/
Navigation
Chronicles
Browse all articles
Explore
Semantic exploration
Research
Entity momentum
Nexus
Correlations & relationships
Story Arc
Topic evolution
Drift Map
Semantic trajectory animation
Posts
Analysis & commentary
Pulse API
Tech news intelligence API
Browse
Entities
Companies, people, products, technologies
Domains
Browse by publication source
Handles
Browse by social media handle
Detection
Concept Search
Semantic similarity search
High Impact Stories
Top coverage by position
Sentiment Analysis
Positive/negative coverage
Anomaly Detection
Unusual coverage patterns
Analysis
Rivalry Report
Compare two entities head-to-head
Semantic Pivots
Narrative discontinuities
Crisis Response
Event recovery patterns
Connected
Search: /
Command: ⌘K
Embeddings: large
TEXXR

Chronicles

The story behind the story

days · browse · Enter similar · o open

Sources: OpenAI's losses roughly doubled to ~$540M in 2022, and Sam Altman privately suggested the company may try to raise as much as $100B in the coming years

OpenAI's losses roughly doubled to around $540 million last year as it developed ChatGPT and hired key employees from Google

The Information Erin Woo

Context & Ripple Effects

In May 2023, The Information put a number on ChatGPT's cost: OpenAI's 2022 losses roughly doubled to ~$540M, driven by model development and hiring away key Google staff, even before ChatGPT generated meaningful revenue. The striking part is Altman's private framing — that OpenAI may need to raise as much as $100B in coming years — which reads less like fundraising chatter and more like a thesis that frontier AI is a capital problem first.

That framing aged into the company's operating reality rather than an outlier remark: within months Altman was telling staff OpenAI had reached $1.3B in annualized revenue growing 30% in three months, and by late 2025 OpenAI was projecting $115B of cumulative cash burn through 2029 — roughly $80B above what it had told investors previously.

First-order effects

  • OpenAI's cost structure is now public knowledge for every investor and partner: training compute plus poached-Google salaries doubled losses to ~$540M in a year when revenue barely registered, making any future raise a necessity, not an option.
  • Google loses directly twice — talent to OpenAI's hiring push and competitive secrecy around its own lab economics — while Altman's $100B signal tells capital markets to price OpenAI as a decade-scale infrastructure bet.

Second-order effects

  • Rival labs face a forced escalation: if Altman is privately contemplating nine-figure-millions-to-$100B scale raises, competitors must either match that capital intensity or cede the frontier — a dynamic visible later when Anthropic reached profitability on $11.6B of quarterly sales while OpenAI's losses jumped 32% to $12.3B in a single quarter.
  • Investors gain the first real loss baseline for a frontier lab, letting them anchor valuations to burn rates instead of demos — and pressuring OpenAI to convert its rapid revenue growth ($300M/month by August 2024) into a credible path against mounting losses.

Third-order effects

  • If the pattern holds — losses doubling early, then compounding into tens of billions of projected burn — frontier AI consolidates around a handful of labs able to raise at unprecedented scale, structurally separating well-capitalized players from the rest of the field.
  • A lab whose unit economics depend on repeated mega-rounds becomes dependent on continued investor appetite for unprofitable growth, making capital-market sentiment itself a systemic risk to AI progress timelines.

The trend: Frontier AI labs are converging on a model where capability leadership is purchased with escalating, multi-tens-of-billions capital commitments rather than earned through near-term profitability.

Discussion

  • @mtrc Mike Cook on x
    A lot of people say that private research is more efficient than academia, but I feel like I could get quite a bit done if I was allowed to lose $540m in a year. https://twitter.com/...
  • @amir Amir Efrati on x
    New: OpenAI losses doubled to $540m last year as revenue quadrupled. CEO Altman has discussed possibility of a $100B capital raise (!) More here: https://www.theinformation.com/ ... w/ @erinkwoo https://twitter.com/...
  • @baykenney Matthew Kenney on x
    AI is hard, and $100B seems reasonable given their objective https://twitter.com/...
  • @alexjc @alexjc on x
    Lending credibility to the hypothesis that OpenAI is not a business nor a non-profit, as it does neither. It's a project secretly funded by evil billionaires to destroy society as we know it—whatever the cost. They hired Altman to play puppet CEO role. (Oh wait! Wrong meeting.) h…
  • @kantrowitz Alex Kantrowitz on x
    Kind of wonder if this shiny new tech might just run into a business model problem https://twitter.com/...
  • @azeem Azeem Azhar on x
    OpenAI's losses last year were about the same as WeWorks' last quarter. I know who I would rather back. https://www.theinformation.com/ ...
  • @brianmcc Brian McCullough on x
    There is an angle here. They don't have a moat. You've seen this theory. This is about to become super interesting. https://twitter.com/...
  • @garymarcus Gary Marcus on x
    What if generative AI didn't really make that much money? Just ... reminiscing about the 100B that went into driverless cars & how that so far has not worked out for most of those who invested. (More about that in Episode 3 of Humans vs Machines, w @CadeMetz and Missy Cummings) h…
  • @maartengm Maarten Mortier on x
    I wonder how soon this side of the AI world will realise that it's difficult to become irreplaceable if you use natural text and natural interfaces.. People can very easily swap out your technology. The M2 chip is very hard to mimic, but an LLM? I'm unsure.. https://twitter.com/.…
  • @ericjackson Eric Jackson on x
    Why they were smart to partner with MSFT: https://twitter.com/...
  • @hammer_mt Mike Taylor on x
    They raised $10b, they can lose $500m every year for 20 years... why would they need to raise more? https://twitter.com/...