Stanford's 2026 AI Report Card: A+ in Math, F in Telling Time

Stanford's AI Index came out this week, all 423 pages of it. The big takeaway is that AI is getting frighteningly good at some things and embarrassingly bad at others. A top model just won a gold medal at the International Mathematical Olympiad. That same model can only read an analog clock correctly about half the time.

The report (link below) calls this the *jagged frontier*. AI's capabilities don't grow evenly. It can write code that would impress a senior engineer but still can't reliably fold laundry or tell you what time it is without digital numbers.

The Numbers That Jumped Out

Global corporate AI investment hit $581.7 billion in 2025, more than double the year before. The US spent $285.9 billion of that. China spent $12.4 billion in private investment, though the government also pours money through state funds that don't show up in those numbers.

The US-China performance gap has almost disappeared. As of March 2026, Anthropic's best model leads China's best by just 2.7%. A year ago, that gap was much wider.

The Environmental Cost

Training one model, Grok 4, produced an estimated 72,816 tons of CO2 equivalent. That's like driving 17,000 cars for a full year. The water used just for GPT-4o's daily operations could meet the drinking needs of 12 million people. AI data centers in the US now draw about 29.6 gigawatts of power, roughly what it takes to run the entire state of New York at peak demand.

The Job Market Is Shifting

Employment for software developers aged 22 to 25 has dropped nearly 20% since 2024. Older developers? Their numbers are still growing. The entry-level jobs that used to be the first rung on the ladder are disappearing. Companies are finding that a senior engineer with AI tools can do the work of an entire junior team.

A third of organizations expect AI to cause workforce reductions in the next year. But most companies are still in the experimental phase. They're spending billions on AI tools that aren't actually integrated into daily workflows yet.

The Transparency Problem

The most powerful AI models are now the least transparent. The Foundation Model Transparency Index scores dropped from 58 last year to 40 this year. The companies building the best models are sharing less information about their training data, compute, and methods.

You can't measure risk if you don't know what's inside.

What Normal People Think

There's a massive gap between experts and the public. 73% of AI experts think the technology will have a positive impact on jobs. Only 23% of the general public agrees. Experts focus on what AI can do. Regular people focus on what it might do to them.

Globally, 53% of people have used generative AI, adoption faster than PCs or the internet. Singapore leads at 61%, followed by the UAE at 52%, and China at 51%. European countries trail behind: Germany at 32%, France at 30%, and the UK at 29%. The US ranks 24th globally at just 28%.

In Africa, Nigeria and South Africa are standout cases. Nigerians are among the most excited about AI globally, with 61% expressing optimism. In South Africa, adoption is strong, ranking among the fastest in learning AI engineering skills. Across the continent, workplace AI use is high, with Nigeria joining India, China, the UAE, Egypt, and Saudi Arabia where over 80% of workers use AI regularly on the job.

Southeast Asia is the most optimistic region. Over 80% of respondents in Malaysia, Thailand, Indonesia, and Singapore believe AI will profoundly change their lives within three to five years. India is a different story. AI anxiety there jumped 14 percentage points in one year, the largest increase globally, while excitement barely moved.

South America shows mixed signals. Brazil leads adoption at 48%. Colombia saw the largest negative sentiment shift of any surveyed country. But overall, the continent's adoption patterns mirror global trends, with higher optimism in younger populations.

The Bottom Line

The report doesn't give easy answers. AI is accelerating in some directions and hitting walls in others. The US leads in investment and top models but is losing its grip on talent. The jobs being lost are entry-level, not executive. The people most optimistic about AI are the ones who don't have to worry about being replaced.

The gap between what AI can do in a lab and what it can do in the real world is still huge. Winning at math Olympiads is impressive. Reading a clock is useful. We haven't quite figured out how to bridge those two things yet.