Overview
Michael Dempsey is a Managing Partner at Compound[1][3], where Dempsey invests in founders and companies at early stages in emerging areas including crypto, AI/ML, robotics, and healthcare/bio[4]. Dempsey's investment career began at Crane Partners, a multi-strategy hedge fund, where Dempsey worked as an Investment Analyst from November 2011 to March 2014, investing across multiple asset classes including public equities, derivatives, and cross-border private equity[4][7]. Dempsey subsequently joined CB Insights from April 2014 to January 2016, where Dempsey helped scale the research and data analysis team[4][6]. Dempsey holds a Bachelor of Arts in Economics from New York University[8].
Profile introduction
I invest in founders and companies at the early stages in emerging areas, including but not limited to Crypto, AI/ML, Robotics, Healthcare/Bio, and more. I began my career investing multiple asset classes ranging from public equities and derivatives, cross-border private equity, and eventually seed-stage startups at Crane Partners, a multi-strategy hedge fund. From there I helped scale the research and data analysis team at CB Insights. I graduated from New York University. Note: If we've never met please include a written note with your LinkedIn request or email me.
Career history
- Managing PartnerSep 2016 to PresentCompound
- Research & Data AnalysisApr 2014 to Jan 2016CB Insights
- Investment AnalystNov 2011 to Mar 2014Crane Partners
Education
Bachelor of Arts, EconomicsNew York University
Insights & ideas
The through-line
Across nearly every appearance, Michael Dempsey returns to the idea that venture investing in frontier technology should be built on developed theses rather than reactive deal-chasing, and that this discipline is what has let him run a research-centric firm, Compound, across very different technical domains: AI/ML, robotics, biology, crypto, and materials [3][5][7]. The throughline has widened over time. Early on the focus was narrow and creative, machine learning applied to animation and creative studios [8]; over the following years the lens opened to decentralized energy grids, local software, and robotics [7], and more recently to the macro question of who actually captures value in the current AI buildout and whether the infrastructure spending underway is sustainable [4][6]. What stays constant is the posture: skeptical enough to interrogate hype, but committed enough to have been investing in AI theses well before the current cycle [4].
On thesis-driven vs founder-driven investing
Dempsey draws a distinction between investing by developing a thesis about where a technology or market is headed and then finding founders to back within it, versus investing primarily on the strength of a founder regardless of sector, and this contrast is presented as central to how Compound operates and to Dempsey's own advice for other investors and emerging founders [3]. The related question of how Compound actually builds those theses, and where the firm sees company-building opportunities as a result, runs through his account of the firm's process [3][7]. He connects this directly to a broader philosophy of "ways to win at venture," suggesting that a research-first approach to picking sectors is itself a competitive strategy rather than just an analytical exercise [3].
On developing theses across frontier technology
Dempsey describes Compound's research process as spanning multiple frontier categories at once, including AI/ML, robotics, biology, and crypto, with an explicit emphasis on understanding blockchain-native challenges and commercial opportunities alongside more conventional AI and robotics bets [5]. He discusses how he thinks through specific theses in robotics, decentralized energy grids, and "local software," treating each as its own research problem rather than folding them into a single generic technology narrative [7]. The account of his process for exploring an entirely new space, going back to his early work on machine learning in creative tools, shows the same method applied at smaller scale: identifying an underexplored technical capability, in that case ML for animation and creative studios, and building out a view on how it would diffuse into products before others focused there [8].
On value capture in deep tech and AI
A recurring concern is that deep tech startups routinely mishandle value capture, meaning they build real technical capability but fail to structure their business so that they, rather than platforms or downstream players, keep the economic upside [4]. This sits alongside a broader discussion of the present AI market, where Dempsey takes part in weighing the "glass half full or empty" question about generative AI, engaging with skepticism from voices like Goldman Sachs' Jim Covello on one side and the rapid revenue growth of companies like OpenAI and Anthropic on the other [6]. He also discusses AI applied to advanced materials and climate technology as a distinct, less-discussed value-capture opportunity relative to consumer-facing generative AI [6], and takes on the question of potential overcapacity in AI infrastructure and how major tech firms are shifting strategy in response [6].
On AI in creative work
Dempsey's earliest recorded thesis work concerns machine learning's application to animation and creative studios, where he discusses automation in animation pipelines, the prospect of AI replacing human creative labor, and the democratization of machine learning tools for creatives, citing tools like RunwayML as part of that shift [8]. This was framed as part of a wider view of what "creative studios of the future" would look like once ML tooling became accessible outside specialist labs [8].
On the AI investment landscape
Beyond value capture, Dempsey engages directly with the state of AI investing: the scale of funding flowing into labs such as Anthropic from Google and Amazon, how those valuations are evolving, and the tension between continued infrastructure buildout and questions about whether the spending is justified by benefit [6]. His view sits deliberately between boosterism and the skeptic case, engaging seriously with both the growth data from leading AI labs and the doubts raised by Wall Street critics [6].
On running Compound and venture strategy
Dempsey's own track record, described as having grown every fund he has run, is tied explicitly to the research-driven approach rather than opportunistic dealmaking [2]. His discussion of "ways to win at venture" and advice for emerging investors extends this into a broader philosophy about how smaller or newer firms should differentiate themselves against larger competitors, namely through depth of sector research rather than access alone [3].
Takeaways
- Compound's process is built around developing sector theses first (AI/ML, robotics, biology, crypto) and finding founders within them, which Dempsey distinguishes from purely founder-driven investing [3][5].
- He identifies value capture, not technical capability, as the point where most deep tech startups fail [4].
- He treats decentralized energy grids and "local software" as live investment theses alongside more obvious AI and robotics bets [7].
- On generative AI, he engages both the skeptic case (echoing critiques like Jim Covello's) and the bull case (rapid OpenAI/Anthropic revenue growth) without settling fully on either [6].
- He flags AI applied to advanced materials and climate tech as an underappreciated value-capture opportunity relative to consumer AI [6].
- His early thesis work centered on machine learning in animation and creative studios, including tools like RunwayML, and the question of AI replacing human creative labor [8].
- He raises the possibility of overcapacity in AI infrastructure spending and shifting strategy among major tech firms as a live risk in the current cycle [6].
Media & appearances
- PersonableApple PodcastsManaging Partner Compound: How He Has 5X Every Fund He Has Run | Michael DempseyMichael Dempsey, Managing Partner at Compound, former analyst at Crane Partners and CB Insights, NYU graduate, and one of the most respected research driven investors in frontier technology, joins Personable to reveal how he has 5X every fund he has eve
- The Peel with Turner NovakApple PodcastsThe Past, Present, and Future of AI, Robotics, Venture Capital, Crypto | Why Deep Tech Startups Mess up Value Capture | Michael Dempsey, Managing Partner, CompoundMichael Dempsey is the Managing Partner at Compound, a thesis-driven, research-centric investment firm. We spent two hours talking through the past, present, and future of a bunch of topics in technology and investing. Michael started investing in AI
- RiskReversal PodApple PodcastsIs the GenAI Glass Half Full Or Empty with Deirdre Bosa, Michael Dempsey & Jonathan Godwin | Okay, Computer.Dan Nathan is joined by Deirdre Bosa, host of CNBC's TechCheck, to discuss the latest in AI investments, including Google and Amazon's funding in Anthropic, and how these valuations are evolving. They also banter about a New York Times article scrutinizing the future impact of generative AI. The episode highlights the rapid revenue growth in companies like OpenAI and Anthropic, growing interest and skepticism towards AI from big players like Goldman Sachs, with insights from Jim Covello. The conversation touches on consumer applications of AI, such as the use of models from companies like OpenAI and Anthropic by platforms like Perplexity. Bosa explores the strategic moves by Microsoft to train its own AI models and the broader implications for the AI market. After the break, Dan Nathan is joined by Michael Dempsey, managing partner at Compound VC, and Jonathan Godwin of Orbital Materials, discussing their innovative approach using AI for advanced materials and climate tech. The episode wraps with a look at potential overcapacity in AI infrastructure and shifting strategies in major tech firms. Further Reading Will A.I. Be a Bust? A Wall Street Skeptic Rings the Alarm. (NYT) AI’s $600B Question (Sequoia Capital) Gen AI: too much spend, too little benefit?
- INDIE AUDIOApple PodcastsIntersection of INDIE & Deep Tech — A Conversation with Michael Dempsey, Compound Managing PartnerMy friend, Michael Dempsey, Managing Partner at Compound, and I have an ongoing thread about the intersection of indie ideals and his areas of focus in deep tech. Most of these conversations happen over meals or online, so we decided to dive in a bit wi
- Scenius StudioApple PodcastsCompound - Michael Dempsey (Episode #41)In this episode, Ben sits down with Michael Dempsey, Managing Partner of Compound. At Compound, Michael oversees investments in AI/ML, robotics, biology, and of course crypto. His cross-industry knowledge and deep understanding of blockchain-native challenges and commercial opportunities makes him stand out from the crowd. Let's get into it. Connect with the guest Compound https://www.compound.vc/ Follow Compound on Twitter/X https://twitter.com/compoundvc Follow Michael on Twitter/X Disclaimer Ben Jacobs is a partner at Scenius Capital Management. All views expressed by Ben and the guests of this podcast are solely their opinions and do not reflect the opinions of Scenius Capital Management. Guests and the host may maintain positions in the assets or funds discussed in this podcast. You should not treat any opinion expressed by anyone on this podcast as a specific inducement to make a particular investment or follow a particular strategy but only as an expression of their personal opinion. This podcast is for informational purposes only.
- "Turpentine VC" | Venture Capital and InvestingApple PodcastsE27: Compound’s Michael Dempsey on Thesis-Driven vs Founder-Driven InvestingIn this episode of Turpentine VC, Michael Dempsey, Managing Partner at Compound, joins Erik Torenberg to discuss thesis-driven vs founder-driven investing, Compound’s AI thesis, company building opportunities, ways to win at venture, advice for emergi
- Moth MindsApple PodcastsMichael Dempsey: Managing Partner at CompoundMichael Dempsey is the managing partner of Compound, a thesis-driven, research-centric investment firm. Prior to founding Compound, Michael worked at the predictive analytics startup CB Insights and at Crane Partners, a multi-strategy hedge fund. Our conversation covered how Michael views venture capital and frontier technologies, how Compound develops theses, and how he thinks about several of those theses in the areas of robotics, decentralized energy grids, local software, and AI. — Moth Minds is a podcast where Molly Mielke of Moth Fund interviews high-agency humans.
- Live from the Future (Frontier Tech | Emerging Tech | Deep Tech)Apple PodcastsFuture of AI&ML in Creative Tech w/ Michael Dempsey @ Compound (0004)Future of AI&ML in Creative Tech: Machine Learning in Animation, Artificial Intelligence for Creative Studios, Democratization of Machine Learning for Creatives Michael Dempsey is a partner at Compound and invests in frontier technologies such as robotics, autonomous vehicles, machine learning, immersive computing, & AI/ML for creative work which is our topic today. Prior to joining Compound, Michael was an early employee at CB Insights where he focused on researching frontier tech and then he made the jump into venture when he joined Rothenberg Ventures in SF. /// IN THIS EPISODE WE COVER /// 1. What does the future of AI&ML in Creative Tech look like 2. Automation with ML in animation and creative studios 4. AI replacing human creatives 5. Shadows and studio of the future 6. RunwayML and democratization of ML 7. VC research process for new spaces /// NOTABLE MENTIONS /// Animation Is Eating The World - https://www.michaeldempsey.me/animation/ Shadow - https://shadows.co/ RunwayML - https://runwayml.com/ Morphin - https://getmorphin.com/ Facemoji - http://facemoji.me/ The Mission of a Lifetime - https://www.goodreads.com/book/show/41454623-the-mission-of-a-lifetime @Gwern - https://twitter.com/gwern
In the news
- There are pivotal times in the company building journey where you must light a fire. Sometimes that means burning a bridge so there's no going back. Other times it means burning down what you've built and starting over. We have found @compoundvc that sometimes founders need https://t.co/8jGCj0NtLV
- This tweet is a Rorschach test for if you should be incredibly bullish or incredibly bearish.
- I often feel we’ve lost the concept of Chapters in our lives
- I think a few things happened in tech from 2022-2026 that created a perfect storm alongside the breakout of scaling laws and AI performance that led to a miserable approach/result in how AI made its way into the public consciousness. In no particular order we: - Embraced a more https://t.co/Msm5KKUwm4
- I think it's really important as a private markets investor to understand what remotely variant view you have on the various branches of technology right now. And if you don't, then you probably need to be investing out of a multi-billion dollar fund that can be levered beta or
- kith + senra podcast is some form of superbowl for the most insufferable tech bro you know
- Early evening walk. https://t.co/39D2QVU7Rk
- Every friend group wants one of their friends to apply for the head of presidio job and nobody is stepping up. So today I’d like to announce
This page shows public professional information only, each fact cited. Is this you? send a correction, or ask for removal within 24 hours, no questions asked.