Andrew Ng built Coursera into one of the world's largest online learning platforms. Now he is betting $100 million of Coursera's money that the conventional story about AI and jobs is wrong.

The new company is called LearnVector. Coursera announced the investment on Tuesday, putting up $100 million for a one-third stake, which implies a total value of roughly $300 million for a firm that does not yet have a product on the market. Ng, who is also a general partner at AI Fund and sits on Amazon's board, will serve as CEO.

The pitch is a direct challenge to a widely held fear. "Conventional wisdom holds that AI will replace people, and I believe in the opposite," Ng said in an interview. His argument is that AI will not eliminate white-collar work so much as reshape which skills that work requires, and that workers who learn those new skills will be more productive, not redundant.

What LearnVector actually does

The company plans to use AI to build personalized courses for professionals, courses that track a learner's progress and grow harder as their proficiency grows. Ng says the target customers are corporations, governments, and universities, not individual consumers browsing for a new hobby.

First courses are expected to launch early next year. The company is still finalizing pricing, course offerings, and its user interface, which means Coursera is essentially funding a bet on Ng's judgment before the product exists.

That bet carries real weight for Coursera. The $100 million investment represents about 13% of the company's available cash as of late March. For a publicly traded company to stake that much on a startup its own co-founder is running is not a routine portfolio move. Coursera will also partner with LearnVector to sell and develop courses together, so the two companies are tied at the strategy level, not just financially.

The anxiety this is selling into

The timing matters. Amazon, Meta, and Block have all cited AI-driven efficiencies as partial explanations for recent rounds of layoffs. Whether AI is actually the primary cause of those cuts or whether a slower economy shoulders most of the blame is genuinely contested. But the fear is real and spreading among office workers who are watching automation tools absorb tasks they used to own.

Ng's counterargument is that automation does not eliminate jobs wholesale. It eliminates specific tasks within jobs, which then creates demand for new skills to handle what automation cannot. The workers who learn those skills stay relevant. The workers who don't lose leverage. That is not an especially comforting message on its own, but it is a more granular and probably more accurate picture than "AI takes everything."

LearnVector is, at its core, a business built on that distinction. If Ng is right, the addressable market is enormous: every white-collar worker who feels the ground shifting under their career. If he is wrong, or if corporations decide to automate rather than retrain, the company will face an uphill climb selling courses to a workforce that has been cut rather than upgraded.

What is less uncertain is the direction of the broader industry. Every large employer is actively working out how much of its workforce to reskill versus reduce. Governments are funding workforce development programs. Universities are scrambling to make credentials feel relevant again. LearnVector is entering a crowded, anxious, well-funded space, led by someone who helped build the template it is competing against.

The $300 million valuation will look either prescient or inflated depending on whether the product, once it exists, can actually change what workers know how to do and whether employers are willing to pay for that. Those are two very large open questions to carry into an early-2027 product launch.