Nvidia

Inference Will Be $10TRN Market: What No One Sees | Software Investors are F***ed | Naveen Rao

Oct 24, 2026 · 1:13:26

Naveen Rao, founder and CEO of Unconventional AI, argues that energy will be AI's biggest bottleneck and that the inference market will reach $10 trillion within 10 years. He explains that physical efficiency gains have stalled, so value comes from co-designing model, architecture, and hardware—why chips will not commoditize and Nvidia keeps the margin. He tells Harry Stebbings that pure software investors are 'kind of fucked' as SaaS fragments, with AI-plus-human outcome delivery replacing it, and that the bottleneck has shifted from creation to verification. Rao predicts robotics will be 100x bigger than agentic coding, open models will carry 80% of token traffic, and traditional VC metrics like market validation no longer fit deep tech.

Valar Atomics Sues Day One Ventures | Meta's Back to the AI Frontier | NVIDIA's $30B Reflection Deal

Oct 15, 2026 · 1:17:10

Harry Stebbings hosts Jason Lemkin and Rory O'Driscoll to break down Valar Atomics suing its seed investor Day One Ventures over pro rata rights ahead of a $6B Sequoia-led Series B, Meta's return to the AI frontier with Muse Wonder, and NVIDIA's rumored $30B acquisition of open-model lab Reflection. Lemkin says both sides erred—Day One should raise it once and let it go rather than hire Quinn Emmanuel—while Rory notes mega funds demanding pro rata waivers never waive their own. They call Meta's internal-only model performative yet see a monster AI product in Muse. The committee backs Jev's $870M Series A at $7.5B, sees bubble risk in 42 decacorns in one year, calls Haiku's 90% price cut the first deflationary LLM move, and warns the SaaS apocalypse persists without scale.

Cognition vs Factory | Anthropic Under Threat | ElevenLabs Doubles Its Valuation to $22B

Oct 8, 2026 · 1:15:52

Harry Stebbings hosts MongoDB CEO Dev Ittycheria with Jason Lemkin and Scale's Rory O'Driscoll, arguing OpenAI's near-$70BN run rate has closed the gap on Anthropic ahead of its November IPO. The panel debates whether Chris Degnan moving from Factory advisor to Cognition CRO crossed a line, and calls Vinod Khosla's 'struggling second-tier competitor' tweet a gift to rival firms. They argue Reflection's Beam gives enterprises a US open-weight model 3-4x cheaper than Chinese options, and endorse ElevenLabs at its doubled $22BN valuation. They weigh Salesforce's $2BN Listen Labs exit, Vercel's $600M ARR with agents driving 50% of new business, the Grox lawsuit over Nvidia's $17BN license deal, Meta's Muse outshining OpenAI's Dots, and Oura pulling its IPO as bad news for venture.

Menlo’s Venky Ganesan on Going All-In on AI, Anthropic & Why Venture Has Changed Forever

Oct 5, 2026 · 1:11:15

Venky Ganesan, Partner at Menlo Ventures, says venture has fundamentally changed and Menlo is 'going for broke' to back the era's defining AI companies. He recounts a 90% loss on AvenX that taught him to take chips off the table, and calls seed an option bet where Menlo is 'somewhat indifferent' to price to win a seat. He warns measured metrics get gamed, cycles crack through debt defaults, and calls passing a $50,000 Facebook check via Sean Parker his costliest miss. Ownership still matters — Menlo owns under 2% of Anthropic yet put 20% of a fund in it — and he assumes 60% dilution, saying IRR must beat the Mag Seven by 1,000 basis points. He backs partial sells to stay long, prefers capital allocation to product vision at scale, and tells LPs to judge the windshield, not the rearview mirror.

Crusoe CEO: Why Everyone Gets GPU Depreciation & AI Energy Costs Wrong

Oct 3, 2026 · 1:05:02

Crusoe CEO Chase Lochmiller tells Harry Stebbings that data centers lower energy prices, use almost no water, and that most moats in AI are an illusion. He explains Crusoe's pivot from Bitcoin mining to AI infrastructure, accelerated by ChatGPT's November 30, 2022 launch, and how Crusoe built Abilene's first 200-megawatt buildings in one year when 34 competitors bid two and a half years, vertically integrating electrical manufacturing to cut a 100-week power distribution lead time to 28 weeks. He names energy and skilled labor as the true bottlenecks, concedes 50% of planned data centers may never be built, and frames Crusoe's three products—data centers, GPUs, and tokens—through an 'AI super major' analogy to Exxon's vertically integrated hedge. On GPU depreciation, he argues six-year cycles underestimate H100s, whose rates today exceed launch pricing, defends take-or-pay rental contracts, and predicts closed models capture more dollars while open models generate more tokens.

How LPs Allocate to Venture in 2026: What They Want, What They Don’t | Baylor University CIO

Sep 14, 2026 · 1:16:29

David Morehead, Chief Investment Officer of Baylor University's $2.6 billion endowment, argues that velocity of capital matters more than fund multiples — a 3x growth fund compounding every six years beats a 15x venture fund held 18 years. He explains why Baylor holds about 2.5% of its endowment in Anthropic, why it bought the software crash down 50-60% by betting on human behavior, and why growth equity is its largest private allocation. Morehead details his rule that returns cannot be discussed without time, position sizing of $3 million per underlying company, and why managers must stick to the strategy they raised on. He names permitting as the data center bottleneck, is bearish on Europe, calls private credit overhyped, and argues biotech could be the next major opportunity.

Jensen Huang Declares AGI Has Arrived | Tesla Launches Cybercabs | Index Pulls Out of Town

Sep 10, 2026 · 1:22:31

Rory O'Driscoll of Scale and Jason Lemkin tell Harry Stebbings that Jensen Huang's AGI declaration is a 'bullshit term' since only coding matters, while the AI assistant race between Instinct, Grok Bot and Meta hinges on distribution. They debate whether legal AI firms Harvey and Legora can match coding AI — Rory pegs legal's AI take at 10–15% versus 30–50% in coding — and warn goal-seeking agents, as in OpenAI's undisclosed DSC Wiki incident, pose a cybersecurity crisis. They assess Tesla's Cybercab as a Waymo rival, Index exiting Town's round after Instinct objected, and Anthropic abandoning its $6 billion Descartes deal. They close on Robinhood underwriting Aura's IPO, Wonderful's $5 billion raise with $170 million in secondary, and Thinking Machines' $40 billion valuation.

How to Build Your Own Data Center & Why Every Startup Should Do It

Sep 5, 2026 · 1:04:40

Cliff Weitzman, co-founder and CEO of Speechify, tells Harry Stebbings that owning NVIDIA GPUs beats renting because renting costs 1.5x the purchase price yearly, so Speechify pays $100,000 per GPU to get Rubins and B300s early. He covers inference versus training workloads, liquid-cooling and insurance logistics, and NVIDIA's 25% buyback underwriting with Blackstone, BlackRock, Apollo and Goldman Sachs. Weitzman admits his biggest mistake was staying B2C as ElevenLabs leapfrogged Speechify through B2B APIs, and defends entering B2B against ElevenLabs and Bret Taylor's Sierra. He also argues founders should hire math olympiads over coders, details agent orchestration with Claude Code and Cursor, and explains how GPU clusters help him sequence genomes to cure his brother's orphan disease.

NVIDIA Crushes Quarter | OpenAI Cuts Off Cursor | Instinct Hits $2.5B Valuation

Sep 3, 2026 · 1:20:29

Harry Stebbings hosts SaaS investors Jason Lemkin and Rory O'Driscoll breaking down NVIDIA's record $96.2 billion quarter, its $12.9 billion pursuit of Hugging Face, and 70% growth guidance signaling compute demand runs at least another 12 months. Lemkin defends OpenAI's cutoff of Cursor as rational, since coding is the model load for LLMs and Cursor competes directly. The panel argues every startup must now become a compound startup, as AI makes code so cheap to produce that teams ship 100 times more software. Lemkin lays out bull cases for Clay at $7 billion and Linear at $2.5 billion, claiming agents will consume 10 to 100 times more GTM resources than humans ever could. They also flag the Hugging Face agent hack as a wake-up call for CISOs and praise Benioff's pragmatism in embracing multi-surface access and outcome-based pricing at Salesforce.

Only 10% of Neo-labs Will Survive | Factory CTO, Eno Reyes

Aug 29, 2026 · 1:29:23

Eno Reyes, Factory CTO, says the cheapest AI model isn't the cheapest system: outcome-based pricing can make frontier models cheaper, and predicts 99% of workflows will run on open models in three years but the remaining 1% holds most economic value. He calls 'Chinese models' a psyop by frontier labs, says frontier valuations are overweighted, sees Anthropic's $2 trillion Claude Code bet as risky due to model lock-in, and credits Microsoft with the best AI hand. Reyes says model routing is commoditized, warns 80-90% of neo labs could die in 18 months unless tied to durable workflows, and insists enterprises must own their intelligence rather than let labs control it. He rejects pedigree in hiring, wants token budgets funding outcomes, and expects SaaS to sell before cannibalization.

Anthropic's $30T Assumption & OpenAI Confirms IPO | Why Customer Service & Robotics are Overinflated

Aug 27, 2026 · 1:26:44

Jason Lemkin, Rory O’Driscoll and Harry Stebbings say the AI boom now depends on Nvidia and hyperscalers financing frontier models, as shown by Nvidia's $6B Poolside purchase and $1B at a $12B pre-money valuation. They dissect OpenAI's promised 2027 IPO, arguing it has no choice as Anthropic outraces it and 'it's all about code'; they call Anthropic's $30T TAM claim delusional and OpenAI's $200 plan a losing consumer business. Also covered: Mercor's $20B round, Hugging Face's $13B sale risk, Citadel unwinding 80% of Leopold's fund, Stripe's 41% growth, agents' data leaks, and why customer support, humanoid robotics and AI services are the dumbest categories. Their verdict: we're less than a third of the way through the cycle and addicted to tokens.

The AI Bubble WILL Burst | Should we be fearful of Chinese Open-Source | Jerry Murdock

Aug 22, 2026 · 1:10:09

Jerry Murdock, co-founder of Insight Partners, warns the AI bubble may burst by March if the Iran war continues, citing credit market complacency and Japan’s $1 trillion Treasuries. He predicts at least half of Neo Clouds disappear within 36 months, favors Fireworks over Base10, and argues open-source models plus ASIC chips will drive customization while frontier labs retain complex tasks. On security, he calls sandboxes the most underrated layer, naming Docker and E2B as leaders, and expects OpenRouter’s 5% markup to be disrupted by inference exchanges. He adds private equity is highly at risk to any financial dislocation, and in quick-fire bets Nvidia exceeds $10 trillion in five years while picking Meta as the Mag 7 short.