TikTok CEO Promises AI Science Breakthroughs Within 2–3 Years

TikTok CEO Shou Zi Chew believes that within the next two to three years, artificial intelligence will make a number of major discoveries in fundamental science possible. In his words, physics, biotechnology and materials science are highly likely to get a sharp boost. He gave a timeline and named the fields. He named no projects, no models and no verifiable results. This is a forecast, not an announcement.

What Chew said and what he did not

The head of the largest short-video platform made an assessment that reaches well beyond his own business. He was not talking about recommendation feeds or video generation, but about science itself. Everything known from the report fits into a short list:

• horizon: two to three years;
• fields: physics, biotechnology, materials science;
• level of confidence: "highly likely";
• criterion for a "major discovery": not defined;
• specific systems, labs and partners: not named.

Without a criterion, the forecast is hard to disprove. If any paper in which a model helped find a candidate material or protein counts as a discovery, one has already happened. If you are waiting for a new physical theory, a three-year deadline looks far bolder.

Why these three fields

Chew's choice is not random. AI already has notable results in these disciplines. Google DeepMind's AlphaFold learned to predict the three-dimensional structure of proteins from their amino acid sequence, and in 2024 that work won the Nobel Prize in Chemistry (Demis Hassabis and John Jumper shared it with David Baker). Another DeepMind system, GNoME, proposed about 2.2 million new crystal structures in 2023, of which roughly 380 thousand were rated stable. Then comes the bottleneck: candidates have to be synthesized and tested in a lab, which is expensive and slow.

This is where AI helps most. It narrows the search space: instead of millions of options, experimenters are left with hundreds. Physics is harder. It has plenty of data from simulations and accelerators, but a breakthrough there usually means a new idea, not a better approximation.

Who is speaking and why

Chew is not a researcher. He is a finance and management professional who was CFO of Xiaomi before TikTok. His statement is closer to the position of an industry leader describing where the technology is heading than to a scientist's expert assessment. ByteDance, TikTok's parent company, develops its own models, including the Doubao chatbot, and benefits when AI is seen as a tool for major discoveries, not just a generator of content and recommendations.

That does not make the forecast wrong. But it is not a reason to revise your plans either: timelines in such statements are traditionally optimistic, and verifying scientific results takes years.

How to approach this

If you work in the industry, look at specifics, not horizons. Useful questions include: Does the model have published results confirmed by independent labs? How many candidates reached synthesis? How many of them worked? For biotech and materials companies, the sensible bet right now is to integrate AI into the candidate-screening pipeline, not to wait for a ready-made "discovery machine." For investors, there is one sign of maturity: the move from predictions to confirmed experiments. If such cases become noticeably more common in a year or two, Chew's forecast will start to come true. If not, what remains is a handsome but, so far, empty claim.