Nvidia CEO Jensen Huang could not get TSMC to return his calls, according to a report from Business Insider Africa. The breakthrough came only when TSMC founder Morris Chang picked up the phone himself. That single detail, if accurate, explains how a young fabless chip company secured the foundry capacity that later powered its rise.

The anecdote is thin on verified specifics. The report offers no dates, no wafer volumes, and no contract values. Readers should treat it as corporate lore rather than documented history. But the story fits a well-established pattern in the semiconductor supply chain, and the pattern deserves scrutiny on its own terms.

Why would TSMC ignore Nvidia?

In the 1990s, Nvidia was an unproven startup. TSMC, founded by Chang in 1987, was already sorting customers by volume and credibility. Foundry capacity is not a commodity. A fab slot given to a risky customer is a slot denied to a proven one.

Huang had founded Nvidia in 1993. The company had no track record of high-yielding designs. Its early products struggled. A foundry evaluating that customer file would see risk, not revenue.

The reported detail is that Chang intervened directly. Founders bypassing normal channel filters is rare. It is also the kind of detail that survives in industry storytelling precisely because it is unusual.

What is not said?

The report does not say when the call happened. It does not say who brokered it. It does not quantify the initial wafer allocation or the process node involved. It does not state whether the first deal covered graphics chips specifically.

Every one of those gaps matters for the commercial picture. An early tie-up on a mature node means one thing. Priority access to an advanced node means something else entirely. Without a date, the reader cannot place the anecdote against TSMC's own technology milestones.

The report also does not address what TSMC received in return. The conventional narrative treats Nvidia as the beneficiary. Over the following decades, Nvidia became one of the foundry's largest and most demanding customers. The relationship cut both ways. The anecdote, as told, leaves that part out.

Does the story change anything today?

Not on its own. Nvidia and TSMC now operate one of the closest partnerships in the industry. Nvidia's flagship accelerators depend on TSMC's most advanced process nodes. TSMC's leading-edge volumes depend heavily on Nvidia's orders.

The anecdote's value is illustrative, not informational. It shows that foundry access has always functioned as a gate. The gatekeeper decides which fabless designers reach the market at scale. Today's AI-chip startups face the same gate, with fewer founders left to open it by hand.

Chang retired from active TSMC leadership years ago. The era of founder-to-founder phone calls has largely closed with him. Access now runs through capacity committees, long-term agreements, and co-development contracts. That is a harder door for a newcomer to knock on.

The report gives no figures, so none can be checked. What remains is a claim about a call, a founder, and a customer nobody else wanted. If the story holds, it marks the moment the industry's most consequential supply relationship began with someone simply answering the phone.