OpenAI’s Raising At $100B Valuation, Telegram’s Pavel Durov Arrested, NVIDIA's Just Okay Earnings

Episode Summary

Executive Summary: The episode examines three major tech stories: OpenAI’s unusual $100B fundraising round and its uncertain path to profitability, Pavel Durov’s arrest in France over Telegram’s alleged failure to police illegal content, and NVIDIA’s still-strong but moderating earnings. The hosts argue that strategic motives, platform responsibility, and AI infrastructure competition matter as much as financial metrics.

Main Topics: OpenAI’s $100B fundraising round (Priority: 5/5): The hosts discuss OpenAI’s reported fundraising at a $100 billion valuation, noting the round’s unconventional structure, strategic investors, and concerns that the company is still not clearly on a path to profitability. Strategic investors and OpenAI’s messy cap structure (Priority: 5/5): Apple, NVIDIA, Microsoft, and Thrive Capital are all discussed as potential participants, with emphasis on how OpenAI’s nonprofit structure makes the deal unlike a normal equity investment and more like a strategic alliance. Meta AI and Llama as competitive threats (Priority: 4/5): The conversation compares OpenAI with Meta’s AI efforts, arguing that Meta AI’s usage numbers may be inflated by accidental use, while Llama poses a stronger competitive threat because of open-source distribution and product integration. Enterprise AI and the future of coding (Priority: 4/5): Amazon’s AWS CEO suggests most developers may not be coding in two years, and the hosts debate whether natural language interfaces will expand access to software creation and redefine what it means to be a developer. Pavel Durov’s arrest and platform accountability (Priority: 5/5): The hosts analyze Durov’s arrest in France over Telegram’s alleged role in illegal content distribution, arguing that while platform liability is a difficult issue, Telegram’s reported lack of moderation and response made enforcement more justifiable. NVIDIA earnings and the AI boom (Priority: 4/5): NVIDIA’s results are described as excellent but less explosive than before, signaling normalization rather than collapse. The hosts also discuss Blackwell chip complexity and whether the company can sustain its growth trajectory.

Key Arguments: OpenAI’s fundraising is strategically important, but its nonprofit-plus-profits structure makes it unlike a normal venture investment and creates unclear investor incentives. The leak of the valuation likely serves signaling purposes to shape market perception and attract capital, which is a warning sign rather than a healthy market dynamic. Microsoft may invest less than before because it is building its own AI stack while still maintaining a strategic relationship with OpenAI. Thrive Capital’s repeated involvement suggests some insiders are doubling down even as broader Silicon Valley skepticism grows. Meta AI’s usage figures may be overstated because many users encounter it unintentionally inside Instagram and Facebook flows. Llama is the more serious competitive threat to OpenAI because open-source model distribution changes the economics of AI deployment. AI is likely to change coding by making natural-language software creation more accessible, reducing the need for traditional manual coding skills. Telegram’s founder should not be treated as exempt from accountability if the platform consistently ignores clear criminal abuse and refuses to cooperate with authorities. NVIDIA’s results show normalization, not weakness: growth remains strong, but expectations were so extreme that even excellent quarters now look merely good. Blackwell’s complexity raises manufacturing risk, but it may also be necessary for NVIDIA’s next wave of growth.

Data Points: OpenAI valuation: $100 billion to $125 billion - Reported fundraising valuation range for OpenAI’s new round OpenAI annualized revenue: $3.4 billion - Cited as OpenAI’s annualized revenue in the discussion OpenAI projected loss: Up to $5 billion this year - Based on reporting referenced during the conversation ChatGPT weekly active users: 200 million - Official usage figure cited in the episode OpenAI GPT-4 build cost: More than $100 million - Mentioned in the context of OpenAI’s spending and losses Meta AI monthly active users: At least 400 million - Reported by The Information and discussed as potentially inflated Meta AI daily active users: 40 million - Attributed to two current employees, with caveats about accidental usage Llama downloads: Almost 350 million - Zuckerberg’s broadcast message cited in the conversation Llama monthly usage growth: 10x since start of the year - From Zuckerberg’s internal/broadcast messaging Telegram employee count: About 30 employees - Used to emphasize the scale mismatch between Telegram’s workforce and its user base Telegram users: Around 900 million - Referenced while discussing moderation challenges Durov net worth: About $15.2 billion - Mentioned to frame his public profile and wealth NVIDIA quarterly performance: Sales and profit doubled year over year - Summarizing the earnings results NVIDIA guidance beat: 2% ahead of Wall Street expectations - Next quarter revenue guidance compared with consensus Prior NVIDIA guidance beat: 28% above consensus - Compared with the same period last year Blackwell chip cost: $40,000 - Value at risk for a defective Blackwell chip Internet Explorer market share: 94% in 2004 - Historical example used to discuss browser dominance Telegram moderation staffing: 30 employees for 900 million users - Used to argue that minimal moderation is structurally insufficient

Pivotal Quotes: "This is the weirdest company in existence." — Alex: Reflection on OpenAI’s unusual nonprofit structure and unconventional fundraising model "If you go forward 24 months from now... it’s possible that most developers are not coding." — Matt Garman (quoted by Alex/Ranjan): Discussing AI’s potential to change software development at AWS "I think that to not do anything sets a horrible precedent, a much worse precedent than a founder getting arrested." — Ranjan: Arguing that Telegram’s lack of moderation justified serious legal consequences

Implications: The episode suggests AI competition is shifting from pure model performance to distribution, partnerships, and platform control. It also signals growing pressure on tech founders to police harmful content, while NVIDIA’s slowdown hints the AI boom may be normalizing rather than ending.