Speaker
Gavin Baker
Appearances over time
4 episodes
Episodes
4
Is AI a Bubble? | Gavin Baker on Data Centers, GPUs, and the AI Economy
Socialists Sweep NYC, China Catches Up in Coding, AI Memory Crunch, Micron's Blowout Quarter
All-In's Best Ideas Pitch Competition: 4 Investors Present Their Top Trades Live
Inside the Private Stock Market Boom: SpaceX, Anthropic, OpenAI & the Rise of Secondaries
Podcasts
Quotes & moments
At the peak of the 2000 internet/telecom bubble, 97% of fiber laid in America was dark — unused and unlit — the defining sign of overbuilding.
The biggest public GPU spenders have seen approximately a 10-point increase in return on invested capital since ramping up AI CapEx, signaling strong returns so far.
Cisco peaked at 150–180x trailing earnings during the 2000 bubble; NVIDIA trades at roughly 40x — a much more moderate valuation.
Large institutional funds face strict SEC constraints, self-imposing limits of 3% to 7% to manage risk.
Gavin Baker predicts that within the next 3 years, a number of high-profile custom AI chip (ASIC) programs will be canceled, especially if Google begins selling TPUs externally.
Google required three chip generations to make the TPU viable — a cautionary benchmark for other companies attempting custom silicon.
Baker analogizes ChatGPT to Netscape Navigator — at that equivalent moment in the internet era, Google had not yet been founded, underscoring how early AI still is.
AI coding tool Cursor has accumulated 1 trillion coding tokens, creating a data moat that makes it increasingly difficult for incumbent software companies to catch up.
Standing up a 1 gigawatt terrestrial data center costs approximately $35 billion in NVIDIA semiconductors plus $25 billion in power and cooling equipment.
With a reusable Starship, it would cost approximately $5 billion in launch costs to put a gigawatt of compute into orbit, making the total around $40 billion versus $60 billion terrestrially.
Google's Gemini has taken 15–20 points of web traffic share in just 2–3 months, suggesting Google may already be the largest AI product by traffic.
Gavin Baker stated his belief that Anthropic would trade at approximately $3 trillion as a public company, with revenue projected well over $100 billion by year end.
Pre-AI SaaS companies routinely achieved 80–90% gross margins, a level Baker says frontier AI companies structurally cannot match due to compute intensity.
Gavin Baker cited an uncontested study showing that electing a Republican DA causes all-cause mortality for young Black men in that city to drop by 7%.
Baker argues the debate between humanoid and non-humanoid robots is settled in favor of humanoids, since they can learn by watching video and accept direct human demonstration.
Staying private shield CEOs from hard questions, creating an echo chamber of sycophantic private boards. Public markets provide clean, brutal truth and pressure testing that ultimately makes companies stronger.
The alliance between Forge and Charles Schwab will democratize access to legendary private assets like SpaceX. This integration represents a major structural shift toward retail-accessible late-stage private equity.
Many venture capitalists focus exclusively on buying, forgetting that returning liquidity to LPs is the goal. Taking chips off the table during high-priced private secondary rounds is a fiduciary duty.
Mutual funds like Fidelity are capped at 15% private asset ownership, but voluntarily restrict exposure further. When private giants finally IPO, a massive wave of public market dry powder will flood back.
Sierra is building native agents for customer service and sales, redefining software. Although they face platform risk from base models, their sophisticated agentic layers make them highly attractive.
Revolut operates with a next-generation software stack, systematically unbundling traditional banks across Europe and the US. Their robust margins and massive scale make them a prime secondary target.
As chips disaggregate into specialized prefill and decode tasks, AI data centers must reinvent networking. ARIA and DriveNets represent the highly specialized infrastructure layer that will capture massive spend.
Neurorobotics is a German industrial robotics company capturing huge logistics revenues out of the limelight of Silicon Valley. They represent exceptional execution with over 100 million in revenue.
Zipline bypassed regulatory hurdles in US airspace by proving its autonomous drone tech delivering critical medicine in Africa. In doing so, they dropped maternal mortality rates by up to 95% in key regions.
Barry Diller's $48/share bid provides a floor, but MGM's real upside is a $2B EBITDA Osaka casino opening in 2030 and 300,000 sq ft of dark casino space in Dubai. Aaron Cowen says the stock could hit $100-$150 from the current high $40s.
Vegas assets (~$60/share) + Japan ($50/share) + Dubai optionality ($40-50/share) = $100-$150 stock from the high $40s. And Barry Diller, who is a financial buyer not a strategic one, has now shown his cards. Don't tender.
Power demand tracks GDP until a technological breakthrough hits — like appliances, then computers, now AI. The US had 2x China's power capacity two decades ago; China now has 3x ours. Talen Energy sits at the epicenter of a mandatory infrastructure buildout with no room for failure.
Talen Energy trades at a $25B enterprise value versus a $45B replacement cost. At 7x free cash flow on a $50/share FCF base, with comparable infrastructure at 15x, the equity should more than double just doing nothing. If power prices rise and data center deals accelerate, you get $1,050.
The PJM grid operator needs 106 gigawatts of new power in just one region of the US within 10 years — roughly the size of Japan's entire grid today. Supply chain constraints make this virtually impossible, locking in scarcity for existing producers like Talen.
Private secondary market trading has surged past its 2021 peak, with transactions competing with IPOs as the primary liquidity mechanism. Employee-backed secondaries reached 31% of all primary venture activity in 2025.
Analysis
What they talk about
- Business 46%
- Technology 42%
- Society & Culture 8%
- Government 4%
Connections
Shows they appear on and people they share episodes with. Drag to explore.
