Forget Growth. Your Valuation Depends On Your AI Story | Tomasz Tunguz, GP @ Theory Ventures
Tomasz Tunguz, a venture capitalist at Theory Ventures, discusses how AI is fundamentally reshaping SaaS valuations, sales metrics, and go-to-market strategies. He argues that the market now values AI narrative and enterprise budget increases over traditional growth metrics, while warning that observable productivity gains may actually reflect demand-side changes rather than supply-side innovations.
Summary
Tomasz Tunguz joins the Topline podcast to explore how the AI boom is warping traditional SaaS investment dynamics and sales metrics. The discussion begins with public market valuations, where companies like CrowdStrike trade at 34x forward revenue while not being the fastest growers in their categories. Tunguz explains that the market has shifted from paying purely for growth to paying for the "AI story"—companies either reselling AI tokens at margins or providing infrastructure to those selling tokens. He uses examples like Twilio and DigitalOcean, which have exploded in valuation due to AI positioning.
A significant portion of the conversation addresses how enterprise security budgets have expanded enormously following high-profile AI vulnerabilities like the Hugging Face attack. Tunguz and the hosts debate the nature of AI agent behavior using reinforcement learning concepts, discussing whether recent AI incidents reflect genuine intelligence or merely monomaniacal pursuit of programmed goals. This leads to broader questions about AI sandboxing and safety.
The conversation shifts to Field Development Engineers (FDEs)—a critical but often overlooked investment trend. AI companies have committed nearly $10 billion annually to placing their own engineers inside customer organizations. Tunguz frames this as necessary education spending, helping enterprises transition to AI-native workflows. He identifies four FDE models: self-funded (like Palantir), spun-out entities (Google), partnerships (Microsoft), and internally managed. The tension emerges between software-only efficiency and FTE-driven customer retention and upsell—suggesting that FDEs function like sales engineers did in previous eras, necessary for complex implementations but unnecessary for simpler products.
When examining sales team productivity, the hosts challenge whether the astronomical quota increases (from $2-4M historically to $150M+ in some cases) reflect genuine rep productivity gains or simply massive demand-side budget increases. Tunguz and Asit Zaman both acknowledge that it's primarily demand-side: enterprise AI budgets have exploded, not rep efficiency. This distinction matters enormously for investors evaluating whether competitive advantages are sustainable.
On the market opportunity side, Tunguz discusses three initial AI use cases with strong product-market fit: coding, chat, and support. He's bullish that these haven't saturated and that security represents a fourth major use case, with markets growing alongside software production. The monetization model for security shifts from seat-based to data-volume-based pricing across endpoint, network, database, and code layers.
The hosts explore whether Tunguz uses AI differently in his own work. He describes an iterative process where AI helps him improve blog writing quality rather than reduce effort—consistent with his chess grandmaster analogy: top performers don't train less with AI; they train as much but hold themselves to higher standards. His analysis shows 134 edits per post regardless of time spent, but a 20% quality increase when graded by external criteria.
Regarding founder selection and investment strategy, Tunguz notes that while technical innovation remains important, go-to-market innovation has become significantly more valuable. He observes that successful companies in both the SaaS era and now innovate on distribution strategy—citing examples like Dropbox, Zoom, and open-source companies like HashiCorp and Confluent. However, he notes that go-to-market secrets are closely guarded because outbound has been so challenging.
Interestingly, CRO talent at top AI companies (Dali at OpenAI, ServiceNow veterans at Anthropic and others) comes from the SaaS era's best practitioners, suggesting organizational design may be more continuous than revolutionary. However, rep preparation and CRM automation are changing how daily work functions.
The conversation culminates in a "Bulls vs. Bears" segment on 50-100x multiples for AI infrastructure companies in 12 months. Tunguz leans bearish, citing concerns about rising treasury rates and inflation, which could constrain funding for the massive data center buildouts AI requires. However, he acknowledges the countervailing force: AI models keep improving at geometric rates, and the economy has become so dependent on AI investment that policy-makers cannot allow it to fail. The hosts highlight a structural tension: data center capacity must grow to meet modeled future demand, but this requires continuous funding through private credit, potentially foreign treasury purchases, or government backing—all creating inflation risks that could eventually force rate increases that constrain valuations.
About this episode
<p>Forget Growth. Your Valuation Depends On Your AI Story | Tomasz Tunguz, GP @ Theory Ventures</p> <p>Tomasz Tunguz, General Partner at Theory Ventures, joins Sam Jacobs, AJ Bruno, and Asad Zaman with public software multiples sitting at 4 to 4.5 times forward revenue, down from the 100x ceiling of 2021. Topics include why the market now pays for an AI story rather than growth, the $10 billion AI companies committed in a single year to putting forward-deployed engineers inside customer orgs, and the 41 days a model company gets to commercialize a state-of-the-art release before it is knocked off the perch. Plus, why the mid-market is being abandoned for enterprise deals that close in 45 days, what quota inflation at AI-native companies is actually measuring, and a bull-versus-bear call on 50 to 100x multiples for AI harness companies twelve months from now.</p> <p>Key Takeaways:</p> <p>- The valuation follows the AI story, not the growth rate. Average public software trades around 4 to 4.5 times forward revenue today, and the category leaders carrying 30x and up are usually not the fastest growers in their category. As Tomasz put it: "what the market is asking for, and this is both true in the public and the private market, is great, you have an existing business, now show me that you can sell tokens." Revenue growth is still the highest single correlate to multiple, but the token story is what re-rates a company before the revenue shows up.</p> <p>- Before you copy an AI-native company's quota model, work out where the number is actually coming from. "It's not the supply side has changed and suddenly become 10x more productive. It's the demand side budgets have increased by a factor of 10. And that's what's driving quotas," said Tomasz. Quota-to-OTE ratios that topped out between 2.5 and 4 at Oracle and IBM five years ago are now routinely 1.5 at early startups, and a single insurance account can carry a quota in the tens to hundreds of millions, which is why he no longer grades companies on AE to SDR or AE to CSM ratios at all.</p> <p>- Using AI to save time is the wrong target for top-tier performance. Tunguz rebuilt his blog workflow so every edit runs through AI, and the total edit count held flat at 134 per post regardless of how long he had been at it, while a graded review of ten years of posts showed a 20% quality increase in 2026. "if you're a chess grandmaster and you want to be better with AI, you don't train less. You train just as much, but you hold yourself to a higher standard." The hour to an hour and a half per post did not change; the research, citations, and depth of analysis did.</p> <p>- Product advantage is thinning, so the distribution move is what investors are underwriting now. "the go-to-market innovation is now significantly more important than it was. And if you can find a founder who can execute a beautiful go-to-market judo move and produce a lot of leverage for the company, then it's incredible," said Tomasz Tunguz, General Partner at Theory Ventures, placing Dropbox, Zoom, Confluent, and HashiCorp as the previous era's version of the same pattern. </p> <p>Connect with the Hosts & Guests:</p> <p>Host: Sam Jacobs, CEO at Pavilion - https://www.linkedin.com/in/samfjacobs/<br /> Host: AJ Bruno, CEO at QuotaPath - https://www.linkedin.com/in/ajbruno3/<br /> Host: Asad Zaman, CEO at STA - https://www.linkedin.com/in/azaman1/<br /> Guest: Tomasz Tunguz, General Partner at Theory Ventures - https://www.linkedin.com/in/tomasztunguz/</p> <p>Topline is more than a YouTube Channel:</p> <p>Subscribe to Topline Newsletter: https://toplinemedia.substack.com/<br /> Tune into Topline Podcast, the #1 podcast for founders, operators, and investors in B2B tech: https://www.joinpavilion.com/topline-podcast<br /> Join the free Topline Slack channel to connect with 600+ revenue leaders to keep the conversation going beyond the podcast: https://www.joinpavilion.com/topline-slack</p> <p>Chapters:<br /> 00:00 Introducing Tomasz Tunguz<br /> 02:16 Market Stopped Paying For Growth<br /> 04:56 Show Me You Can Sell Tokens<br /> 06:33 Inside The Hugging Face Hack<br /> 10:07 Explaining RL With A Roomba<br /> 13:43 Is Security The Next AI Use Case<br /> 17:49 Fertility, Consumption, And Tokens<br /> 22:36 Overleveraged On The AI Buildout<br /> 26:08 Writing Every Post Through AI<br /> 30:55 The $10 Billion FDE Boom<br /> 39:11 FDE Moat Or Toll Booth<br /> 44:39 Quotas Are A Demand-Side Story<br /> 52:54 Does Customer Success Die<br /> 55:55 Enterprise Shift And GTM Judo<br /> 01:03:41 Bulls and Bears</p> <p>Sources & Attribution:</p> <p>2015 quota figure: https://tomtunguz.com/how-big-should-your-sales-pipeline-be/<br /> SaaS AE compensation: https://blog.bridgegroupinc.com/saas-account-executive-compensation<br /> 2024 AE metrics and compensation benchmark: https://blog.bridgegroupinc.com/2024-ae-metrics-compensation-benchmark<br /> AI investment map: https://www.forbes.com/councils/forbestechcouncil/2026/02/27/ai-is-redrawing-the-tech-investment-map-in-2026/<br /> The $10B FDE Boom: https://tomtunguz.com/the-10b-fde-boom/<br /> Clip: https://www.youtube.com/watch?v=8DbCQn86f9w<br /> Clip: https://www.youtube.com/watch?v=bSo4V6tY9XQ&t=1385s<br /> Clip: https://www.youtube.com/watch?v=36mE3cjYldU&t=560s</p>
Key Insights
- Tunguz argues that the market is valuing the AI narrative and potential token consumption rather than current AI-driven revenue, accepting the story before seeing monetization in financial results.
- He claims that CrowdStrike's 34x multiple is driven not by their growth rate but by the market recognizing that enterprise security budgets have no limit due to AI-related threats, using Palo Alto's market cap growth from $16B to $300B as evidence.
- Tunguz observes that when AI companies reach $100M in ARR, the go-to-market composition matters less than macro-level unit economics, making granular operational metrics less predictive than in previous eras.
- He identifies that the astounding quota increases (from $2-4M historical ranges to $150M+ in some cases) reflect demand-side budget increases rather than supply-side rep productivity improvements—a critical distinction for evaluating business durability.
- Tunguz argues that Field Development Engineers represent a necessary, long-term phenomenon similar to management consultants, not a temporary patch, because models continuously improve and require ongoing re-education of customer organizations.
- He contends that software market growth is not constrained by population decline because parallelization and agent consumption create geometric growth in token consumption per capita, offsetting demographic headwinds.
- Tunguz presents data showing that despite using AI to improve blog writing quality by 20%, his edit count remains constant at 134 per post, suggesting AI elevates standards rather than reduces effort for skilled practitioners.
- He claims that go-to-market innovation is now significantly more important than technical innovation in driving venture returns, using HashiCorp's six-product IPO and Confluent's open-source-to-enterprise motion as examples.
- Tunguz observes that top AI companies recruit CROs from the SaaS era's established players (ServiceNow, PTC, Oracle veterans), suggesting organizational design principles remain continuous despite technological disruption.
- He argues that the 50-100x multiples on AI infrastructure companies are unsustainable if treasury rates continue rising, as rising rates constrain the debt and private credit funding required for multi-year data center buildout plans.
- Tunguz identifies that the macro economy faces a structural tension: AI requires perpetually increasing data center capacity funded through debt to meet modeled future improvements, creating a financing problem that cannot be solved by operational cash flows alone.
- He contends that the market's dependence on AI infrastructure has created a 'too big to fail' dynamic where government policy and central bank intervention may become necessary to prevent AI investment contraction, risking inflation as the resolution mechanism.
Topics
Transcript
Today's guest is Tomas Tunguz, a venture capitalist who has worked with at least eight unicorns. In a past life, he ran a billion-dollar business unit at Google as a product manager. In today's episode, he shares how the AI boom is warping every traditional sales metric, as well as broader investment dynamics. But he also warns CROs not to be fooled by the astronomical numbers certain reps are posting right now. If not the supply side has changed and suddenly become 10x more productive, it's the demand side budgets have increased by a factor of 10. And that's what's driving quotas. So in today's episode, we cover why product moats are disappearing and why that means VCs are hungry…
Full transcript available for MurmurCast members
Sign Up to AccessMore from Topline
SPOTLIGHT: Surviving the Inbox Apocalypse | Evan Huck, CEO & Co-Founder @ UserEvidence
Evan Huck, CEO of UserEvidence, discusses building a customer evidence platform while balancing startup growth from $4M to $11M revenue. He shares his career journey from SDR to founder and explains how a long-term brand investment strategy, after eight quarters of flat growth, finally drove significant pipeline acceleration.
How To Make Your ARR 20% More Valuable To Investors | Keenan, CEO @ A Sales Growth Company
Keenan, CEO of A Sales Growth Company, argues that not all revenue is created equal—revenue generated through heroic, reactive sales practices can be valued 15-20% lower by investors than revenue from systematic, predictable processes. He emphasizes that understanding the fundamental physics of selling and the actual business problems customers face matters far more than standardizing behaviors, and that most sales organizations suffer from broken systems rather than training deficiencies.
SPOTLIGHT: AI Isn't Just Faster Translation, It's a $40B Tug-of-War for Global Attention. | Bryan Murphy, CEO @ Smartling
Bryan Murphy, CEO of Smartling, discusses how he led his AI translation company through a rapid organizational transformation after recognizing GPT's potential to dramatically improve translation quality. He shares his framework for managing innovation through structured R&D processes with confidence levels and time-boxing, while maintaining focus on customer outcomes.
The Unwritten Rules of CRO Survival | Rick Smolen, CRO @ ShipHero
Rick Smolen, CRO of ShipHero, discusses how AI acts as a leverage tool that can either dramatically improve performance or cause catastrophic failure in go-to-market teams. He emphasizes that AI should enhance human work rather than replace critical thinking, and argues that the best revenue leaders will become exponentially better while average performers will deteriorate without proper AI implementation discipline.
SPOTLIGHT: He Missed Quota 5 Times in 25 Years. Here's the System Behind That.| Mike Carpenter, CEO & Co-Founder @ Xfactor
Mike Carpenter, CEO of X-Factor and former CrowdStrike president of global sales, discusses his methodology for consistently exceeding sales targets by finding "X-Factors" or "cheat codes"—unconventional, data-driven strategies that differentiate performance. He explains how this philosophy, combined with rigorous AI-powered analysis, led to missing quota only 5 times across 25 years of leadership, and how X-Factor.io applies this approach to help companies unlock hidden revenue opportunities.