Uncapped with Jack AltmanFormer CAA Talent Agent Turned Investor with $70B in AUM on AI and Venture Strategy | Ep. 29
CHAPTERS
- 0:00 – 1:51
Cycle check: why the AI moment feels different this time
Jack asks Thomas for a temperature check on valuations and bubble talk in fall 2025. Thomas frames the current cycle through prior “seminal moments” (iPhone/Apple, NVIDIA’s data center inflection) and argues AI infrastructure spending is entering a higher-stakes phase.
- •Valuations returning to 2021-like multiples for certain AI-flavored companies
- •Historical analogs: iPhone-era Apple earnings beats; NVIDIA data center re-rating alongside ChatGPT moment
- •Oracle’s recent move as a signal of a new phase in AI infrastructure
- •Shift from optional spending to “existential” competition and urgency
- •Need for increased vigilance as capital intensity rises
- 1:51 – 5:31
AI infrastructure arms race: leverage enters the build-out
Thomas explains what changed with the Oracle announcement: AI build-out is no longer funded only by hyperscalers’ cash flows. Free-cash-flow-negative players are now making massive commitments, raising the stakes and competitive intensity across the ecosystem.
- •Earlier AI capex funded by highly profitable Big Tech cash flows
- •New dynamic: leveraged/FCF-negative investors (e.g., OpenAI) making huge infrastructure bets
- •Cloud competition intensifying beyond the prior three-player oligopoly
- •Oracle’s potential path to meaningful cloud share; GPU-specialist clouds like CoreWeave
- •Investment posture changes when companies spend beyond internally generated cash
- 5:31 – 6:45
Deploying capital from inception through IPO: thematic, wide-aperture portfolio construction
Jack probes how Coatue thinks about allocating across public vs private and seed vs growth. Thomas describes a thematic approach guided by a “wide aperture lens,” then mapping themes to the best vehicles across stages and markets.
- •Capital deployment driven by themes rather than stage labels
- •Research synergies between public and private company work
- •AI as a multi-layer opportunity spanning infrastructure, models, and applications
- •Conviction varies by layer; fuzziness increases higher in the stack
- •Goal: find the best beneficiaries of major tech transitions
- 6:45 – 8:28
Layer 1—Semis, data centers, and power: the physical constraints of AI
Thomas outlines AI infrastructure as a foundational layer: chips, data centers, and power. He cites public winners and private innovators, emphasizing that AI’s growth is gated by real-world supply chains and energy availability.
- •AI requires semiconductors; Thomas’s roots as a semiconductor analyst
- •Public examples: NVIDIA; Broadcom/Avago under Hock Tan
- •Private innovation example: Cerebras
- •Infrastructure needs extend beyond compute to data centers and grid capacity
- •High-conviction investing opportunities exist across public and private markets
- 8:28 – 9:41
Power as an AI bottleneck: nuclear, gas turbines, and capacity limits
In response to Jack’s question on power, Thomas highlights near-term investable technologies and constraints. He points to nuclear and gas generation build-outs, plus the practical reality that critical capacity is already sold out for years.
- •Nuclear as a key pillar; “behind-the-meter” deals tying plants to hyperscalers
- •Constellation Energy examples and reactor reactivation dynamics
- •GE Vernova and natural gas generation as near-term scalable supply
- •US gas abundance vs. turbine capacity shortages (multi-year sold-out timelines)
- •Why power capture is central to AI’s continued scaling
- 9:41 – 10:45
Layer 2—Foundation models: market coalescing around a few labs
Thomas moves up the stack to models, where he believes a small set of companies will dominate. He describes how foundational models enable transformational applications and set the baseline for the rest of the ecosystem.
- •Expectation of consolidation around a handful of model providers
- •Named leaders: OpenAI, Anthropic, Google; watching Meta/Microsoft/Amazon
- •Models as a durable layer powering many downstream products
- •ChatGPT as a step-change in software expectations and workflows
- •High conviction relative to more uncertain application-layer outcomes
- 10:45 – 13:00
Layer 3—Data layer disruption: the “system of record” loses its lock-in
Thomas argues the era of SaaS platforms locking data inside proprietary systems is ending. Using Workday’s move to integrate with Snowflake/Databricks, he predicts value shifts toward open data layers and agent-centric enterprise software.
- •Workday’s announcement as a key signal: integrating with Snowflake/Databricks
- •Data portability enables combining HR/finance data with broader enterprise datasets
- •Future differentiation shifts from storage to “Fortune 500-grade” agents
- •Implication: traditional SaaS lock-in weakens; databases/data platforms gain centrality
- •Open question: is software “dead” or being reinvented around agents?
- 13:00 – 18:02
Enterprise memory: default-on recording and agent-driven compliance/coaching
Thomas predicts that within a few years most enterprise interactions will be recorded by default, feeding agent systems. He offers a compliance framing: proactive, agent-led remediation can reduce catastrophic blowups from bad behavior or regulatory risk.
- •Prediction: default-on recording for meetings, email, Slack, and even in-person interactions
- •Rationale: enterprise intelligence comes from interactions, not just form fields
- •Compliance officer concerns vs. the proactive remediation model
- •Multi-step escalation concept: agent coaching → mandatory training → HR escalation
- •Cultural shift: people increasingly assume everything is recorded
- 18:02 – 18:46
Gong and the ROI of recorded corpuses: faster ramp times and best-practice extraction
Thomas gives a concrete example of how recorded interaction data becomes more valuable with generative AI. Gong’s re-acceleration illustrates how organizations can codify top-performer behaviors and improve productivity through automated analysis.
- •Gong as a sales intelligence example that benefits from generative AI
- •Extracting what great reps do differently (messaging, competitor framing)
- •Outcome: reduced ramp time and improved onboarding/coaching
- •A preview of how similar approaches spread across enterprise functions
- •Recorded data becomes a strategic asset when paired with strong models
- 18:46 – 24:17
What wins in venture: being the roaming “fisherman,” not fighting for one spot on the river
Jack asks about winning venture strategies; Thomas explains Coatue’s differentiated approach. Rather than anchoring to one stage, Coatue “travels” across stages and markets, using theme-driven research to find opportunities wherever they appear.
- •Coatue’s organic compounding and long-term orientation
- •River analogy: established firms have the best “spots”; Coatue chooses mobility
- •Cross-stage investing as a feature, not a compromise
- •Public/private research flywheel: insights transfer both directions
- •Growing acceptance that venture can scale and still work with different return shapes
- 24:17 – 28:56
Early-stage operating dynamics: tribal VC, patience, and ‘you didn’t miss the company—just a round’
Thomas explains how he operates in early-stage ecosystems that can be tribal and zero-sum. He shares a key pattern from his own P&L: many of his biggest winners were initially “no’s,” and later entry still worked—often after mutual understanding improved.
- •VC tribalism: he aims to be welcomed across “villages” and add value
- •Analysis of top winners: first chance to invest was often a “no” (either side)
- •Lesson: relationships and understanding often take time to develop
- •Contrast with single-shot Series A dynamics that amplify zero-sum behavior
- •Adapting strategy as sectors/geographies shift (e.g., China no longer available)
- 28:56 – 34:26
Navigating conflicts at scale: the ‘second call’ role, disclosure, and hard compliance walls
Thomas offers a nuanced view on conflicts: some are real (board/large ownership in direct competitors), but many perceived conflicts can be managed. He argues Coatue’s value can increase with broader visibility—if paired with strict disclosure and information controls.
- •Positioning: traditional VCs aim to be the founder’s first call; Coatue often aims to be the second call
- •Clear red line: won’t back a direct competitor when deeply involved (e.g., 20% owner/board)
- •Benefits of broader investing: domain expertise and network visibility
- •Principles: disclose conflicts early; protect trust via strict info security
- •SEC-regulated context increases the need for rigorous internal controls
- 34:26 – 36:44
Wide-aperture investing as curiosity: when breadth helps—and when specialization wins
Thomas defines “wide aperture” as unbounded curiosity across technologies, geographies, and stages. He notes the trade-off: breadth can fail in domains like early crypto where deep specialization provided an edge.
- •Curiosity as a core trait of great investors (examples: Druckenmiller, Andreessen)
- •Freedom to learn first, then determine relevance later
- •No rigid filter by asset class or stage
- •Downside case: crypto rewarded early deep specialists
- •Personal strategy: bet on strengths while acknowledging trade-offs
- 36:44 – 40:29
Evaluating founders: CAA lessons and why ‘star quality’ is real
Drawing on his time at CAA, Thomas describes a specific, hard-to-fake presence he looks for in founders. He connects charisma and force of personality to market insight, using examples from entertainment and tech leaders to show how it manifests.
- •“Star quality” as a tangible, rare trait
- •Examples from acting: Tom Cruise presence; Colin Farrell magnetism pre-fame
- •Founder analog: Evan Spiegel’s worldview shaping Snap’s product philosophy
- •Magnet vs. polarizing leaders—both can win if the pull is strong
- •Example of polar founder energy: Travis Kalanick and Uber dominance
- 40:29 – 48:23
How Coatue decides: momentum-based collaboration vs. single-meeting investment committees
Jack asks about decision-making mechanics and firm design. Thomas explains that Coatue’s decisions build through iterative internal momentum—early and broad feedback—rather than a single “oracle” meeting with definitive debate.
- •Skepticism of traditional investment-committee “smoke up the chimney” moments
- •Deals progress via repeated check-ins and cross-team collaboration
- •Early solicitation of expert input as a cultural expectation
- •Ideas often die by loss of momentum rather than a dramatic committee ‘no’
- •Meritocratic structure: faster impact for young talent; low tolerance for ‘dead weight’
- 48:23 – 54:25
Why golf matters: integrity, presence, mentorship, and a four-hour relationship container
Thomas explains how golf changed his life, primarily through relationships and mentorship formed on the course. He also highlights golf’s built-in integrity tests, the ability to compete across ages/skill levels, and the rare benefit of being offline for hours.
- •Golf as a relationship engine: mentors, friends, and deep conversations
- •Integrity and self-scoring as a character test
- •Presence: four hours without phones and with a defined end point
- •Handicap system enables competition across skill, age, and gender
- •Value of great teachers; learning accelerates with expert coaching
- 54:25 – 57:19
Working with family: trust, no politics, and the challenge of never fully turning it off
Thomas reflects on building Coatue alongside his brother Philippe. The upside is unusually high trust and minimal politics; the downside is that the business can dominate family life, requiring deliberate effort to create off-switch activities.
- •Constant communication; shared founder burden
- •Key advantage: no questioning motives → dramatically less politics
- •Economics rarely discussed because baseline trust is high
- •Downside: hard to separate work and family, especially during crises
- •Coping mechanism: shared activities (e.g., golf) to reduce work intensity
- 57:19 – 1:02:43
Mentorship and early-career advice: focus as a luxury and the ‘gift-wrapping’ lesson
Thomas closes with guidance for young professionals: treat early focus as a rare advantage and use it to develop craft-level excellence. His CAA gift-wrapping story illustrates how obsessive attention to detail can differentiate you and earn bigger opportunities.
- •Mentorship as a core life principle (both mentor and mentee roles)
- •CAA experience and learning integrity through close apprenticeship
- •“Focus is a luxury” early in a career—use it deliberately
- •Gift-wrapping anecdote: meticulous execution created a reputation and new role
- •Analyst takeaway: clarity, concision, and detail mastery beat volume and noise
