TikTok CEO Shou Chew Expects an AI Scientific Breakthrough Within Two to Three Years

Shou Chew, CEO of TikTok, believes AI will lead to a major scientific breakthrough within the next two to three years. He named physics, biotechnology, chemistry and materials science as possible areas. He made the forecast on Monday, October 12, at the DBS Bank Asia Future Forward conference in Singapore. CNBC reported on the talk, and The Next Web picked up the story.

What we know about the statement

The source gives only an excerpt of the talk, so it is important to separate what was said from what was inferred. Chew did not name a specific discovery, lab or model that would deliver it. Nor did he explain what he counts as a "breakthrough": a new class of drugs, a material with record-setting properties, or a result that changes a theoretical model. For now, this is an executive's estimate, not an announcement of a development.

Key parameters of the statement:

• Who: Shou Chew, CEO of TikTok.
• Where: DBS Bank Asia Future Forward conference, Singapore.
• When: Monday, October 12, 2026.
• Horizon: two to three years.
• Fields: physics, biotech, chemistry, materials science.
• Primary source: CNBC report.

Why these fields

The list is not random. These are disciplines where AI already has notable results. The best-known example is AlphaFold from Google DeepMind. Its protein structure prediction earned Demis Hassabis and John Jumper the 2024 Nobel Prize in Chemistry, which they shared with David Baker. In materials science, DeepMind reported in 2023 that it had predicted 2.2 million crystal structures, of which about 380,000 were labeled stable. Some specialists disputed the novelty and practical value of those results. The dispute points to the central difficulty: a model produces candidates quickly, but lab verification takes months or years.

That makes Chew's phrase "real scientific breakthrough" more cautious than it may seem. In these fields, a breakthrough is measured not by a benchmark but by a confirmed experiment: a synthesized substance, a clinical signal, a reproduced result.

Where the forecast is weak

First, two to three years is a short time for science. The path from a discovered molecule to an approved drug usually takes many years, and AI mostly speeds up the early stages. A breakthrough that could be recognized as such by 2028–2029 is more likely to come from computational chemistry, physics or materials, where verification is faster.

Second, Chew runs a short-video platform, not a research lab. The available text does not make clear whether he ties the forecast to work at parent company ByteDance, which has its own models, or is speaking as an industry observer. These are the words of an observer with a large audience and access to Asia's business forums, not inside knowledge of an upcoming result.

Third, "breakthrough" has no generally accepted definition. If the bar is not set in advance, any high-profile achievement can be counted as confirming the forecast.

How to use this forecast

For readers who follow the industry, the statement is useful as a sentiment indicator: executives at consumer tech companies increasingly talk about AI as a tool for science, not only for recommendations and chatbots. It is more practical to watch verifiable signals than forecasts:

• publications in peer-reviewed journals with experimental confirmation of AI predictions;
• partnerships between AI labs and pharmaceutical companies and materials manufacturers;
• clinical trial filings for molecules found with the help of models;
• independent reproductions of results.

If such confirmations appear within two to three years in even one of the four named fields, Chew's forecast can be considered fulfilled. For now, it is a bet on how fast AI will move from "proposed candidates" to "proved a result."