Why Data Is the Next $1 Trillion Market

Episode Summary

Executive Summary: The roundtable focused on how venture is becoming increasingly power-law driven, with liquidity concentrated in a small set of elite private companies, while secondary markets, M&A, and AI data access reshape startup economics. The guests debated SPV/secondary transparency, the importance of proprietary data, the role of model pricing in margins, and whether SaaS firms can adapt—or get acquired—in the AI era.

Main Topics: Secondary markets, SPVs, and transparency (Priority: 5/5): The hosts examined the USVC/Anduril dispute as a window into the messiness of secondary transactions, highlighting SPV stacking, weak disclosure, and the lack of public-market-style guardrails in private markets. Power-law concentration in venture (Priority: 5/5): Both guests argued that venture capital and LP attention are increasingly concentrated in a tiny set of elite companies, creating pressure to win exposure to only a handful of assets and raising the stakes for secondary liquidity. M&A as an increasingly important exit path (Priority: 4/5): The discussion covered GPT-0's sale to Superhuman and broader acquisition activity, suggesting that strategic buyers are using equity currency to buy talent, product, distribution, and workflows in a re-rated software market. Data as a core AI substrate (Priority: 5/5): The speakers argued that data—not just compute or energy—is becoming a key bottleneck and opportunity in AI, especially for companies that can license proprietary data or leverage unique operational datasets. SaaS resilience vs AI disruption (Priority: 4/5): They debated whether AI is killing legacy SaaS or actually strengthening it by making proprietary workflows and customer data more valuable, with examples like Intercom, Box, Salesforce, and Monday. AI infrastructure and market risk (Priority: 4/5): The conversation shifted to the fragility of the AI capex boom, including dependence on semiconductor growth, hyperscaler spend, private credit, and whether a slowdown could trigger a broader market correction. China open-model restrictions and startup margins (Priority: 3/5): The panel discussed how possible restrictions on open-weight model exports from China could constrain startup model choices, but noted that cheaper closed-source frontier-adjacent models and orchestration layers may partially offset the pressure.

Key Arguments: SPVs and secondary deals can be used legitimately, but when layers stack and ownership is unclear, the market becomes vulnerable to confusion and possible fraud. Private markets increasingly function like public markets for a tiny set of companies because companies such as SpaceX and Stripe provide regular tender offers and liquidity windows. Secondary activity is no longer taboo; it is now an accepted tool because distributions from venture have been scarce, though the gains will accrue unevenly to a small set of funds and LPs. Early-stage investors can often sell small amounts without stigma, but lead investors selling large positions can send negative signals and trigger concerns about misalignment. Acquisition offers increasingly matter because they may be equity-based, so valuation depends on what the acquirer’s stock will be worth in the future, not just today’s headline price. Data licensing and proprietary datasets are emerging as a major AI business opportunity, with companies like Protege helping monetize access to valuable data sources. The best AI businesses may be those with proprietary workflows, customer relationships, and data, since AI can amplify those assets rather than replace them. The AI market remains vulnerable because it depends on sustained capex, compute availability, and a few hyperscalers hitting very high expectations quarter after quarter. If Chinese open-weight models become less available, startups may be pushed harder toward closed-source frontier models, self-hosting, or orchestration strategies to manage costs and avoid data leakage. Smaller venture funds can benefit more from moderate exits because a $500 million outcome can be meaningful to a $225 million fund even if it barely moves the needle for a billion-dollar fund.

Data Points: Footwork last fund size: $225 million - Nikhil Basu Trivedi described Footwork’s latest fund size. GPT-0 revenue: tens of millions - Nikhil said GPT Zero reached tens of millions in revenue before being acquired. GPT-0 profitability: never burned a dollar of invested capital - Nikhil said the company was lifetime profitable and had more cash than it had ever raised. Sandana secondaries return target: 2x to 3x - Michael said secondary funds can reasonably return 2x, 2.5x, or 3x, though not 5x-10x. Example fund impact: $75 million return - A $500 million exit with 15% ownership would return about $75 million to a couple-hundred-million-dollar fund. Chat of SaaS valuation: 3.6x-3.7x forward revenue - Michael cited Meritech’s index for enterprise software multiples. Monday.com valuation multiple: ~2x - Used as an example of public SaaS compression despite strong businesses. Box valuation multiple: 3.4x trailing price/sales - Alex cited Box as an undervalued public-market example. Salesforce growth: 10% to 12% - Michael noted Salesforce’s growth rate while discussing valuation compression. NVIDIA forward P/E: 24x - Michael used this to argue that the market is less absurd than in the 1999 bubble, though still demanding. Samsung EBITDA: $58-$59 billion - Referenced as evidence of huge semiconductor profitability and market surprise. Etched funding: $800 million - Alex mentioned Etched’s large funding round and upcoming chip/data center plans. Sama/Samova post-money: $11 billion - Alex cited a recent financing in the chip/inference ecosystem. Protege revenue: hundreds of millions in year two - Nikhil said Protege has scaled rapidly in data licensing. DigitalOcean cost reduction: up to 50% - Sponsor claim about AI native cloud savings. Workato metrics: 67% lower inference cost, 79% lower latency, 2x faster to production - Sponsor ad example used to support AI-native cloud positioning.

Pivotal Quotes: "The bigger companies are getting bigger can coexist with it, it is easier than ever to start a company, and it is more possible than ever for two people to build a business that gets big quickly and gets wildly profitable quickly." — Nikhil Basu Trivedi: On why power-law concentration and startup accessibility can both be true at once. "If you're going to be buying into one of the six companies that Nikhil's talking about, you have to have some confidence that ultimately down the road, you can actually get liquidity yourself." — Michael Kim: On the need for liquidity and tender offers in private markets. "The fundamental job we all have is to back those companies that are power-law companies." — Nikhil Basu Trivedi: On why early-stage venture must identify outliers early rather than chase consensus names.

Implications: Expect more M&A, more secondary activity, and more value concentrated in data-rich, AI-adapted companies. Startups need proprietary data, capital discipline, and model flexibility to stay competitive as liquidity and power cluster around a few winners.