Wei Zhou and the education that never ended

Wei Zhou had already earned a doctorate when he went to law school. That might have been a tidy ending to one career and the beginning of another. Instead, the two professions began sharing a desk. While studying at Stanford, he joined Affymetrix, a young company trying to bring semiconductor methods to the study of DNA. The scientist had acquired a new vocabulary. The law student had found a problem large enough to keep the scientist occupied.

There is something appealingly inconvenient about this route. Professional biographies usually make advancement look like a staircase: qualification, appointment, promotion, larger office. Zhou’s looks more like a university timetable assembled by someone who refuses to choose a major. Research, intellectual property, international operations, legal practice, entrepreneurship. Each turn adds another kind of work to the same underlying fascination with how to read the genome.

Today he is founder and chief executive of Centrillion Technologies in Palo Alto. The company makes high-density DNA chips. That description is compact enough for a business card; the work behind it involves chemistry, optics, computation, and manufacturing. It also gives his seemingly wandering education a practical destination. To build a tool that crosses disciplines, it helps to have crossed a few yourself.

A doctorate, then a different set of books

At Virginia Tech, Zhou earned his Ph.D. as a John Lee Pratt Fellow. Postdoctoral training at Emory followed. He then studied law at Stanford, where he served as an associate editor of the Stanford Law Review. His education also includes Harvard Business School’s General Management Program. A laboratory, a law review, and a management classroom ask different things of a person. Together, they make an unusually varied preparation for running a technology company.

The interesting part is the change in scale. A researcher can spend years making one question precise. A lawyer must also consider who owns an answer, who can use it, and on what terms. A manager has to arrange people and resources so that an answer becomes something other people can buy. Those are distinct jobs. Zhou’s career puts them close enough together to see the joins.

At Affymetrix, he became vice president of intellectual property and advanced technology, holding that role until 2008. The pairing is revealing. A technology company needs inventions, but it also needs an account of what those inventions are and how they fit together. Intellectual property sits at that awkward meeting point between the thing being built and the commercial world waiting outside the laboratory.

The temporary assignment that got bigger

Affymetrix also gave Zhou a second education in business. He took responsibility for its China operations after initially working on intellectual property. He later recalled being asked to oversee the market on an interim basis. A temporary assignment is one of corporate life’s more optimistic phrases. Sometimes it means a short visit. Sometimes it supplies the next chapter of a career.

He helped build the company’s China business. The move carried him from managing ideas to managing an operation, with the different demands that entails. Knowing what a technology can do is one problem; helping it travel into another market is another. The two can meet in the same executive’s calendar without becoming the same skill.

Before starting Centrillion, Zhou was a partner at Wilson Sonsini Goodrich & Rosati, advising technology businesses and investors. He also served as a senior adviser to Covington & Burling. Legal practice placed him among the people creating companies and the people financing them. Founding a company would put his own decisions on the other side of that conversation.

2009: the adviser becomes the founder

Centrillion began in 2009. Its technical ambition belonged to the same world that had drawn Zhou to Affymetrix: making molecular information easier to examine through better physical tools. In a 2015 research-funding announcement, the company described work on phased genome sequencing using specially fabricated substrates. Filip Crnogorac was the project’s principal investigator. Semiconductor methods were part of the proposed route forward.

That detail matters because the word “data” can make the physical work disappear. Before a computer can analyse a measurement, someone has to make the measurement possible. Surfaces have to behave as intended. Molecules have to reach the right places. A procedure has to survive repetition. The elegant chart at the end depends on a procession of less photogenic practical decisions.

In Centrillion’s process, light selects locations where the next DNA building block can attach. Repeating the cycle builds designed sequences at known positions. The company reports capacity for more than 2.4 billion distinct DNA features on a wafer. The useful idea is parallel work: many positions can be patterned during one exposure. Density becomes a manufacturing question as well as a biological one.

The week he tried to buy his old workplace

In March 2016, Zhou’s connection to Affymetrix took a very public turn. He was president of Origin Technologies, an entity formed by former Affymetrix executives to acquire the company. Origin initially offered $16.10 per share in cash. It subsequently raised the proposal to $17. Thermo Fisher Scientific’s competing transaction offered $14 per share. The former employee was now helping propose the future ownership of his former employer.

The higher number did not settle the matter. Affymetrix’s board considered financing, timing, regulatory, and other execution risks, ultimately recommending against Origin’s proposal. Origin withdrew. Zhou disputed the board’s assessment of the risks and described the outcome as disappointing. Thermo Fisher completed its acquisition on March 31, 2016.

It is a useful interruption to the smooth founder narrative. Experience, conviction, and a higher bid can still end with someone else completing the transaction. The episode brought together several parts of Zhou’s working life: technology, former colleagues, legal negotiation, and cross-border finance. For a short period, those strands appeared in public filings rather than a biography.

There is no need to turn the defeat into a secret triumph. The offer was withdrawn. Centrillion continued. A career can contain both facts without one being recruited to explain away the other. The uncompleted transaction belongs in his story because it shows the scale of a decision he was prepared to pursue, and the limits that another board placed on it.

A book you would rather keep in order

Zhou has used a reader’s image for a technical problem: the difficulty of reconstructing long DNA sequences from fragments. In 2018, he described the original ambition this way:

“instead, we should read it page by page.”Wei Zhou · 2018 interview

In November 2021, Zhou, Maurizio Righini, and Justin Costa published work on “DNA bridges.” The researchers attached long DNA molecules to a surface, then removed portions of that surface so the molecules remained suspended. Their experiments showed that enzymes could access the suspended DNA and incorporate labelled nucleotides. The paper presented a possible foundation for further sequencing work, rather than announcing a finished universal reading machine.

The distinction gives the story some welcome texture. An aspiration can be expressed through a book metaphor. Testing it requires a particular arrangement of surfaces and molecules. Between those two descriptions lies much of the actual work of invention: taking a sentence that sounds obvious and discovering how many conditions must be satisfied before it becomes an experiment.

The small print travels

Another route into use arrived through a January 2023 licensing and technology-transfer agreement with 10x Genomics. The transaction included specified Centrillion chip technology and the acquisition of Centrillion’s Taiwan chip pilot facility. Centrillion retained engineering and bioinformatics capabilities there. The relationship has a visible commercial trace: 10x Genomics’ HD 3′ Gene Expression product page identifies its technology as manufactured under license from Centrillion.

For someone whose career includes intellectual property, that is a particularly apt place for the story to arrive. A licence turns a technical asset into permission for another organisation to build with it. The public-facing product carries another company’s name while a short acknowledgement records part of its technological ancestry. A few words of small print can contain years of work.

Zhou’s activities extend beyond Centrillion. He also cofounded the bioinformatics software company Sentieon, incorporated in July 2014, and served as HYSTA’s chairman in 2014. In 2025, an IEEE conference programme listed him on an industry panel about AI standards, needs, and challenges. The settings differ, but they keep him near the conversation between technical possibility and practical organisation.

His current company places increasing emphasis on biological data for AI. Seen across the whole career, the continuity is the attention paid to the tools that make knowledge possible. The doctorate supplied one preparation, law another, business a third. None appears to have persuaded him that the education was finished. Somewhere between the books, the contracts, and the wafer, there was always another subject to learn.