NEA's Tiffany Luck on AI IPOs, Personal Agents, and the ROI Reckoning

Tiffany Luck, a Partner at New Enterprise Associates focused on early-stage AI, APIs, and B2B SaaS, appeared on TechCrunch's Equity podcast on June 17, 2026, to discuss AI IPO prospects, the emerging personal agents category, and how enterprise buyers are beginning to demand measurable returns on AI investment.

Luck joined NEA in 2023 after a prior tenure as a Partner at GGV Capital. She sits on several portfolio company boards, giving her a ground-level view of how AI spending decisions are playing out inside real organizations — not just in pitch decks.

The ROI Reckoning

The timing of the podcast aligns with a growing tension in enterprise AI adoption. Uber burned through its entire 2026 AI budget by April of this year — exhausting roughly a full year of planned spend in four months, largely on Claude Code usage. By early June, the company had moved to cap individual employee AI spending. Uber's COO acknowledged the difficulty of drawing a direct line between rising token costs and customer-facing outcomes — a candid admission that many finance and engineering leaders are wrestling with privately.

Uber is not alone in confronting this. Meta built an internal leaderboard it called "Claudeonomics" to track employee AI tool usage, then shut it down in April 2026 after data from it began circulating publicly. The episode illustrated just how sensitive internal AI consumption metrics have become — organizations are simultaneously struggling to quantify value and reluctant to have the conversation in the open.

This is the environment in which Luck's "ROI reckoning" framing lands. Enterprise buyers who moved fast in 2024 and early 2025 — often without formal measurement frameworks — are now facing renewal cycles and budget reviews where the question is no longer "are you using AI?" but "what did you get for it?"

Worth flagging: from an investor's vantage point, this inflection is not necessarily bad news. Startups that can credibly instrument and demonstrate productivity lift — lines of code shipped, tickets closed, deal cycles shortened — will have a structural advantage in a market where buyers have become more discerning. Luck's focus on APIs and B2B SaaS suggests she is backing companies that sit precisely in this measurement layer.

Personal Agents and the IPO Pipeline

The personal agents category is moving from research curiosity to product category. Where earlier AI assistants were largely reactive — answering queries, summarizing documents — personal agents are expected to act on behalf of users across multi-step workflows, maintaining context and handling authorization chains autonomously. The infrastructure requirements are non-trivial: reliable tool-use, memory management, permissioning models that enterprises will actually accept, and latency profiles that don't erode the user experience.

For a venture investor whose portfolio spans early-stage AI and API infrastructure, the agent layer is naturally interesting — it is a forcing function for new API surface area, new orchestration primitives, and new identity and security tooling. What Luck thinks about the near-term IPO pipeline for AI companies is a signal worth watching. The 2025 market saw a handful of AI-adjacent listings, but a genuine cohort of pure-play AI software companies going public would represent a meaningful datapoint for how public investors are prepared to value inference-dependent revenue at scale.

What Founders Should Know

Luck gave a preview of her investment lens at TechCrunch's All Stage event in Boston in 2025, where she outlined five questions she expects founders to answer in a pitch. The Equity appearance extends that thinking into the current market — where the questions of distribution, measurement, and defensibility have sharpened considerably since early generative AI enthusiasm peaked.

The broader context here is that the venture community is itself recalibrating. The cohort of AI companies funded in 2022–2024 is approaching the point where some will need to demonstrate durable unit economics rather than usage growth. Luck's portfolio focus on B2B SaaS and API infrastructure puts her in close contact with that reckoning — she is backing the picks-and-shovels layer at the same moment enterprises are scrutinizing whether the gold rush justified its costs.

The Equity podcast episode is available via TechCrunch.