Anthropic: is expected to go public as early as **October 2026** with a $965 billion valuation, while **OpenAI** filed for an IPO in June 2026 and may list in **2027**.\n- OpenAI CEO **Sam Altman** has called any valuation below **$1 trillion** a "nonstarter," drawing comparisons to SpaceX's successful IPO, which soared past $2 trillion.\n- OpenAI generated roughly **$13 billion** in revenue in 2025 and is pacing toward **$24 billion** in 2026, but recorded a net loss of **$38.5 billion** in 2025.\n- **Palantir CEO Alex Karp** blasted the token-payment model, telling CNBC that "something has gone completely wrong" with the industry's pricing structure.\n- Enterprise customers including **Uber, Microsoft, Salesforce, and Meta** have begun rationing AI usage to control costs, as token-based pricing has proven more expensive than anticipated.\n- **Chinese open-weight models** like **Z.ai's GLM-5.2** and **DeepSeek** are emerging as cheaper alternatives, four to six times less expensive than frontier AI models.\n\n**Why this matters:** The AI industry's dominant pricing model is facing pushback just as its biggest players seek public market validation. If enterprise customers migrate to cheaper alternatives, the billion-dollar valuations of OpenAI and Anthropic could be at risk.
OpenAI officially filed for an IPO in early June 2026, but Altman has been candid about his reluctance to take the company public. "Am I excited to be a public company CEO? 0%," he revealed in December 2025. Despite this, the company needs the capital. With a net loss of $38.5 billion in 2025 and $8.5 billion in Q1 2026 alone, going public offers a faster path to raising cash compared to private funding rounds.
Altman likely drew inspiration from SpaceX's IPO, which saw public demand so strong that the company's market cap quickly topped $2 trillion. Interestingly, over 90% of SpaceX's claimed total addressable market is tied to AI, making the comparison between the two companies more relevant than it may initially appear.
The aggressive adoption of agentic AI in workplaces this year has given rise to the slang term "tokenmaxxing" — what happens when engineers, under pressure to demonstrate AI integration without clear guidelines, use AI models excessively. Executives have realized this is not an efficient business practice.
Companies including Uber, Microsoft, Salesforce, and Meta have taken steps to ration their employees' use of advanced AI because the pay-per-use token structure favored by Anthropic and OpenAI has proven more expensive than it's worth. Palantir's Alex Karp went further, describing excessive AI use without regard for value creation as "kind of like a porn addiction."
Beijing startup Z.ai's GLM-5.2 model is now ranked among the top 10 large language models by Artificial Analysis and ranks as the second-best model for web development on Code Arena. Crucially, it is four to six times cheaper than frontier AI models from US labs. Some US and international enterprise customers have already reported switching to cheaper Chinese models like DeepSeek, cutting back on payments to OpenAI and Anthropic.
The *Financial Times* reported that OpenAI has held talks with the Trump administration about granting the US government a 5% stake in the company. However, the proposal reportedly depends on other AI labs like Anthropic agreeing to similar terms. Experts warn that recent export controls on advanced US AI models may inadvertently accelerate international adoption of Chinese-developed models.
When will OpenAI go public?\nA: OpenAI filed for an IPO in June 2026, but the company has not set a firm date. Reports suggest it may not happen until 2027. CEO Sam Altman has said the company is "not in a hurry" to execute the public sale.\n\nQ: What is the "tokenmaxxing" controversy?\nA: Tokenmaxxing refers to the excessive use of AI models by engineers under pressure to demonstrate AI integration, without clear guidelines on efficiency. This behavior has led to skyrocketing costs under the pay-per-use token model, prompting companies like Uber and Microsoft to ration AI usage.\n\nQ: Why are Chinese AI models gaining traction?\nA: Chinese open-weight models like Z.ai's GLM-5.2 and DeepSeek offer comparable performance to frontier US models at four to six times lower cost, making them attractive alternatives for cost-conscious enterprises.\n\nQ: How does OpenAI's financial performance look?\nA: OpenAI generated approximately $13 billion in revenue in 2025 and around $2 billion per month in 2026, pointing to strong growth. However, the company recorded a net loss of $38.5 billion** in 2025 due to massive capital expenditures on data center infrastructure.
For investors:: Both OpenAI and Anthropic represent high-growth opportunities, but profitability remains elusive due to massive infrastructure costs. Compare their metrics with SpaceX's AI division for a relative valuation gauge.\n- **For enterprise customers:** Evaluate whether token-based pricing from frontier AI labs is cost-effective for your use case. Chinese open-weight models may offer significant savings without major performance trade-offs.\n- **For AI professionals:** The shift toward cost-conscious AI adoption means demonstrating ROI is more important than ever. "Tokenmaxxing" without clear business value could harm your case for AI investment.\n- **How to prepare:** Track your organization's AI spending per task, compare costs across models (including open-weight alternatives), and establish clear guidelines for when to use expensive frontier models versus cheaper options.
Will OpenAI and Anthropic justify their trillion-dollar valuations when they go public, or will cheaper alternatives erode their market share? Do you think the token-payment model is sustainable in the long run? Let us know in the comments below!
Share this with others who need to stay ahead of this trend!
Sources:
On July 24, 2026, a coalition of 25 major technology companies — including **Nvidia, Microsoft, Meta, Palantir, Google, IBM, and Hugging Fac...
On Saturday, July 25, 2026, OpenAI experienced a significant global outage that left thousands of users unable to access ChatGPT, the compan...
A 55-year-old Florida pastor, Scott Winters, has filed a lawsuit against OpenAI and its founder Sam Altman, alleging that medical advice fro...
On July 20, 2026, a federal court in San Francisco gave final approval to what is now the largest copyright class action settlement in US hi...
⚠ Disclaimer: Yanuki provides article summaries and links for reference only. Yanuki does not endorse, verify, or guarantee the accuracy of third-party sources. Please review original sources and verify information independently. Managed by the Yanuki Data Engine. Full Disclaimer