
A 0% chance of Jensen Huang returning to Nvidia was Rohan Goswami’s assessment when host Ed Elson put the question to him in the prediction segment of Prof G Markets. Two sets of verifiable facts informed his conclusion. At a recent industry gathering attended by around 400 AI and technology company CEOs, Goswami had found that the prospect of Huang returning was a leading subject of discussion. His previous reporting, meanwhile, indicated that hours of board deliberations had preceded Nvidia’s decision to change its leadership. Among the directors were people whose working relationships with Huang extended back more than 30 years to the company’s founding. For Goswami, the extent of that process suggested the board was unlikely to undo its decision simply in response to pressure from outside.

The circumstances of Huang’s departure left more scope for a comeback in the view of fellow guest Deirdre Bosa. She pointed out that he had chosen to resign, rather than being dismissed directly by the board. Some observers also drew encouragement from Sam Altman’s reinstatement as OpenAI’s CEO only days after his removal. Yet Bosa accepted that expressions of public support were not necessarily a sign that directors would reconsider. Those expressions included a petition for Huang’s reinstatement signed by Elon Musk, as well as Donald Trump’s comment that a return “might not be a bad idea”.
As for a possible new venture with $50 billion, that was a forecast Goswami offered personally during the programme, not a proposal announced by Huang. The sum nevertheless had a specific reference point: Nvidia’s $50 billion AI financing platform, unveiled personally by Huang in August 2026.
Nvidia Advances CEO Search and Rules Out Huang’s Return
The board had no intention of reinstating Huang, Nvidia Chairman Harvey Jones told a conference call explicitly. A search for a long-term CEO was taking place, he explained, under the direction of executive search firm Heidrick & Struggles. Far from leaving matters at the vague assurance that “all options remain open.”, Nvidia had committed itself to a position and acted on it. With an external search firm appointed, the selection procedure was already in progress.
Bosa’s reference to Altman therefore involves a materially different sequence of events. His return to OpenAI within days of being dismissed followed the board’s decision to reverse its original action. At Nvidia, a voluntary resignation by Huang was followed instead by a formal recruitment process and a public rejection of the prospect of his return. There is a wider purpose when a chairman tells analysts on an earnings call that there is “no intention of bringing Huang back,” beyond addressing speculation. Customers, employees and executive recruiters are also being assured that the search is for a new CEO, not a holding exercise until the founder returns.
By taking those formal steps, the board gave further weight to Goswami’s argument: having advanced so far, it was unlikely to be persuaded to reverse its decision by outside pressure. The search for an executive is already in progress, regardless of Musk’s petition or Trump’s words of support.
Huang’s Next Venture Speculation Centres on Financing
The future of Nvidia’s $50 billion AI financing platform after its founder’s exit, rather than his estimated $145 billion net worth, is the more important consideration behind Goswami’s use of $50 billion when discussing Huang’s next move.
Designed to assemble capital from multiple sources, the platform that Huang personally unveiled in August 2026 sought to provide $50 billion of financing for AI computing deals. Its chief strength lay in the credibility and network Huang brought as chief dealmaker, rather than simply the capital available. Confidence in his judgement was a major factor in participants’ willingness to fund data centres and chip assets through it.
The platform’s uncertain position without Huang has already been flagged by market institutions. In Reuters’ reporting, the loss of its chief dealmaker could complicate Nvidia’s efforts to fund some of its riskier infrastructure investments. An estimate from UBS analyst Tim Arcuri put a figure on the possible consequences: $50 billion or more of market funding could be lost if Nvidia reduced its financing support.
Goswami’s projected $50 billion thus corresponds in scale to the financing shortfall that could emerge at Nvidia’s platform. A new Huang business starting with $50 billion might tap a capital network moving over from the existing platform, rather than have to source all its funding anew. For AI infrastructure projects previously funded through Nvidia’s platform, and for investors who joined its ecosystem out of confidence in Huang’s judgement, that would present a decision. They could accompany him to a new entity or remain at Nvidia within a financing arrangement in which he no longer participated.
$50 Billion Would Put a New Venture Above Typical Funding Levels
The funding currently available to AI chip start-ups provides the benchmark for judging what $50 billion of initial backing would mean for a new company established by Huang.
November saw British AI inference chip company Fractile raise $720 million in a Series C round that valued it at $7 billion. A valuation of $2.9 billion was reached by South Korea’s FuriosaAI after its latest round. Both businesses are among those attracting closer attention in the AI chip start-up market, where rounds span hundreds of millions to billions of dollars. Initial capital of $50 billion would exceed those sums many times over.
Rather than describing an ordinary AI chip start-up, then, Goswami’s forecast implies a business with sufficient scale to alter competition immediately. Funding of $50 billion could finance the full chain of activities: designing chips, developing software, securing foundry capacity and establishing a customer ecosystem. Repeated fundraising in the early stages, with the associated dilution of control, would not be necessary to support that investment.
Financial resources do not, on their own, establish a barrier to competition. Different technological routes are also in contention, as illustrated by Fractile’s roadmap: stacking DRAM and logic chips in 3D to achieve higher bandwidth and speed than conventional HBM. Spending more than competitors on the same route is not sufficient for success. A business founded by Huang would have to compete in a market where Nvidia’s share exceeds 80%, but where investors and customers are actively looking for alternatives. A broader selection of AI chips is something Microsoft has publicly said it wants, while funding continues to flow to newcomers such as Fractile and FuriosaAI.
In that setting, particular importance attaches to Goswami’s mention of “friends who have supported him all along”. A new entity would need Huang to reassemble the relationships accumulated over three decades. That means investors prepared to fund him, customers ready to trial new chips and supply chain partners prepared to commit production capacity to a new entrant.