not establish the eventual cash distributions to each fund's investors.

The discussion turns to a public claim about Discord's CEO transition. Stebbings frames it as a claim; O’Driscoll says he does not know the exact circumstances, so the panel does not establish why the change occurred.

O’Driscoll says boards should set expectations and discuss concerns with a founder before making a leadership change. In his experience, many transitions can be approached cooperatively.

Both investors stress the risk and disruption of replacing a founder CEO. Lemkin cites his own analysis of business-to-business IPOs and argues that a forced transition can be especially difficult; his sample and interpretation are his own.

O’Driscoll says a board may nevertheless decide that a change is required by its obligations. The guests agree that no one can observe the alternative history of a company after such a decision.

Lemkin sketches a hypothetical Discord trade-off between retaining the founder and pursuing a higher valuation with a new leader. He says he would prefer the founder in that scenario; it is not a factual account of the board's choice.

The investors say founders should understand a prospective investor's approach to leadership changes. O’Driscoll favors retaining founders by default while allowing for exceptions.

Stebbings turns to Ramp's reported financing at a $16 billion valuation, comparing it with Brex and Mercury. The panel's valuation and revenue figures are contemporaneous reports and estimates, not comparable audited results.

Lemkin notes that a $200 million round at that valuation would cause relatively modest dilution. O’Driscoll asks a different question: whether the price leaves sufficient return for the incoming investors.

O’Driscoll describes Ramp as combining card spending and software and notes that its eventual valuation may depend on what happens as growth slows. He says the price feels high but does not claim to know the eventual outcome.

Lemkin argues that private markets sometimes apply similar revenue multiples to businesses with different gross margins and public comparables. The guests caution that fintech and software revenue can have different economics.

O’Driscoll points out that a card business can require capital to fund the period between merchant payment and customer repayment. The specific financing mix is not established in the discussion.

The panel compares Ramp, Brex and Mercury's products and customer relationships. They differ over how much value to assign to banking relationships and distribution.

They discuss how competitive financing rounds can encourage investors to pay more to gain access. The speakers view such terms as a sign of market demand, not evidence that every valuation will hold.

Stebbings cites a reported Pentagon contract with OpenAI. The guests discuss government procurement and the company's commercial relationships; the contract figure and scope are described as reports, not verified spending.

Lemkin argues that a large platform may want to serve many kinds of customers. The discussion treats government contracting and public positioning as business and policy choices, not settled moral conclusions.

The panel debates Microsoft and OpenAI's relationship. O’Driscoll says he does not know how their negotiations will end and cautions against assuming which party acted first or will prevail.

They discuss reportedly complex rights, profit sharing and an AGI-related trigger. Lemkin acknowledges he does not know the full contract; predictions about the definition or timing of AGI remain uncertain.

The guests speculate about possible restructuring and extended agreements. They do not establish final ownership, contractual rights or any completed renegotiation.

Stebbings says Chime's shares rose about 50% after its IPO and asks whether the market is reopening. This describes the trading period discussed in June 2025, not a current share-price claim.

O’Driscoll says quality companies offered at attractive prices can reopen a market after a quiet period. Strong early trading may draw investors back, although a first-day rise also raises questions about offering price.

The panel compares the recent market with losses some investors experienced after the 2021 IPO cycle. O’Driscoll sees improved sentiment as encouraging, but does not claim that every future offering will perform similarly.

Stebbings asks whether stronger IPO demand will bring companies such as Stripe or Databricks to market. O’Driscoll thinks larger companies may be able to choose their own timing and could still decide to wait.

On Chime's pricing, the guests discuss whether conservative advice contributed to the initial rise. They note that a company's proceeds, shareholder dilution and eventual investor sales are distinct questions.

Lemkin says a recent analyst presentation suggests Databricks is preparing for public-market scrutiny. He acknowledges limited direct knowledge and presents a near-term filing as a possibility, not a confirmed plan.

Stebbings cites a reported Gusto tender at a $9.3 billion valuation and roughly $900 million in annual recurring revenue. The speakers' figures are contemporaneous reports, not an announcement of an IPO date.

O’Driscoll says payroll is a large market with established public competitors. The guests use those companies as valuation comparisons, while recognizing that business models, growth and market prices differ.

O’Driscoll reflects on not investing in an earlier Gusto opportunity. He says he underestimated the founder, customers' willingness to switch and how quickly the price would rise.

Lemkin discloses that he has been a Gusto customer and appeared in a company case study. He likes its software but says older payroll providers also delivered a human service that reduced work for him.

The guests consider Dropbox and Glean as examples of incumbents competing with newer AI companies. Lemkin worries established vendors are moving too slowly.

O’Driscoll argues that an incumbent's strongest advantage is its existing customer base. A new provider may instead offer a product that works across multiple systems, changing how the two compete.

They discuss reported restrictions around Slack connectivity and whether an API fee or different rules might emerge. O’Driscoll's description of possible outcomes is speculation, not a confirmed policy decision.

On Slack's role within Salesforce, Lemkin gives an estimate of revenue and argues that the acquisition may still be valuable. The panel separates his commercial assessment from verified financial disclosures.

In a prediction-market round, the guests debate whether Apple will announce iPhone assembly in the United States. They distinguish an announcement from actual manufacturing and make speculative calls.

Asked whether the S&P 500 will end the year positive, O’Driscoll says the outcome is unknowable and considers the wager's odds. The guests disclose their own equity exposure; their comments are not investment advice.

The panel considers whether a Chinese AI model will lead a benchmark that year. They distinguish a temporary benchmark result from durable business leadership and make predictions, not factual claims about the future.

The guests close with a correction about Tableau's founding era and thank one another for the discussion.