Key Moments
Elon’s Anthropic Deal, The Next AI Monopoly?, “FDA for AI” Panic, Trading the AI Boom
All-In PodcastKey Moments
SpaceX's Elon Musk is leasing massive data center capacity to Anthropic, turning Elon Web Services (EWS) into a hyperscaler competitor and potentially enabling SpaceX's IPO by subsidizing XAI's development with $4-5 billion in annual revenue.
Key Insights
Anthropic has tripled its ARR from $10 billion to $30 billion between January and March and then accelerated to $44 billion ARR in April, with expectations to reach $100 billion ARR by year-end.
The SpaceX-Elon Musk deal is projected to generate an incremental $4-5 billion in revenue for SpaceX this year through Elon Web Services (EWS), complementing their core businesses.
The White House is reportedly considering an 'FDA for AI' model to vet new AI models for safety, with Anthropic's Mythos model cited as a catalyst due to cybersecurity concerns.
The labor participation rate in the US has dropped from a high of 63.3% to 61.9% in March, alongside mixed signals for recent college graduates' employment.
S&P 500 operating margins have increased from 11% in 2023 to 13% in Q1 2024, with a continued forecast of expansion over the next two years.
Despite AI's perceived benefits, it ranked 29th out of 39 issues in importance to voters, with cost of living and the economy being the top concerns.
Elon Musk's strategic data center and compute play with Anthropic
The discussion kicks off with a significant development: Elon Musk, through SpaceX, has leased all of Colossus 1, a massive data center, to Anthropic. This move positions SpaceX as a major competitor in the hyperscaler market, competing with AWS, Azure, and GCP, under what's being dubbed 'Elon Web Services' (EWS). This deal is seen as a critical move to address Anthropic's and OpenAI's compute constraints, which are the primary bottlenecks limiting their revenue growth. According to Brad Gerosner, this deal is estimated to generate an incremental $4 to $5 billion in revenue for SpaceX this year, which will be crucial in subsidizing the significant investments XAI is making in developing its own models, like Grok. Chamath Palihapitiya emphasizes that the revenue performance of AI companies is solely dictated by the availability of compute and power, not demand. He suggests that if these companies had infinite resources, their revenues would be even more parabolic. The escalating demand for AI compute is creating supply constraints, exacerbated by protests against new data center construction, which could further restrict supply. This creates a prime opportunity for Musk, whose data center infrastructure, built ahead of the curve, becomes a critical asset. This strategy not only bolsters SpaceX's valuation narrative ahead of a potential IPO but also allows XAI to train its models without the immediate pressure of generating its own substantial revenue.
Anthropic's 'insane' growth trajectory and the monopoly question
David Sacks highlights Anthropic's unprecedented growth, stating the company tripled its Annual Recurring Revenue (ARR) from $10 billion to $30 billion between January 1st and March 31st of the current year. Remarkably, this growth accelerated in April, pushing their ARR to $44 billion. This kind of exponential growth at such a large scale is unprecedented in Silicon Valley, even for a region accustomed to rapid expansion. Sacks predicts Anthropic is on track to hit an ARR of approximately $100 billion by the end of the year, and forecasts they could potentially reach $1 trillion in ARR by 2027. This trajectory leads Sacks to suggest that Anthropic might become the 'most powerful monopoly ever created in human history,' or what he terms 'AGI' from a monopolistic standpoint. He contrasts this with the hyperscalers' revenue growth, which, while substantial, is in the 20% to 30% range, significantly lower than Anthropic's over 100% growth. The potential for Anthropic to reach $1 trillion in ARR would dwarf even the combined market cap of the current 'Mag 7' tech companies, underscoring the immense value being created. While Chamath acknowledges the potential for unlimited TAM, he cautions that such forecasts might be ambitious, yet agrees that the trajectory is undeniably exponential and could lead to a dominant market position.
The emergence of 'Elon Web Services' and its market implications
Brad Gerosner elaborates on Elon Musk's strategy, referring to the five-layer cake of SpaceX's business: launch, connectivity (Starlink), compute, hyperscaler, space data centers, and applications/models. The deal with Anthropic directly addresses the compute and hyperscaler layers, validating Musk's long-term vision and potentially justifying SpaceX's high valuation. By securing Anthropic as a major client, SpaceX can monetize its data center investments, generating significant revenue that subsidizes XAI's development and reduces its reliance on immediate revenue generation. This move effectively turns XAI into a competitor in the hyperscaler market, a space currently dominated by tech giants with market caps in the trillions. Chamath further emphasizes the strategic advantage of Musk's early investment in data center infrastructure and securing power. This foresight positions him as a 'kingmaker' in the AI landscape. The ability to offer compute services makes it easier to justify SpaceX's valuation, especially as it prepares for an IPO. The scale of these data centers, with facilities like 'Macro hard' and 'Macro harder' boasting 1.2 gigawatts of power, highlights the massive infrastructure underpinning this strategy. The deal with Anthropic, providing H100 GPUs ideal for inference, demonstrates a symbiotic relationship where Anthropic gains much-needed compute, and SpaceX monetizes its assets.
The 'FDA for AI' debate and regulatory concerns
A significant portion of the discussion revolves around the White House contemplating an 'FDA for AI,' reportedly spurred by concerns over Anthropic's 'Mythos' model and potential cybersecurity risks. This proposal suggests a review process for new AI models akin to drug approvals. However, the panel largely expresses skepticism about this approach. Brad Gerosner argues that an FDA-like approval regime would be disastrous, potentially leading to Washington picking winners and losers, slowing down innovation, and creating regulatory capture. He clarifies that his conversations with officials like Kevin Hassett indicate a preference for coordination and government capacity building, not pre-approval. Chamath adds that this 'FDA for AI' narrative is partly 'fake news,' amplified by media interpretations, and that many senior officials do not support such a stringent regulatory approach. He suggests that the real issue is a negative 'vibe shift' around tech and AI, and a lack of positive messaging from the industry about AI's benefits. Sacks echoes these concerns, warning that such regulations could create a moat around a duopoly and stifle competition.
Counterarguments against the monopoly narrative and calls for competition
Brad Gerosner pushes back against the 'monopoly' label for Anthropic and OpenAI, pointing out that their current revenues, while growing rapidly, are still relatively small compared to established giants like Google, which has substantial AI revenues and significant free cash flow. He stresses the importance of competition in driving AI advancement and urges Washington to stay out of the way. He views the rapid innovation and competition among the major AI labs as crucial for America's leadership in the field. Chamath agrees that competition is the 'northstar' and that regulatory capture by monopolists is a concern. He highlights that Anthropic banning competitors from using its models, as with OpenClaw, could be viewed as anti-competitive behavior. While acknowledging the need for guardrails, he argues that the 'FDA for AI' rhetoric might be a tactic to create a stronger moat around emerging monopolies, distracting from the real objective of fostering competition and innovation.
Reframing AI's narrative: from fear to innovation and societal benefit
Chamath and David Sacks advocate for a shift in how AI is perceived, moving from a narrative of fear and job displacement to one of innovation and societal benefit. They propose strategies for the AI boom to directly benefit the broader population. This includes encouraging major AI companies to engage in significant 'giving' initiatives, such as 'IPO K' or 'IPO for kids,' where a portion of IPO proceeds goes into public investment accounts for citizens. This would ensure that the immense wealth generated by AI is more broadly distributed. Sacks also emphasizes AI's potential to revolutionize healthcare and education, reducing suffering and costs. He suggests that capitalists should engage more directly with issues like the minimum wage and universal healthcare, arguing that increased consumer spending power benefits the economy. The panel agrees that while AI development is crucial, the narrative must include concrete examples of how it improves lives, such as through advanced medical diagnostics, personalized education, and solving complex societal problems. The current focus on potential AI-driven job losses is seen as overshadowing the immediate and potential positive impacts on productivity and economic growth.
Economic outlook: AI's contribution to growth and market performance
The discussion turns to the broader economic picture, with panelists noting the strong performance of the market, driven in large part by AI-related companies and cloud computing. AWS, Azure, and Google Cloud are reported to have significant ARR growth, with Google Cloud showing a remarkable 63% growth rate. Tech stocks, including Meta and NVIDIA, are trading at multiples that, while high, are not seen as bubble territory by some. Chamath notes that AI is contributing substantially to GDP growth, with 75% of Q1 GDP growth attributed to it. He highlights a construction and blue-collar boom with significant wage increases, challenging the narrative that AI benefits only a select few in Silicon Valley. The consensus among the panelists is that current economic policies are working, leading to accelerating GDP and controlled inflation, with AI being a significant tailwind. Despite some criticisms regarding tariffs or geopolitical conflicts, the overall sentiment is that the US is in a strong economic position, leading the world in innovation.
The ROI of AI: measuring productivity gains and future economic impact
A key point of contention emerges regarding the tangible return on investment (ROI) from AI, particularly concerning productivity gains and their impact on corporate margins. Chamath argues that while there's a lot of experimentation and spending on AI tokens, there's not yet clear evidence that this has directly lifted operating margins across the S&P 500. He posits that a critical 'reckoning moment' may be two to three years away, when it will become clear whether AI leads to reduced operating expenses or expanded revenues with stable or increased costs. Conversely, Sacks and Gerosner point to evidence of margin expansion in companies like Microsoft and Google, as well as the broader S&P 500, and the fact that enterprises continue to spend heavily on AI, suggesting a belief in future ROI. They cite examples of companies reducing headcount while increasing output, and ad creative becoming more effective and cost-efficient due to AI. Gerosner believes the ROI is 'fate complete,' especially in startups, where AI enables smaller teams to achieve more. However, Chamath maintains that until companies can trace a direct 'x spent, y made' where y is demonstrably greater than x and lifts margins, the full economic benefit remains unproven, though he acknowledges the potential is immense.
Mentioned in This Episode
●Software & Apps
●Companies
●Organizations
●Concepts
●People Referenced
Common Questions
The deal provides Anthropic with crucial compute capacity, addressing their shortage and allowing them to scale their AI models. For Elon Musk, it monetizes his data center infrastructure and supports his broader AI ambitions, potentially providing a structural base for companies like XAI.
Topics
Mentioned in this video
A major cloud computing service provider, mentioned as a competitor to Elon Musk's EWS.
A major cloud computing service provider, mentioned as a competitor to Elon Musk's EWS.
Platform where bets are placed on future events, used to bet on Trump's AI review timeline.
AI model from XAI, the development of which is subsidized by Elon Musk's compute infrastructure.
AI model from Anthropic experiencing increased usage and rate limits, with adjustments made by the company.
Microsoft's cloud computing service, mentioned as a competitor to Elon Musk's EWS.
New base model for GPT-5.5, indicating optimism for continued improvements at OpenAI.
AI coding assistant that XAI has partnered with, indicating XAI's efforts to improve its coding product.
Anthropic's AI model, which reportedly caused concern at the White House regarding safety.
Mayor of Los Angeles, described as extremely left-wing.
CEO of Apple, whose innovation strategy was contrasted with Elon Musk's approach.
Political candidate in Los Angeles whose campaign ads and debating skills were discussed.
White House Chief of Staff, who released a statement that downplayed reports of an 'FDA for AI'.
Politician who has expressed support for certain Trump administration AI initiatives and has called for a ban on data centers.
Mentioned in reference to a city councilwoman's political alignment.
Commentator credited with initially raising the idea of an 'FDA for AI'.
CEO who is relocating his company's efforts to Florida due to safety concerns in Chicago.
Founder of SpaceX and CEO of Tesla, discussed for his AI compute strategy and data center deals.
Director of the National Economic Council, confirmed reports about the White House's AI safety considerations.
Guest on the podcast who commented on AI safety and the balance between innovation and regulation.
Historical figure known for monopolistic practices, used as an analogy for AI companies.
A software that Anthropic banned from using its models, an action perceived as anti-competitive.
Tech company whose innovation, particularly Apple TV, was discussed in contrast to Elon Musk's ventures.
Cybersecurity company mentioned as a leader in AI-powered defense.
Elon Musk's AI company, which is trading models at Colossus 2 and facing financial challenges.
AI company whose revenue performance is linked to compute and power supply, and is a key competitor in the AI race.
Memory chip manufacturer whose stock valuation is noted as being very low.
AI company that has leased data center capacity from Elon Musk, experiencing increased demand and rate limits.
Provider of GPUs, mentioned as part of Anthropic's compute expansion and relevant to AI hardware.
Company founded by Elon Musk, discussed for its data center capacity and its potential IPO valuation.
Tech company whose valuation and performance were discussed in relation to the market and AI boom.
Company led by Elon Musk, whose factory-building expertise is compared to data center construction.
Company founded by John D. Rockefeller, used in a thought experiment about renaming for public perception.
Cybersecurity company mentioned as a leader in AI-powered defense.
Memory chip manufacturer whose stock valuation is noted as being very low.
Tech company whose stock valuation is noted as being very low.
Country whose minimum wage policies were studied as a potential model.
Country whose minimum wage policies were studied as a potential model.
Country whose minimum wage policies were studied as a potential model.
Country whose minimum wage policies were studied as a potential model.
City mentioned as a potential location for data centers, with a proposal for free electricity.
More from All-In Podcast
View all 408 summaries
96 minHow We Grew Koch Inc. to $150 Billion Without Going Public: Charles & Chase Koch
72 minSpencer Pratt on Fixing LA: Wildfires, Homelessness, Corruption & the Fight to Take It Back
81 minOpenAI Misses Targets, Codex vs Claude, Elon vs Sam Trial, Big Hyperscaler Beats, Peptide Craze
69 minCA Governor Candidate Steve Hilton on Why California is Destroying Itself & How a Republican Can Win
Ask anything from this episode.
Save it, chat with it, and connect it to Claude or ChatGPT. Get cited answers from the actual content — and build your own knowledge base of every podcast and video you care about.
Get Started Free