How do OpenAI and Anthropic differ in their business models?

OpenAI is primarily consumer-focused (3/4 subscriptions, 1/4 API), while Anthropic is the opposite (1/4 consumer, 3/4 API). OpenAI also has a more conservative revenue recognition, whereas Anthropic recognizes gross tonnage.

Is OpenAI losing market share in the consumer AI space?

Yes, OpenAI's market share for ChatGPT has reportedly dropped from 100% in 2023 to an estimated 75% in 2025, with Apple, Meta, and Windows expected to capture significant portions.

What are the potential economic models for consumer AI services?

It's likely that consumers will pay for value-added AI services similar to streaming subscriptions, potentially earning more than even current mobile plans, or through advertiser-supported models.

How is the increasing efficiency of AI impacting business valuations?

The potential for AI to disrupt businesses every 5-6 years is causing a rerationalization in capital markets, leading to lower valuations for SaaS companies and a flight to predictable, durable cash flows from large tech companies.

Will company brands become obsolete due to AI-driven abundance?

One perspective is that brands may go to zero as AI enables cheaper, faster, and better products, making abundance more valuable than brand affiliation. However, brands with strong network effects and defensible ecosystems may persist.

What is the legal fallout for Meta and YouTube regarding platform harms?

Meta and YouTube have been found liable for issues related to child exploitation and the harmful effects of addictive platform design, resulting in significant damage awards.

What is the balance between personal responsibility and corporate liability for platform harms?

While acknowledging corporate responsibility, the discussion emphasizes individual and parental responsibility in choices regarding screen time and social media use.

How can parents manage children's social media and AI app usage?

Parents can use parental controls and age assurance technology. Banning AI chat apps for kids is seen as a mistake, as learning to use these tools is crucial for future success.

What are the new challenges in enforcing age restrictions for online platforms?

Existing age verification methods like COPPA are easily bypassed. Social pressure among peers also makes it difficult for parents to maintain rules about which apps their children can use.

What is the role of the new PCAST council?

The PCAST council, with 'doers' and 'builders' as members, will advise on science and technology policy, including AI, nuclear power, quantum computing, and biotech, with a focus on the industrial race against China.

Key Moments

Anthropic's Generational Run, OpenAI Panics, AI Moats, Meta Loses Major Lawsuits

All-In Podcast

Entertainment8 min read81 min video

Mar 27, 2026|113,921 views|2,945|481

chamath david sacks david friedberg jason calacanis all in podcast tech news politics election stocks stock market palihapitiya

Save to Pod

Key Moments

TL;DR

Anthropic is experiencing a 'generational run' with rapid product releases and enterprise adoption, while OpenAI faces declining consumer market share and shifts focus, amidst increasing regulatory scrutiny for Meta.

Key Insights

Anthropic has launched multiple significant products in quick succession, including co-work, Opus 4.6, and computer use agentic systems, contributing to an estimated $6 billion in annual run rate added in February alone.

OpenAI's consumer market share for ChatGPT has declined from 100% in 2023 to an estimated 75% in 2025, even as the overall market grows.

Meta was found liable in two separate lawsuits, one resulting in a $375 million damages award in New Mexico for child predator access on Facebook and Instagram, and another for designing addictive platforms that harmed a young user's mental health.

The private equity approach to AI involves rolling up existing businesses like accounting and healthcare firms and integrating AI to manage change and drive efficiency, with some firms betting on their ability to own this transformation process.

A new President's Council of Advisors on Science and Technology (PCAST) has been appointed, featuring a mix of industry leaders and scientists, with a stated focus on areas like AI, nuclear power, quantum computing, and advanced semiconductors, in a race against China.

China's scientific output has surpassed the US, publishing 50% more research papers last year across various disciplines, indicating a significant shift in global scientific leadership.

Anthropic's rapid product cycle fuels enterprise growth

Anthropic is currently on a significant 'generational run,' marked by a series of high-impact product releases and strategic moves into the enterprise market. Since January, they've launched 'co-work' for business users, enabling complex tasks like cron jobs and integrating with platforms like Gmail and Notion. Their Opus 4.6 model has been hailed as a major step forward, with 'Jensen' Michael Dell describing it as an inflection point and the first agentic model capable of delivering real productivity gains in teams. This momentum continued in February with the release of claw code plugins, which, while causing some disruption in the SaaS market, also contributed to an estimated $6 billion in annual run rate added within that month alone. Most recently, Anthropic announced 'computer use,' a new agentic system that allows enterprise-grade open-claw functionality and enables users to control their desktop computers from their phones. This aggressive product development strategy, particularly their bet on coding as a key use case, has proven to be a smart business move, providing a gateway into enterprise IT budgets and driving rapid revenue growth, which then extends into other product areas like co-work and agent development. While some have philosophical objections to Anthropic's past advocacy for a 'permissioning regime' for chips and models, viewing it as a form of regulatory capture that could stifle new entrants, their product innovation and market traction are undeniable.

OpenAI faces consumer market share decline and strategic shifts

OpenAI, despite its strong consumer recognition with ChatGPT, is experiencing challenges in maintaining its market dominance. While ChatGPT remains a 'verb' in consumer parlance, its market share has reportedly fallen from 100% in 2023 to an estimated 75% by 2025, even as the overall market grows. Major players like Apple, Meta, and Windows are underrepresented but poised to capture significant market share, potentially pushing ChatGPT's consumer share below 50%. This shift is prompting OpenAI to re-evaluate its strategy. News has emerged of OpenAI shutting down its Sora video app project, which had attracted a potential $1 billion investment from Disney, and is reportedly cutting back on side projects to focus on the enterprise market, aiming to compete more directly with Anthropic. Additionally, OpenAI has offered private equity investors a guaranteed 17.5% return in a joint venture to deploy AI, suggesting a pivot towards more robust revenue generation models.

Meta faces significant legal challenges over platform harms

Meta has been hit with two major legal verdicts impacting its social media platforms, Facebook and Instagram. In New Mexico, a jury found Meta liable for enabling child predators to access minors on its platforms, leading to a $375 million damages award. An undercover investigation created fake child profiles that were contacted by predators, and a former Meta engineer testified about his own daughter receiving sexual solicitations on Instagram. In a separate case in Los Angeles, a jury found Meta and YouTube negligent for designing addictive platforms that harmed a young user's mental health. The plaintiff, who began using YouTube at age six and Instagram at age nine, testified that features like notifications and algorithms led to compulsive use, causing depression and anxiety. These verdicts highlight growing concerns about the design and impact of social media on young users and could set precedents for future litigation.

The evolving economics of AI: Consumer vs. Enterprise

The discussion around AI's economic future is bifurcated between consumer and enterprise models. While OpenAI has dominated the consumer space, its market share is eroding, leading to questions about its long-term monetization strategy. Some argue that consumer AI services will likely remain free, supported by advertising, with companies like Apple and Google leveraging their existing ad networks to integrate AI. Others, however, believe consumers will pay for premium AI services, citing the success of subscription models like Spotify and Netflix, and suggesting that AI's utility in managing daily life could justify significant subscription fees. Conversely, the enterprise market, where Anthropic has seen considerable success, is viewed as more stable and easier to monetize due to stickier customer relationships and higher willingness to pay. The unique position of companies like Google, with their vast existing data and free cash flow, allows them to operate separate consumer and enterprise AI plays, a luxury startups lack.

Private equity's role in AI adoption and 'change management'

A notable trend emerging in the AI landscape is the involvement of private equity (PE) firms. Instead of solely developing new AI technologies, some PE firms are adopting a strategy akin to 'roll-ups,' acquiring and consolidating existing businesses in sectors like accounting, legal services, and healthcare. Their bet is on their ability to manage the 'change management' process of integrating AI into these established industries. The argument is that simply introducing AI to businesses doesn't guarantee success; it's the effective deployment and integration that drive efficiency. By owning both the business and the change management process, these PE firms aim to unlock the latent value of AI. This approach addresses the statistic that a significant majority of enterprise AI pilots fail, suggesting a gap in AI implementation that PE firms believe they can fill.

Valuation recalibration: Amidst AI's potential abundance

The rapid advancement of AI, particularly the potential for 'super intelligence' and 'infinite abundance,' is forcing a recalibration of how companies are valued. There's a fundamental debate: are we on a path to a world where AI creates limitless possibilities, or is it simply the next generation of software? This question heavily influences investment decisions. Some data suggests a 'rerationalization' in public markets, with SaaS companies experiencing marked decreases in valuation multiples as the market anticipates increased disruption from AI. Conversely, large tech giants like Apple, Microsoft, Meta, and Alphabet are seeing their valuations increase, driven by the belief that their cash flows are 'monopolistically durable.' NVIDIA, despite its strong performance, is being valued more cautiously. This leads to a market dynamic where 'moats' like brands, network effects, and strong management teams are being scrutinized for their resilience against pervasive AI disruption, with some predicting that brands, in particular, may lose their pricing power.

The impact of AI on jobs and the future of work

The accelerating pace of AI development and adoption is having a profound impact on the job market and the very concept of work. One perspective suggests that as AI becomes more capable of performing tasks previously thought to require human expertise, traditional equity-based compensation models for employees might shift. If businesses are disrupted every few years, the long-term value of equity diminishes, potentially leading employees to demand higher immediate salaries. The analogy is drawn to the 'SAS apocalypse' scenario, where the rapid disruption of software models creates uncertainty in capital markets. The question arises: in a world of potential 'digital abundance' driven by AI, how does one value anything, especially future cash flows? This uncertainty could lead to a flight towards assets with perceived durability, such as physical experiences and foundational industries like energy and mining, or companies that can effectively integrate AI to expand productivity manifold.

US re-engages with science and technology policy on the global stage

In parallel with the AI and legal developments, the US is also refocusing on its science and technology policy, particularly in the context of global competition. A new President's Council of Advisors on Science and Technology (PCAST) has been appointed, featuring a high-profile mix of industry leaders and scientists, including figures like Mark Andreessen, Jensen Huang, and David Sacks, who will co-chair the council. This body aims to advise on critical areas such as AI, nuclear power, quantum computing, advanced semiconductors, and biotech. A key motivation for this renewed focus is the accelerating scientific output from China, which now publishes 50% more research papers than the US annually. This shift signifies a recognition that a race is underway not only in fundamental research but also in the industrialization of new technologies, necessitating a more 'doer'-centric approach to policy-making. The goal is to bolster US industrial policy and competitiveness in these critical fields.

Mentioned in This Episode

●Products

●Software & Apps

●Companies

●Organizations

●Drugs & Medications

●Concepts

●People Referenced

Common Questions

Anthropic has had a strong year with products like Claude 4.6, Co-work for business, and agentic systems. They bet on coding as a key use case, which has integrated them into enterprise IT budgets and led to revenue growth.