Broadcom’s share price climbed close to five percent on Thursday after chief executive Hock Tan outlined an ambitious outlook for artificial intelligence chip revenue that exceeded even optimistic expectations from Wall Street analysts.
The technology executive told analysts during an earnings discussion that he expects the company’s artificial intelligence semiconductor revenue to be “significantly in excess of $100 billion” by 2027 as demand for custom silicon accelerates.
That projection immediately drew attention from investors and analysts alike, with several firms revising forecasts upward after the company indicated it was approaching roughly ten gigawatts of AI compute capacity across six major customers.
Analysts at JPMorgan subsequently suggested Broadcom could generate between $12 billion and $15 billion in revenue per gigawatt by 2027, leading them to lift their long-term AI revenue estimates to at least $120 billion.
The forecast reinforced the view that hyperscale technology companies are rapidly expanding infrastructure built around custom artificial intelligence processors and networking hardware to support large-scale machine learning workloads.
Industry observers noted that the company’s role in designing and enabling custom chips for hyperscalers positions it as a critical supplier within the expanding artificial intelligence infrastructure ecosystem.
Strong Earnings Highlight AI Momentum
The bullish outlook came alongside quarterly results that surpassed analyst expectations, with revenue linked to artificial intelligence products more than doubling due to strong demand for accelerators and networking components.
According to analysts at Goldman Sachs, Broadcom’s leadership across AI networking systems and custom silicon design allows hyperscale customers to achieve significantly lower inference costs when running large language models and similar workloads.
They wrote that the company “enables the lowest inference cost for its hyperscaler customers, and we see it delivering ongoing cost reductions on pace with market leader Nvidia.”
Despite persistent concerns across the semiconductor industry regarding shortages of high-bandwidth memory required for advanced artificial intelligence chips, Tan indicated Broadcom had already secured sufficient memory and wafer supply commitments through 2028.
The assurance appeared to ease investor fears that supply constraints might limit the company’s ability to scale production during one of the fastest expansions in computing infrastructure history.
Tan also addressed questions surrounding the profitability of AI-focused hardware deployments, explaining that efficiency improvements have allowed the company to maintain margins consistent with its broader semiconductor business model.
“We have gotten our yields, we’ve gotten our cost to the point where the model we have in AI will be fairly consistent with the models we have in the rest of the semiconductor business,” Tan told analysts.
Competition And Market Dynamics
Competition within the artificial intelligence semiconductor market has intensified as major cloud computing companies increasingly explore designing their own custom processors rather than relying entirely on third-party chipmakers.
However, Tan argued that the complexity of competing with established leaders such as Nvidia ultimately strengthens Broadcom’s strategic position within the rapidly evolving AI infrastructure market.
Large language model developers, he suggested, require extremely high-performance chips in order to remain competitive against rival AI platforms and rapidly improving hardware architectures.
“Large language model makers cannot afford to have a chip that is just good enough,” he said.
“You need the best chips around because you’re competing against other LLM players, and most of all, you’re also competing against Nvidia, who is by no means letting their guard down.”
Ripple Effects Across The AI Supply Chain
Broadcom’s positive outlook also lifted shares of several companies connected to artificial intelligence infrastructure and data center networking hardware.
Chip connectivity specialist Credo saw its shares rise around ten percent, while Amphenol climbed roughly four percent as investors bet that copper-based connectivity solutions will remain central to large-scale AI server deployments.
By contrast, companies focused on newer optical interconnect technologies experienced declines, with Lumentum and Coherent each falling more than four percent during the same trading session.
The contrasting market reaction highlighted how shifting infrastructure preferences within hyperscale data centers can rapidly influence valuations across the broader semiconductor and networking supply chain.
