The Numbers That Should Terrify Investors

In December 2025, Deutsche Bank analyst Jim Reid released a note that should have sent shockwaves through Silicon Valley. According to internal OpenAI documents shared with investors, the company expects to generate $345 billion in revenue between 2024 and 2029. The problem? Expenditures during the same period are projected to reach $488 billion, primarily to pay for computing power costs.

The math is brutally simple: $143 billion in cumulative negative free cash flow before OpenAI sees its first profitable quarter. And that staggering figure does not include the company's recently announced $1.4 trillion commitment to data center investment.

Internal OpenAI documents predict the company will lose $14 billion in 2026 alone, roughly three times worse than early estimates for 2025. The company is reportedly burning $12 billion per quarter, with some reports indicating $15 million per day is consumed by Sora alone, OpenAI's video generation tool that has yet to prove meaningful commercial viability.

The "Code Red" Memo: Panic at the Top

On December 1, 2025, Sam Altman sent an internal memo to all OpenAI employees declaring a "Code Red" status. The irony was not lost on industry observers: three years earlier, Google had issued its own "Code Red" when ChatGPT threatened to upend the search business. Now the tables had turned.

"We are at a critical time for ChatGPT," Altman wrote in the memo. The declaration meant all other projects would take a backseat to the flagship product. Advertising plans were delayed. Other initiatives were shelved. Every available resource was marshaled toward a single goal: stopping the bleeding.

The memo came just two weeks after Google launched Gemini 3 on November 18, 2025, what Google described as the quickest model integration into Search in company history. The new model immediately topped LMArena's leaderboard across most benchmarks, delivering superior performance in reasoning tasks, coding accuracy, and multimodal comprehension.

Market Share Collapse: From Dominance to Desperation

The numbers tell a story of precipitous decline. According to Similarweb data released in January 2026, ChatGPT's market share has cratered from 87.2% one year ago to just 64.5% as of January 2, 2026. That 22.7 percentage point collapse represents the steepest decline for any dominant technology platform in recent memory.

The raw visitor numbers paint an even bleaker picture. ChatGPT's seven-day average visitors fell from nearly 203 million to 158 million between early December 2025 and January 2026. Meanwhile, Gemini's referral traffic nearly quintupled year-over-year, while ChatGPT's grew by just over half.

User engagement metrics show deeper interactions on Gemini, with average visit duration of 7 minutes 20 seconds and 4.3 pages per visit, compared to ChatGPT's 6 minutes 32 seconds and 3.8 pages. When users are spending more time with your competitor and exploring more of their product, the writing is on the wall.

The Distribution Problem OpenAI Cannot Solve

Google possesses a competitive advantage that OpenAI simply cannot replicate: the Android operating system. Similarweb data reveals that twice as many US Android users engage with Gemini through the operating system directly compared to the standalone app. Gemini does not need to convince users to download anything. It is already there, integrated into the phones of billions of users worldwide.

This built-in distribution channel means Google can push Gemini updates to users instantly, without the friction of app store downloads or subscription barriers. Every Android phone becomes a potential Gemini user without any additional effort. OpenAI, by contrast, must fight for every download, every signup, every subscription.

The Subscription Crisis: Growth Has Stalled

The market share decline compounds an existential challenge: converting free users to paying subscribers. Despite 800 million weekly users, only about 5 percent pay for ChatGPT Plus or higher tiers, roughly 40 million subscribers. And those paid subscriptions have plateaued across major European markets since May 2025, with no recovery in sight.

OpenAI needs to generate $200 billion in annual revenue by 2030 to justify their projections and current $500 billion valuation. That requires 15x growth in five years while costs continue to explode. The math simply does not work if subscription growth has stalled before ChatGPT even reached profitability.

The Talent Exodus and Internal Turmoil

Noble's analysis highlights something the press releases never mention: the talent exodus. Key researchers and engineers have departed OpenAI at an accelerating rate, many citing concerns about the company's direction, its pivot toward commercialization over safety, and the increasingly autocratic leadership structure.

The Elon Musk lawsuit continues to hang over the company, challenging the very foundation of OpenAI's nonprofit-to-for-profit conversion. Whether or not Musk prevails legally, the discovery process has already revealed embarrassing internal communications and raised legitimate questions about the company's governance.

Meanwhile, the GPT-5 launch was widely considered a disappointment. Developer forums are filled with complaints about GPT-5.2's "extremely high hallucination rates during certain periods." Users report wasting hundreds of dollars in API tokens attempting to correct recurring errors. The inconsistency makes the problem particularly dangerous: the model sometimes works correctly, leading users to trust outputs that later prove fabricated.

The Energy Problem Nobody Talks About

Noble raised another issue that receives insufficient attention: the vast energy requirements to keep AI companies operational. "It will cost five times the energy and money to make models two times better," he noted. "The low-hanging fruit is gone. Every incremental improvement now requires exponentially more computers, more data centers, more power."

HSBC's analysis is even more dire, predicting that by 2030, OpenAI's cash burn could exceed $210 billion. This is not a company approaching profitability. This is a company accelerating toward a cliff while the investors in the backseat cheer the speedometer.

When Legends Sound the Alarm

George Noble's career began in the early 1980s at Fidelity, where he worked alongside legendary investor Peter Lynch and ran the company's first international fund, the Fidelity Overseas Fund, which was the number one fund in the United States for several years. He subsequently launched two billion-dollar hedge funds. This is not a commentator seeking attention. This is a market veteran with a track record of identifying bubbles before they burst.

His advice to investors is unambiguous: "If you're exposed to the Magnificent 7 through AI infrastructure bets, consider trimming. The smart money is rotating into sectors where valuations actually reflect fundamentals."

Perhaps most tellingly, Michael Burry, the contrarian investor immortalized in "The Big Short" for predicting the 2008 financial crisis, responded to Noble's critique by issuing his own stark warning that the artificial intelligence frenzy represents a bubble of unprecedented scale, one destined to burst.

The Bottom Line

OpenAI is not a company in temporary difficulty. The evidence suggests a fundamental business model problem that no amount of investor enthusiasm can solve. When your market share drops 22 points in a year, when your traffic growth is one-sixth of your competitor, when you are burning $12 billion per quarter with no clear path to profitability, when industry legends are warning of imminent collapse, the question is not whether a reckoning is coming.

The question is whether you will be holding the bag when it arrives.

OpenAI declared "Code Red" in December. Perhaps it should have declared bankruptcy instead.