Broadcom’s $100 Billion AI Bet: CEO Hock Tan Signals Bold Push to Challenge Nvidia’s Dominance

In the high-stakes race to power the artificial intelligence revolution, Broadcom Inc. is making one of the boldest predictions yet.

Chief Executive Hock Tan says the semiconductor giant can generate more than $100 billion annually from AI chip sales by 2027, a dramatic leap that would position the company as one of the most formidable challengers to industry leader Nvidia.

Speaking during an earnings conference call with analysts, Tan said Broadcom already has a clear path to the milestone.

The projection underscores how aggressively Broadcom is pivoting toward artificial intelligence as the technology reshapes the global semiconductor industry.

Just two years ago, such a forecast might have sounded unrealistic. Yet Broadcom’s AI business has already surged at a remarkable pace.

To reach a $100 billion annual run rate, the company will need to multiply its current quarterly pace several times over — but executives say the demand pipeline is already forming.

Investors responded quickly. Shares of Broadcom Inc. (AVGO) jumped around 6.8% in pre-market trading following the announcement.

While Nvidia dominates the market for general-purpose AI accelerators, Broadcom has carved out a different strategy: custom-built AI semiconductors designed for specific tech giants.

These chips — along with high-speed networking hardware — help massive data centers train and run advanced AI models.

Broadcom’s client roster includes some of the biggest names in the AI race:

Tan revealed that OpenAI could begin shipping Broadcom-designed chips in volume next year, eventually delivering more than one gigawatt of computing capacity.

Meanwhile, demand tied to Google’s TPU (Tensor Processing Unit) chips — which Broadcom helps develop — is expected to accelerate further by 2027.

Broadcom’s latest earnings highlighted how quickly AI is reshaping its business.

For the fiscal first quarter ending February 1:

The AI segment alone grew 106% year-over-year, fueled by demand for custom accelerators and AI networking infrastructure.

That growth is forcing investors to rethink Broadcom’s role in the semiconductor ecosystem.

Despite the strong results, skepticism remains.

Wall Street has grown cautious about the massive spending spree on artificial intelligence infrastructure, particularly after even stellar earnings from Nvidia recently triggered a stock selloff.

The key question hanging over the industry:
Will today’s AI boom last long enough to justify the trillions being poured into data centers and chips?

Broadcom itself had faced doubts earlier this year, with its shares falling more than 8% before the latest rally.

Broadcom’s long-term strategy relies heavily on deep partnerships with leading AI developers.

Tan dismissed recent speculation that Meta might be pulling away from the collaboration.

Alongside its AI ambitions, Broadcom also delivered a solid financial outlook.

The company expects around $22 billion in revenue for the fiscal second quarter, beating Wall Street’s average forecast of $20.5 billion.

Broadcom also approved a $10 billion share buyback program, adding to the $7.8 billion in stock repurchases already completed this year.

Even if Broadcom hits its ambitious targets, Nvidia will still tower over rivals. Analysts expect Nvidia to generate more than $330 billion in AI data center revenue by fiscal 2027.

But Broadcom’s strategy — building custom silicon for the world’s largest AI labs — could reshape the competitive landscape.

As AI infrastructure expands from cloud companies to governments, research labs, and enterprises worldwide, the semiconductor battle may no longer be a one-company show.

And if Hock Tan’s prediction comes true, the AI chip race could soon have two giants instead of one.