Podcast Title: Better Offline
Host/Author: Cool Zone Media and iHeartPodcasts
Episode: OpenAI Is Not A Real Company
Release Date: February 26, 2025

Introduction

In the February 26, 2025 episode of Better Offline, host Ed Citron delves deep into a scathing critique of OpenAI, challenging the prevalent narrative surrounding the generative AI industry. The episode, titled "OpenAI Is Not A Real Company," scrutinizes the financial viability, business models, and overall impact of OpenAI and its peers within the tech landscape.

Ed Citron's Critique of OpenAI

Financial Unsustainability

Ed Citron begins by questioning the financial underpinnings of the generative AI sector, particularly focusing on OpenAI's economic model. He asserts, "Generative AI lacks the basic unit economics, product, market fit or market penetration associated with any meaningful software boom" (07:00). Citron emphasizes that despite the widespread hype, the industry remains financially precarious.

He presents OpenAI's financials starkly:

• Revenue vs. Expenditure: In 2024, OpenAI reportedly generated around $4 billion in revenue but incurred $9 billion in operational costs, resulting in a $5 billion loss (13:01).
• Venture Capital Dependency: The company’s survival hinges on continuous venture capital influx and support from hyperscalers like Microsoft, which provides significant compute credits to offset costs.

Burn Rate and Operational Costs

Citron highlights the exorbitant costs associated with running generative AI models:

• Compute and Storage: OpenAI spends approximately $2 billion on running models and $3 billion on training them, covering cloud compute, GPUs, and other infrastructure necessities (12:00).
• Talent Acquisition: Salaries alone consumed over $700 million in 2024, excluding stock-based compensation, which further strained financial resources (16:08).

Revenue Streams and User Engagement

Despite OpenAI's claims of substantial user bases, Citron challenges these figures:

• User Metrics: OpenAI boasts 400 million weekly active users, a number Citron deems misleading without clear definitions (22:00). He estimates actual monthly active users to be between 500 to 600 million, revealing a dismal conversion rate of approximately 2.58% from free to paying subscribers (28:08).
• Subscription Model Flaws: With 15.5 million paying customers generating around $3 billion, each subscription still results in a loss due to high operational costs (16:08).

Product Strategy and Market Fit

Citron critiques OpenAI's product offerings, arguing that they lack meaningful differentiation and fail to achieve product-market fit:

• AI Products: OpenAI’s attempts at monetizing products like Operator and Deep Research are labeled as compute-intensive and ineffective, further inflating costs without generating significant revenue (25:00).
• Commoditization of Models: Competitors like DeepSeq and Perplexity have replicated similar models at reduced costs, eroding OpenAI’s competitive edge (28:08).

Supporting Data and Analysis

Cost Breakdown

Citron provides a detailed breakdown of OpenAI’s expenditures:

• Training Costs: $3 billion on model training alone.
• Operational Compute: $2 billion dedicated to running the models.
• Total Spend: OpenAI’s total spending surpasses its revenue by a billion dollars, highlighting unsustainable financial practices (12:00; 30:14).

User Metrics Discrepancies

He challenges OpenAI’s user statistics by comparing them with independent estimates:

• Sensor Tower Data: Suggests that ChatGPT’s monthly active users are around 339 million on mobile platforms.
• Traffic Analysis: SimilarWeb reports 246 million unique monthly visitors to chatgpt.com in January 2025 (28:08).

Subscription Viability

Citron underscores the inefficacy of OpenAI’s subscription models:

• Revenue Loss: Each paying user contributes to continued financial losses due to high maintenance and operational costs.
• Low Conversion Rates: With a conversion rate below 5%, the subscription model fails to compensate for the overall financial deficits (28:08).

Product Strategy Flaws

Operator and Deep Research

Ed Citron scrutinizes OpenAI’s flagship products:

• Operator: Intended as an autonomous agent, it reportedly functions only 30% of the time, leading to increased user frustration and additional compute costs (25:00).
• Deep Research: Aimed at generating comprehensive reports, this product delivers low-quality outputs, necessitating further user engagement and exacerbating financial losses (25:00).

Lack of Innovation

Citron argues that OpenAI has stagnated in its product development:

• No Killer Apps: The absence of groundbreaking applications beyond ChatGPT limits market penetration and revenue diversification.
• Stagnant Model Evolution: Despite promises of GPT 4.5 and GPT 5, the enhancements offer negligible improvements, failing to drive significant user adoption or revenue growth (16:00).

Future Outlook and Predictions

Sustainability Concerns

Citron remains pessimistic about OpenAI's future:

• Burn Rate Projection: Forecasts for 2025 indicate a potential loss of $14.4 billion if current spending patterns continue (30:14).
• Dependency on Partnerships: Any disruption in OpenAI’s relationships with key partners like Microsoft and Oracle could irrevocably impact its financial stability (16:00).

Industry Implications

He extends his critique to the broader generative AI industry:

• Bubble Risk: The unsustainable financial models and lack of viable revenue streams suggest a looming bubble that could negatively affect tech valuations and the stock market (07:00).
• Media Complicity: The tech media’s unchecked promotion of OpenAI fosters an unrealistic perception of the industry’s success, obscuring underlying financial frailties (07:00).

Conclusion

Ed Citron's Better Offline episode "OpenAI Is Not A Real Company" presents a formidable argument against the perceived success and sustainability of OpenAI and the generative AI industry at large. Through meticulous financial analysis, user metric scrutiny, and product strategy evaluation, Citron exposes significant vulnerabilities that question the long-term viability of leading AI entities. His warnings highlight the potential repercussions for the tech industry, urging listeners to critically reassess the hype surrounding generative AI.

Notable Quotes

• Ed Citron on Generative AI's Economics:
  "Generative AI lacks the basic unit economics, product, market fit or market penetration associated with any meaningful software boom." (07:00)

• On OpenAI's Financials:
  "OpenAI spent $9 billion to make just under $4 billion in 2024." (12:00)

• Regarding User Metrics:
  "ChatGPT's monthly active users are somewhere in the region of 500 to 600 million. That's good, right? Its actual users are higher than officially claimed, right? No, it's bad." (28:08)

• On Product Strategy:
  "OpenAI’s models do not scale like traditional software... These platforms aren't innately compute heavy and so you don't need to have the same level of infrastructure to support the same amount of people." (22:00)

• Forecasting Future Losses:
  "OpenAI is on course to burn over $26 billion in 2025 for a loss of $14.4 billion." (30:14)

• On Industry Sustainability:
  "The generative AI industry outside of OpenAI is incredibly small, with little to no consumer adoption and pathetic amounts of revenue compared to the hundreds of billions of dollars sunk into supporting it." (31:00)

Final Thoughts

Better Offline delivers a compelling and critical examination of OpenAI’s role within the generative AI sector. By highlighting financial deficits, questionable business practices, and flawed product strategies, Ed Citron challenges listeners to rethink the narratives surrounding AI advancements. This episode serves as a crucial reminder of the importance of scrutinizing the economic realities behind technological innovations.