OpenAI Wants a $1 Trillion Valuation. But College Students Are Testing At The Level Of 10-Year-Olds.

Sam Altman wants Wall Street to value OpenAI at $1 trillion. Meanwhile, 14% of US college students are reading at or below the level of a 10-year-old, according to the OECD. Those two facts, sitting side by side, may explain why the most anticipated technology IPO in a generation just got pushed into 2027.

OpenAI Files Its S-1

OpenAI confidentially filed its S-1 around May 22, 2026, targeting a Q3 or Q4 2026 debut. Then the market delivered a warning. SpaceX went public in June, ran up past $225, and fell back to roughly $153 within two weeks — a drop of about 32% from its peak that OpenAI’s advisors reportedly cited as a caution. OpenAI is now reported to be leaning toward delaying its listing to 2027, and while it hasn’t confirmed that timeline, Altman has reportedly rejected any valuation below the $1 trillion target as a “nonstarter.”

The numbers underneath that target are staggering. OpenAI recorded a $38.5 billion net loss in 2025 on roughly $13 billion in revenue — though much of that figure stems from non-cash accounting charges, with the operating loss closer to $21 billion. The company doesn’t expect to turn cash-flow positive until 2029 or 2030. Anthropic, the closest peer, filed confidentially and recently raised capital at a $965 billion valuation. It’s still aiming for a late-2026 debut. The bet underwriting both valuations is simple: AI will make humans dramatically more productive. The evidence is not cooperating.

College Students Are Increasingly Testing at 10-Year-Old Levels

Start with the US. In the OECD’s 2023 PIAAC Survey of Adult Skills, which tested 160,000 adults across 31 countries, 14% of US college students scored at or below Level 1 in literacy, nearly double the OECD average. Globally, 8% of college students now read at or below the level of a 10-year-old. The generational reversal is the sharpest signal: older US adults score above the international literacy average, while younger US adults now score below it. Researchers point to pandemic-era learning gaps, looser admissions standards, and the explosion of large language models creating a new floor for academic failure.

The public has noticed. NBC News found in March 2026 that just 26% of registered voters hold a positive view of AI, while 57% say the risks outweigh the benefits. Pew Research reported in September 2025 that 50% of US adults say AI makes them more concerned than excited, up from 37% in 2021. And a Quinnipiac poll in April 2026 found 80% of Americans are concerned or very concerned about AI.

Will AI-Fueled Productivity Gains Arrive?

Then came the concession. At Meta‘s (NASDAQ:META | META Price Prediction) July 2 town hall, Mark Zuckerberg told employees that AI “hasn’t really accelerated in the way that we expected”. It was the first major CEO admission that the productivity payoff is lagging the spend. And the spend is enormous: companies have committed up to $700 billion in AI capex in 2026. Consumer sentiment, meanwhile, sits at a reading of 44.8 in May 2026, deep in recessionary territory. If AI were compounding household productivity, that number would not look like this.

Now, its worth noting these fears could prove temporary. Fear of electricity, the internet, and smartphones all looked overwrought in retrospect. The cognitive decline predates ChatGPT and has clear pandemic roots. And OpenAI’s revenue grew from $6 billion in 2024 to $20 billion annualized in early 2026, an explosive trajectory that does not require public affection to continue.

So here is the question underwriters, pension funds, and retail investors have to answer before Altman rings the bell in 2027: if the smartest technology ever built is meeting the least literate cohort ever measured, who is actually getting more productive?

Contact [email protected] for any questions or corrections.