Broadcom's September earnings produced a striking analyst split. The company reported quarterly revenue of about $29.6 billion and outlined a sharp acceleration in AI chip sales, including a path toward roughly $115 billion of AI revenue in fiscal 2027 and $230 billion in fiscal 2028. That scale would make custom accelerators and networking an even larger part of the investment case.

Yet a stronger long-term AI number did not create a single Wall Street response. MarketScreener and TipRanks summaries of 3 September actions show firms including Cantor Fitzgerald, BMO Capital, Bernstein and Raymond James raising targets, while Truist, Evercore ISI and TD Cowen reduced theirs. Ratings often remained positive even when the target moved down.

The disagreement is about conversion, not only demand

The central question is no longer whether AI infrastructure demand exists. Broadcom's reported AI chip sales and customer pipeline make that difficult to dispute. The harder question is how much of the disclosed opportunity converts into revenue on the timetable investors have already priced into the stock.

Some analysts highlighted supply availability and a broader custom-chip customer base as sources of upside. Others focused on customer concentration, the economics of large AI customers and competitive pressure. Those differences can produce materially different valuation outcomes even when analysts agree on the direction of demand.

A target cut is not automatically a downgrade

This batch of research actions also illustrates a common mistake in reading analyst headlines. TD Cowen could lower a target while retaining a Buy recommendation, just as Evercore could trim a target while keeping an Outperform view. The rating answers a relative investment question; the target is a model output that can change for several reasons.

That is why Global Markets Review now tracks the recommendation, old target and new target as separate fields. A headline that says an analyst 'cut Broadcom' can imply a more fundamental change than the research action actually represents.

What to watch next

The most useful confirmation will come from Broadcom's subsequent AI revenue, customer concentration and margin disclosures. If the company delivers the steep growth path while maintaining attractive economics, analysts using more conservative deployment assumptions may have to revise upward. If supply, customer funding or competition slows conversion, the more cautious targets will look less anomalous.

For investors, the range is information in its own right. Wide target dispersion means the market is debating the shape and value of future cash flows, not merely updating one agreed forecast.