Skip to content
The Twenty Minute VCThe Twenty Minute VC

Howie Liu: Decoding Airtable's $11B Valuation; The Impending AI Revolution in Enterprise | E1053

Howie Liu is the Founder and CEO @ Airtable, the fastest way to build apps for your business. To date, Howie has raised over $1BN with Airtable with the last round valuing the company at $11BN and an investor base including Benchmark, Thrive, Caffeinated, Greenoaks and Coatue to name a few. --------------------------------------- (0:00) Personal Insights and Product Strategy (06:03) AI and Technology (30:52) Business and Growth (43:08) Insights and Reflections (46:41) Quick-Fire Round --------------------------------------- In Todays Episode with Howie Liu We Discuss: 1. Scaling into Enterprise: What are the single biggest challenges when moving from PLG to enterprise? Why does Howie believe you have only truly hit enterprise when you sign $1M contracts? How long did it take for Airtable to sign their first $1M ARR contract? How can founders know when is the right time to scale into enterprise? How does the product need to change with the scaling? 2. Enterprises: Do They Really Love AI: Why does Howie believe that enterprises are not jumping on AI yet? When does enterprise interest turn into enterprise buying and purchasing? What are the single biggest barriers to enterprises buying AI solutions today? Post-purchase, what are the biggest implementation challenges for enterprises with AI? 3. The Changing Sales Process: Are we seeing the bundling of tools within large enterprises today? Which categories and vendors are most vulnerable? Which will survive the cuts? What do vendors need to do to prove to CFOs that they need to remain in their budget? How has the customer success process changed over the last year with tightening budgets? 4. Howie Liu: AMA: Airtable famously got Benchmark to lead their Series C, how did this come to be when they famously always only do Series A? Why does Howie believe that it is total BS to suggest post-PMF, everything is good? What does Howie know now that he wishes he had known when he started Airtable? --------------------------------------- Subscribe on Spotify: https://open.spotify.com/show/3j2KMcZTtgTNBKwtZBMHvl?si=85bc9196860e4466 Subscribe on Apple Podcasts: https://podcasts.apple.com/us/podcast/the-twenty-minute-vc-20vc-venture-capital-startup/id958230465 Follow Harry Stebbings on Twitter: https://twitter.com/HarryStebbings Follow Howie Liu on Twitter: https://twitter.com/@howietl Follow 20VC on Instagram: https://www.instagram.com/20vc_reels Follow 20VC on TikTok: https://www.tiktok.com/@20vc_tok Visit our Website: https://www.20vc.com Subscribe to our Newsletter: https://www.thetwentyminutevc.com/contact --------------------------------------- #HowieLiu #Airtable #HarryStebbings

Howie LiuguestHarry Stebbingshost
Aug 25, 202356mWatch on YouTube ↗

CHAPTERS

  1. 0:00 – 1:51

    Founder hindsight: aligning product strategy with go-to-market from day one

    Howie reflects on what he would tell himself before starting Airtable: product-market fit alone isn’t enough. The real leverage comes from designing a product strategy that naturally pairs with an effective go-to-market model.

    • Advice is clearer in hindsight, but the recurring lesson is GTM-product alignment
    • Different products demand different GTM motions (viral/organic vs outbound vs performance marketing)
    • Airtable initially over-focused on product quality and under-weighted early GTM design
    • Great product can buy time, but eventually distribution/GTM becomes decisive
  2. 1:51 – 3:38

    What Airtable would have emphasized earlier: team-scale use cases and templates

    Pressed on what he would actually change, Howie explains he would have guided adoption toward larger, team-centric workflows earlier. He ties team-centricity directly to monetization and to unlocking more scalable GTM options.

    • Airtable invested early in real-time collaboration infrastructure to support teams
    • Templates initially spanned solo and team use cases; he’d bias harder toward team-scale
    • Broader, unguided adoption created diffusion; clearer guidance could have accelerated enterprise value
    • Team and larger-team use cases monetize better and support outbound/paid GTM economics
  3. 3:38 – 5:43

    Horizontal vs vertical PLG: when to let a thousand flowers bloom

    Harry asks how horizontal PLG companies should choose between broad adoption and focusing on a few verticals. Howie argues the answer depends on how self-evident the product category is and where the market is on the “crossing the chasm” journey.

    • Some “horizontal” tools are effectively a clear category (e.g., Slack, Dropbox)
    • For less self-evident products, early verticalization helps people understand ‘what it’s for’
    • Go deep in a few use cases while preserving long-term platform optionality
    • The right answer emerges when you get concrete about use case, roadmap, and GTM model
  4. 5:43 – 8:44

    Why GenAI could be even more profound than cloud computing

    Howie compares cloud’s mostly ‘binary’ shift (on-prem to cloud) with GenAI’s compounding expansion of use cases. He argues GenAI’s breadth across knowledge work and its accelerating capability curve make it uniquely disruptive.

    • Cloud benefits are straightforward (scalability, less ops), and realized quickly after migration
    • GenAI targets a broad span of knowledge work across functions (legal, finance, marketing, creative)
    • Even today’s outputs can be ‘useful’ at scale; capability improvements likely unlock more
    • GenAI disruption unfolds piece-by-piece across industries, potentially larger in total impact
  5. 8:44 – 10:54

    Enterprise AI is still early: education, limitations, and the path to real use cases

    Drawing on Fortune 500 conversations, Howie says many enterprises are still learning what LLMs can and can’t do. He highlights hallucinations/accuracy, need for citations, and the broader “education phase” before truly strategic deployments take off.

    • Enterprises are still mapping AI affordances, limitations, and safe application boundaries
    • High-accuracy use cases (HR/wealth/legal info retrieval) require sources and correctness
    • Vocabulary and primitives (LLMs, vector DBs) are starting to appear in C-suite discussions
    • Once baseline understanding rises, enterprises can apply AI more intelligently to specific problems
  6. 10:54 – 16:26

    Adoption dynamics and labor fears: what happens when employees resist AI

    Harry raises employee pushback driven by job displacement fears. Howie responds with a sober view of economic disruption, while also arguing AI can lower production costs, raise demand, and potentially expand employment through augmented productivity.

    • GenAI spreads via consumer adoption, not just enterprise rollouts, increasing accessibility
    • Employee resistance can be rational when automation threatens roles (media/content, creative teams)
    • Near-term economic disruption may arrive before society adapts
    • Analogy to PCs: short-term displacement, long-term productivity gains and new employment patterns
  7. 16:26 – 19:31

    What blocks enterprise GenAI adoption: privacy, deployment, IP risk, and safety

    Howie details why even motivated enterprises hesitate: data privacy and hosting constraints, copyright and plagiarism concerns, and the challenge of getting accuracy and safety high enough for serious workflows.

    • Education remains a bottleneck, but governance concerns loom large
    • Many enterprises want self-hosting/on-prem options; hosted-only models can be a barrier
    • Copyrighted training data and plagiarism risk matter for public-facing content generation
    • Safety, reliability, and accuracy thresholds are decisive for production deployments
  8. 19:31 – 21:12

    Services firms as AI winners: the integration ‘handholding’ era

    Harry proposes that services companies will be major winners by integrating AI into enterprises. Howie agrees they’ll be critical early on because implementation details (chunking, embeddings, citations) are complex and no full out-of-the-box platform exists yet.

    • SIs/services will help enterprises implement vector DBs, embeddings, chunking, and citations
    • Technical details strongly influence solution quality and usefulness
    • Enterprises can build internally, but it’s a heavy lift to develop in-house expertise
    • Long-term value split between services vs applications is uncertain, but near-term demand is clear
  9. 21:12 – 24:21

    Incumbents vs startups in AI: expanding the pie and creating new categories

    Howie argues AI can grow the overall market rather than strictly create winner-take-all outcomes. Incumbents can monetize existing distribution (e.g., Copilot ARPU), while startups can reimagine workflows and open new use cases (e.g., new presentation paradigms).

    • AI can increase ARPU on massive installed bases (e.g., Office/Copilot)
    • Incumbents can deepen value (e.g., Adobe Generative Fill) and charge more
    • Startups can build novel experiences unconstrained by legacy expectations
    • Disruption may come from new categories, not direct replacement of existing use cases
  10. 24:21 – 27:30

    Can Airtable move fast as an ‘incumbent’? Speed, tradeoffs, and enterprise-grade constraints

    Asked whether incumbency slows execution, Howie says Airtable is still closer to an upstart relative to giants. He outlines the core tradeoff: building clean-slate products quickly versus integrating changes into a mature platform with security, stability, and SLA expectations.

    • ‘Incumbent vs startup’ is a spectrum; sub-$1B revenue still feels upstart vs titans
    • Resource scarcity remains real even after large funding and headcount growth
    • Clean-slate products can move faster; integrated enterprise-grade shipping is inherently slower
    • Security, stability, and customer expectations constrain ‘move fast’ shortcuts
  11. 27:30 – 30:43

    When will the enterprise AI ‘train’ arrive? An iterative adoption loop and Airtable’s AI primitives

    Howie is skeptical that ‘next year’ is a magic inflection point; instead he sees monthly progress in enterprise sophistication. He describes an iterative feedback loop—ship primitives, test use cases, learn—and shares how Airtable approaches AI Fields and workflow-driven deployments.

    • Enterprise AI adoption is progressing steadily, not tied to a single calendar year
    • Companies start with lower-stakes, higher-upside use cases and iterate
    • Airtable is building flexible primitives (e.g., AI Field) that chain into automations/interfaces
    • They pair primitives with targeted starter workflows (marketing/product) and refine via feedback
  12. 30:43 – 33:28

    PLG to enterprise: defining the senior buyer, avoiding ‘productivity-only’ sales, and surviving consolidation

    Howie explains why transitioning from PLG to enterprise is hard: enterprise buyers consolidate tools and demand org-level value. Products that only deliver individual productivity gains struggle; winners define a strategic buyer and sell differentiated, outcome-linked ROI.

    • Enterprise consolidation pressures vendors: thousands of SaaS tools get rationalized into a few
    • Individual-only value forces a weak ‘productivity sale’ that’s harder to defend
    • Successful enterprise motion requires a defined senior/strategic buyer with a business case
    • Airtable positions as an app platform delivering end-to-end process ROI (not ‘just spreadsheets’)
  13. 33:28 – 39:51

    Bundling vs differentiation: where Airtable sits—and why AI hype doesn’t close deals

    Harry asks about suite bundling dynamics and whether Airtable is vulnerable to Google/Microsoft bundles. Howie argues commoditized, shallow tools get bundled, while differentiated ROI platforms can resist; he adds that ‘sexy AI features’ don’t sell—specific business cases do.

    • Bundling wins for horizontal, commoditized tools (video conferencing, whiteboarding, generic docs)
    • Enterprises choose simplicity/cost-effectiveness when differentiation is weak
    • Differentiated platforms can sell via quantifiable business ROI (process speed, fewer errors, visibility)
    • Leading with AI excitement opens conversations, but deals close with concrete use-case value
  14. 39:51 – 46:50

    Enterprise reality check: logos vs meaningful spend, $1M ARR threshold, and today’s ROI scrutiny

    Howie cautions that small pilots with big-name logos can be misleading; meaningful enterprise commitment starts closer to ~$1M ARR. He also describes the post-2020 shift to tool rationalization, usage audits, and outcome-based proof—changing how customer success must operate.

    • Big logos with tiny seat counts/spend are often just experiments, not true adoption
    • Howie’s heuristic: ~$1M ARR is the threshold for a ‘real’ enterprise account (250–500k is intermediate)
    • Airtable reached $1M+ contracts around 2019, powered by organic internal groundswell
    • Macro shift: enterprises now demand ROI proof, evaluate active-to-paid ratios, and prioritize business outcomes over activity
  15. 46:50 – 56:43

    Quick-fire reflections: investing, startup myths, scaling discipline, valuation pressure, and leadership legacy

    In the closing quick-fire, Howie shares why he won’t concentrate angel bets, what investing taught him about founder struggle, and why ‘PMF solves everything’ is a myth. He closes on durable growth over valuation-chasing and a leadership focus on building a great company rather than personal legacy.

    • Angel investing: diversification matters; it’s too early to know AI category winners
    • Big lesson: every company’s journey is harder internally than it looks externally
    • Bullshit advice: PMF is only the beginning; scaling introduces new, harder problems
    • Valuation: focus on durable, efficient growth; multiples and macro are outside your control

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.