Divya van Mahajan

Is Google Now the AI Leader?

ai-daily-brief-podcast

Is Google Now the AI Leader?

Overview

This episode of the AI Daily Brief (recorded September 4, 2025) examines whether Google has taken the lead in the AI race, using the resolution of Google’s antitrust case (the Chrome divestiture ruling) as a launching point for a broader retrospective on Google’s journey from perceived laggard to potential frontrunner. The episode also covers major headlines including Anthropic’s latest funding round, OpenAI’s internal restructuring, and ongoing talent wars across the major AI labs. The host is Nathaniel Whittemore (implied by show format and style), producer of the AI Daily Brief podcast and video series.

Source video URL not provided.

Prerequisites

  • Familiarity with the major AI labs: Google DeepMind, OpenAI, Anthropic, Meta, xAI
  • Basic understanding of large language models (LLMs) and generative AI products
  • Awareness of the 2022–2023 ChatGPT launch and its disruption of the AI landscape
  • General knowledge of U.S. antitrust law and the DOJ’s case against Google’s search monopoly
  • Familiarity with AI product categories: chatbots, coding agents, image/video generation, multimodal models

Main Points

Anthropic Closes a Record Funding Round

  • Anthropic raised $13 billion at a $183 billion valuation, co-led by Fidelity and Lightspeed Venture Partners, with notable participation from private equity, sovereign wealth funds, and retirement funds.
  • Previous round was $3.5 billion at a $61.5 billion valuation (February 2025), representing a more than 3× increase in valuation in months.
  • Annual Recurring Revenue (ARR) grew from $1 billion to $5 billion year-over-year; the company now claims 300,000 business customers.
  • Large accounts (spending >$100K/year) grew sevenfold; Claude Code alone accounts for $500 million in revenue, with usage up 10× since its June launch.
  • The scale of this round signals that the AI industry has outgrown Silicon Valley venture capital as its primary funding source.

OpenAI Restructures Around Applications

  • OpenAI acquired analytics/A-B testing platform Statsig for $1.1 billion (all-stock), one of its largest acquisitions to date.
  • Statsig CEO Vijay Raji joins as CTO of Applications, overseeing ChatGPT, Codex, and future product lines; Statsig continues operating independently.
  • Fiji Simo is now officially CEO of Applications; the company is drawing a sharp organizational line between its applications business and its research division.
  • Chief Product Officer Kevin Weil moves to the research side as VP of AI for Science; his former product team (including ChatGPT head Nick Turley) now reports to the CMO.

Talent Wars: Apple, Meta, and the Lab Ecosystem

  • Apple is experiencing significant AI talent departures: its lead AI robotics researcher joined Meta, and three LLM team members left for Anthropic and OpenAI; roughly 10 members of the Foundation Models team have departed recently.
  • Apple is reportedly considering outsourcing its LLM needs to Google or Anthropic, which has worsened internal morale; nearly all remaining researchers are said to be actively interviewing elsewhere.
  • Meta’s Superintelligence Lab (TBD Labs) faces reporting of internal friction: three high-profile recruits departed (two never formally started; one left after a month); Scale AI SVP Ruben Mayer also left after two months (citing personal reasons).
  • ChatGPT co-creator Shangjia Zhao reportedly signed paperwork to return to OpenAI within days of joining Meta before Meta appointed him Chief AI Scientist.
  • Meta is now partnering with alternate data labeling vendors (Surge, Mercor) instead of working exclusively with Scale AI.
  • Host’s assessment: given the extreme aggressiveness of Meta’s recruiting, some attrition is expected (“Occam’s razor”); the real test will be what products Meta ships.

Google’s Fall and Rise: A Retrospective

  • Google declared itself an “AI-first company” in 2016 under Sundar Pichai, yet was caught flat-footed by ChatGPT’s launch in late 2022—just as it had been surprised by Amazon Alexa years earlier.
  • For most of 2023, Google’s AI story was one of catching up: BARD and other announcements failed to approach GPT-4 capability; the Gemini announcement in December 2023 was widely seen as rushed, with top-performing models not yet available.
  • Early 2024 brought further reputational damage: an image generation model was pulled after generating historically inaccurate and offensive images; AI Overviews in search launched with high-profile errors (e.g., suggesting glue as a pizza ingredient).
  • Internally, Google spent 2024 consolidating fragmented AI efforts under DeepMind and Demis Hassabis; the Gemini app team was not fully folded into DeepMind until October 2024.

NotebookLM Audio Overviews: Google’s First Genuine Hit

  • NotebookLM’s Audio Overviews feature (fall 2024) allowed users to auto-generate podcast-style audio summaries from uploaded documents or links—a genuinely new medium for consuming information.
  • Adoption was broad and sustained across use cases: legal (case file summaries), education (study tool), journalism, and general research.
  • This was framed as Google’s first “bona fide hit” in the generative AI era, and generated goodwill suggesting the prior struggles were the anomaly, not the norm.

Google’s 2025 Product Surge

  • Sundar Pichai signaled urgency at a December 2024 town hall: “2025 will be critical… we need to move faster as a company.”
  • At Google I/O (May 2025), Google listed approximately 100 announcements, a large portion AI-related; the company explicitly shifted to a “ship fast” ethos, releasing major models outside of keynote events.
  • Coding focus: Gemini 2.0 (coding-fine-tuned, February), Gemini 2.5 Pro Preview (May), and Gemini CLI (June) all targeted the developer/coding use case. Token processing grew from 480 trillion (May) to 980 trillion (July)—104% growth in two months, attributed largely to coding use case adoption.
  • Multimodal leadership:
    • Veo 3 enabled simultaneous video and audio generation (a first), driving a viral explosion of AI video on social media.
    • Genie 3 (experimental): generates interactive world environments from a prompt.
    • NanoBanana (Gemini 2.5 Flash Image): high-fidelity image editing that opened new economically useful use cases; compared in cultural impact to the earlier Studio Ghibli trend from OpenAI.
  • Andreessen Horowitz’s top 100 Gen AI consumer apps: Gemini (#2), Google AI Studio (#10), NotebookLM (#13), and Google Labs all ranked in the top 50—three Google properties in the top 15.

The Antitrust Ruling and Its Strategic Implications

  • A federal judge ruled that Google will not be forced to sell Chrome, and may continue paying to be the default search engine on third-party platforms (e.g., Apple’s Safari).
  • The sole prohibition: Google cannot arrange exclusive distribution deals.
  • Google must share search data freely with competitors to reduce its moat.
  • The court explicitly noted that AI startups are “already in a better position both financially and technologically to compete with Google than any traditional search company has been in decades, except perhaps Microsoft”—suggesting market forces, not divestiture, are the appropriate remedy.
  • Strategic upside for Google: it can now pursue similar distribution deals for Gemini on the iPhone, potentially replacing or supplementing its search deal with Apple (worth ~$20 billion/year to Apple).
  • Google stock rose ~9% on the news. Analysts noted the ruling allows Google to replicate its search-era distribution strategy for AI.
  • Rumors of Gemini 3 being imminent, with Semi-Analysis reporting it is shaping up to be highly performant on coding and multimodal capabilities.

Is Google Actually in the Lead? The Counter-Arguments

  • ChatGPT’s dominance in usage: Gemini has only ~12% of ChatGPT’s web visits; for a large segment of users, “ChatGPT” is synonymous with AI itself.
  • Meta’s potential: Despite talent turbulence, Meta’s aggressive acquisition of talent and compute means outcomes could look very different in 12 months.
  • xAI’s trajectory: Elon Musk’s xAI has made rapid progress and is on an “impressive trajectory.”
  • Anthropic’s momentum: Fastest-growing products in the space this year, particularly in the coding category.
  • Chinese AI labs: Continued release of competitive, often open-source models represents an ongoing wildcard.

Key Concepts

  • ARR (Annual Recurring Revenue): A measure of predictable, recurring subscription-based revenue normalized to a one-year period; used here to track Anthropic’s growth from $1B to $5B.
  • Acquihire: An acquisition primarily motivated by gaining the target company’s talent rather than its products or technology; referenced in the context of Meta’s Scale AI deal.
  • Multimodal AI: AI systems capable of processing and generating multiple types of data simultaneously (text, image, audio, video); identified as Google’s primary area of competitive advantage.
  • Veo 3: Google’s video generation model notable for producing synchronized video and audio in a single generation pass.
  • NotebookLM Audio Overviews: A Google feature that auto-generates podcast-style audio discussions from user-supplied documents; credited as Google’s first major generative AI consumer hit.
  • NanoBanana (Gemini 2.5 Flash Image): Google’s image editing model noted for high-fidelity, prompt-faithful image manipulation.
  • Genie 3: Google’s experimental world model capable of generating interactive environments from a text prompt.
  • Gemini CLI: Google’s command-line coding agent, part of its broader developer-focused AI push.
  • TBD Labs (Meta Superintelligence Labs): Meta’s internal AI research organization formed through aggressive recruitment from other major labs.
  • AI Overviews: Google’s feature that places AI-generated summaries at the top of search results; launched with notable early errors in 2024.
  • Polymarket: A prediction market platform; cited here as showing Google at a 75% probability of having the best AI model by end of 2025.
  • Statsig: An A-B testing and product analytics platform acquired by OpenAI for $1.1 billion to build out its applications infrastructure.

Summary

The episode argues that Google has undergone one of the most dramatic reversals in the AI race—from being mocked as a laggard caught flat-footed by its own technology in 2022–2023, to emerging in mid-2025 as the consensus frontrunner on prediction markets and in analyst commentary. This turnaround was driven by painful internal consolidation under DeepMind, a “ship fast” product culture exemplified at Google I/O 2025, breakout consumer hits in NotebookLM and NanoBanana, dominance in multimodal AI through Veo 3 and Genie 3, and strong coding infrastructure via Gemini CLI and Gemini 2.5 Pro. The antitrust ruling preserving Chrome and allowing continued distribution deals removes a major strategic overhang and positions Google to extend its reach through the iPhone as a Gemini distribution channel. However, the host stops short of declaring Google the definitive leader, noting ChatGPT’s commanding usage share, Anthropic’s explosive growth in coding, Meta’s unproven but resource-heavy superintelligence lab, and the persistent competitiveness of xAI and Chinese labs—framing the AI race as genuinely open, with Google nonetheless in its strongest relative position in years.