Broadcom is shifting its strategic centre of gravity from acquisitions to organic expansion in artificial intelligence, marking a notable change for a company long defined by Chief Executive Officer Hock Tan’s dealmaking discipline and post-merger cost control.Tan has indicated that Broadcom’s strongest growth prospects now lie inside its own AI semiconductor pipeline rather than in another large takeover. The message lands at a crucial point for the Palo Alto-based group, whose rise into one of the world’s most valuable chip and infrastructure software companies was built through a succession of major transactions, including CA Technologies, Symantec’s enterprise security business and VMware.
The pivot reflects the scale of AI demand now flowing through Broadcom’s custom accelerator and networking businesses. The company reported second-quarter fiscal 2026 revenue of $22.19 billion, up 48 per cent from a year earlier, with AI semiconductor revenue reaching $10.8 billion, a 143 per cent year-on-year increase. Broadcom expects AI semiconductor revenue to rise to about $16 billion in the third quarter, implying growth of more than 200 per cent from the same period last year.
That growth has changed the logic of capital allocation. Broadcom’s acquisition model relied on buying large, mature technology assets, cutting overlapping costs, improving margins and converting products into recurring platforms. AI offers a different route: large cloud customers are committing heavily to custom silicon, high-speed networking and tightly integrated data-centre infrastructure, giving Broadcom a chance to grow by deepening existing customer relationships rather than buying new ones.
Broadcom’s custom AI chips, often designed for hyperscale customers, are aimed at companies seeking alternatives or complements to general-purpose graphics processors. Its networking components also sit at the centre of AI clusters, where moving data rapidly between processors is critical to model training and inference. That gives the company exposure not only to AI compute but also to the infrastructure layer required to scale it.
The company has told investors it has visibility toward more than $100 billion in AI chip revenue in fiscal 2027, supported by demand from a limited group of large customers. That concentration remains both an advantage and a risk. Deep partnerships with leading cloud and AI platforms can produce large, multi-year revenue streams, but any shift in customer sourcing, chip design strategy or spending pace can quickly affect sentiment.
Market reaction underscored that tension. Broadcom shares fell sharply after its latest earnings update, despite record revenue, strong free cash flow and guidance for third-quarter revenue of about $29.4 billion. Investors had expected a more aggressive upward revision to AI targets, and the decision to maintain rather than raise longer-term guidance disappointed a market already pricing in rapid AI growth.
The sell-off also reflected concerns about margins. AI custom silicon can generate huge revenue, but investors are watching whether a larger mix of customer-specific chips and systems could dilute Broadcom’s traditionally high profitability. The company’s adjusted EBITDA reached $15.24 billion in the second quarter, equal to 69 per cent of revenue, showing strong operating leverage, but margin expectations remain under scrutiny as AI becomes a bigger part of the business.
Broadcom’s software division remains a major stabilising force. Infrastructure software revenue was $7.18 billion in the second quarter, up 9 per cent from a year earlier, aided by VMware’s integration and Broadcom’s push towards subscription-led enterprise offerings. The VMware acquisition, completed in 2023, gave Broadcom a deeper role in private cloud infrastructure, an area the company argues will benefit as enterprises build AI workloads outside public cloud environments.
Tan’s reduced appetite for acquisitions does not mean Broadcom is abandoning its dealmaking DNA. The company still has a history of using large transactions to reshape its revenue base. But the AI cycle has created an internal growth opportunity of a size that could rival what past acquisitions delivered, without the regulatory delays, integration risk and customer disruption that often accompany major takeovers.
Topics
Technology