OpenAI has reportedly seen renewed investor demand in private secondary markets over the past month, fueled by the launch of its GPT-5.6 family of AI models and increased traction with its Codex AI coding agent. This resurgence follows a difficult start to 2026, where concerns over slowing growth, executive departures, and a lawsuit from Elon Musk weighed on share demand. Despite the recovery, Anthropic remains the preferred investment in the AI sector, continuing to attract the strongest interest from buyers.

Market Sentiment and Valuations

Private market investors buy shares from employees and early backers because neither OpenAI nor Anthropic is publicly traded. These transactions occur in fragmented secondary markets where valuations are estimates rather than continuously traded market prices. According to Caplight data cited in a report, Anthropic is currently valued at about $1.2 trillion in secondary markets, while OpenAI is valued at approximately $933 billion, marking a rise of about 20% over the past three months.

Glen Anderson, CEO of Rainmaker Securities, noted that buyer demand still favors Anthropic by roughly five buyers for every two seeking OpenAI shares. However, he observed that OpenAI is now being "bid" on far more frequently than it was a month ago. Dave McClure, founder and managing partner at Practical Venture Capital, confirmed the "resurgence" in investor interest for OpenAI.

Product Launches Drive Confidence

Several secondary market participants identified OpenAI's recent product launches as the key catalyst behind the renewed interest. The company previewed GPT-5.6 Sol in June before expanding availability this month alongside GPT-5.6 Terra and the lower-cost GPT-5.6 Luna. While independent benchmarks rank GPT-5.6 Sol among the industry's top models, Anthropic's latest Mythos and Fable models still lead some evaluations.

Mark Klein, CEO of Neostellar, pointed to OpenAI's disclosure that Codex and ChatGPT Work have reached 9 million active users. Some investors now view OpenAI as the more attractive value play, though McClure cautioned that competition from Anthropic, Chinese AI startups, and open-source models could continue to pressure the company's growth and margins.

Business Users Fuel Adoption

New data reported by OpenAI this month indicates a significant shift in the user base for its Codex coding agent. Non-developers now account for roughly 20% of Codex's 5 million weekly users and are adopting the tool three times faster than engineers. This trend validates the premise that business users, not just engineers, represent a substantial market for coding agents.

Adisorn Ratanakovit, founder and CEO of enterprise AI automation platform Complete.dev, noted that this data confirms a pattern he observed three years ago. "My take is that OpenAI and Anthropic built these tools with engineers in mind, and business users started showing up on their own," said Adisorn. "That's the shift I think this data confirms, and it's useful to see it validated at this scale."

Scott Galloway predicted Monday that OpenAI will replace Sam Altman as CEO within six months, telling Prof G Markets listeners the AI giant needs an operator to steady the company. The New York University professor stated that Sam Altman is on the way out, characterizing him as an innovator rather than an operator suited for the current challenges facing the firm.

Galloway’s specific forecast involves OpenAI acquiring enterprise AI firm Sierra and installing co-founder Bret Taylor as chief executive. Taylor currently chairs OpenAI’s board, meaning the prediction involves him purchasing his own startup and swapping roles with the current CEO. Galloway cited Taylor’s previous tenure as Twitter’s board chair in 2022, where he oversaw the completion of Elon Musk’s $44 billion buyout, as evidence of the steady hand required.

The prediction arrives during a difficult stretch for OpenAI. Apple Inc. reportedly sued OpenAI over alleged trade secret theft, while Applications President Fiji Simo announced her departure and the company shut down its Atlas browser. Competition is also intensifying, with ChatGPT’s market share reportedly falling from 76% to 50% in a year due to pressure from rivals like Gemini, Claude, and Chinese models such as Moonshot’s Kimi K3.

Pricing pressure is emerging as a significant threat. DeepSeek is reportedly pricing output at 87 cents per million tokens compared to $45 for OpenAI’s GPT-5.6. Bret Taylor pushed back on the cheap-token narrative on CNBC, arguing that frontier models are more token efficient and may complete tasks at a lower real cost despite higher headline prices.

Financial modeling by Prof G Markets co-host Ed Elson highlights potential risks for investors. In a worst-case scenario where OpenAI’s hardware business fails, ad revenue misses by 90%, and model prices fall 80%, the company could be losing $165 billion a year by 2030. This poses a concern for Microsoft Corp., OpenAI’s largest backer, and Oracle Corp., whose debt S&P recently downgraded, citing OpenAI as a key credit risk.

Prediction markets reflect this skepticism. Traders on Polymarket see an 86% chance that Anthropic, OpenAI’s main rival, will IPO before OpenAI. A separate market assigns OpenAI only a 10% chance of having the best AI model by the end of the year.