Give 2 million MIM tokens for equity! The Tokenmaxxing phenomenon behind OpenAI’s massive issuance of new startup tokens

OpenAI CEO Sam Altman announces providing YC startup teams with $2 million worth of tokens in exchange for equity, aligning with the "Tokenmaxxing" trend in the tech industry. He has previously advised entrepreneurs to focus on innovation rather than copying ChatGPT.

OpenAI distributes $2 million worth of tokens to YC startups

OpenAI CEO Sam Altman announced at the Y Combinator (YC) event on May 19 that they will provide all current YC participating startup teams with $2 million worth of OpenAI token usage (note) in exchange for equity in these startups.

This not only encourages young startups to develop using OpenAI’s models but is also compared within the industry to the practice of early investors providing funding to each startup.

Image source: X OpenAI distributes $2 million worth of tokens to YC startups **Note: Tokens are the smallest units of processing text in large language models, often called "符元" or "詞元" in Chinese. OpenAI’s **__API usage is billed per token, with costs varying depending on the model and whether the token is input, output, or cached.

The phenomenon of Tokenmaxxing behind OpenAI’s token distribution

Foreign media outlet The Information quoted sources familiar with OpenAI, stating that this proposal is part of a pilot program launched by OpenAI and YC for spring and summer participants, expected to cover hundreds of early startups.

The investment uses a simple agreement for future equity (SAFE) structure, allowing these startups to immediately access token usage now, with OpenAI’s equity share to be determined in future funding rounds.

OpenAI’s token distribution plan aligns with the "Tokenmaxxing" (token maximization) phenomenon in the tech industry.

Tokenmaxxing means that many employees in large tech companies are trying to maximize the consumption of tokens and credits, thereby demonstrating extensive use of AI.

However, for startups, tokens are an increasingly expensive resource, with actual costs depending on the chosen model and the frequency of cached or repeated inputs.

For example, $2 million can buy approximately 1 trillion GPT-5 tokens. In comparison, some power users within Meta consumed 73.7 trillion tokens over 30 days.

Additionally, Peter Steinberger, founder of OpenClaw who recently joined OpenAI, revealed he spent $1.3 million in just one month to purchase up to 603 billion OpenAI tokens.

• Further reading: Meta bets big on AI: Zuck uses Claude to code, employees launch token consumption battles to hit KPIs

Altman’s entrepreneurial philosophy: Don’t just build the next ChatGPT

Regarding how to create a startup product, Altman has given clear advice.

Last year, during a YC event, he said that he would feel sad if he heard someone claiming to build the next ChatGPT. Because OpenAI has already done very well in this area and holds a significant lead, entrepreneurs should focus on the vast space of innovation beyond core chat assistants.

Altman emphasized that people are easily influenced by peers, leading everyone to flock to the same concept at the same time.

He believes that the most sustainable successful companies usually do not follow the crowd to do exactly the same thing. This differentiation allows startup teams to buy valuable time, giving them the space to clarify what makes a good product and how to build the technology before having to answer defensive questions.

Image source: YC YouTube Sam Altman shares entrepreneurial insights from last year’s YC event Valuation surpasses $852 billion, paving the way for OpenAI’s IPO

While supporting startups, OpenAI is also preparing for an IPO.

According to CNBC, OpenAI completed a $122 billion fundraising at the end of March, pushing its valuation to $852 billion. Its rival Anthropic recently reported that some investors proposed valuing the company between $850 billion and $900 billion, raising an additional $40-50 billion.

As the IPO anticipation grows, OpenAI faces the pressure to justify this enormous valuation.

Currently, OpenAI’s monthly revenue reaches $40B, with annual revenue last year of $13.1 billion, but it is still consuming large amounts of cash and has not yet turned a profit. To control operating costs, OpenAI has recently begun cutting some large expenditure plans and shutting down products with lower revenue and user benefits, including AI video generation tool Sora App.

Related reports:
ChatGPT is used by 700 million people weekly! OpenAI CFO: AI costs plummet, similar to a steroid-enhanced Moore’s Law