The reason is not that Broadcom has “risen the most,” but that the key threshold for the next generation of AI leaders is shifting from single-chip computing power to the economics of computing power per unit and system-level delivery capabilities. Ahead of Nvidia’s earnings, Goldman Sachs reiterated its Buy rating and $300 price target, forecasting FY3Q27 EPS 6% above consensus, on the rationale that gross margins would beat expectations and the risk to 2027 data center deployments was limited. This certainly shows that NVDA’s near-term fundamentals remain solid. But Goldman Sachs also listed power and chassis supply in 2028 as “major uncertainties,” and noted that AMD’s MI450X Helios rack system and custom ASICs from hyperscalers are creating real competition. This means that the logic behind GPU leadership is expanding from “whose individual chip is faster” to “who can deliver computing power per watt and per dollar with greater certainty.”

Broadcom is perfectly positioned at this turning point. After Google unveiled its eighth-generation TPU, JPMorgan reiterated its Overweight rating, noting that Broadcom still handles Google’s highest-performance TPU project, that the TPU roadmap has no gaps through 2031, and that revenue growth will not slow. The v8i Max inference TPU integrates four compute chips in a single package, doubling computing power, with production ramping up in the second half of 2027; the next-generation 2nm TPU is on track to enter production in mid-2028. The implications of this roadmap are clear: as cloud providers increasingly favor custom ASICs to lower inference costs and secure long-term computing capacity, Broadcom is one of the few companies able to translate “customer roadmaps” directly into its own “revenue trajectory.” Compared with NVDA’s reliance on sustained generational leadership in the general-purpose GPU ecosystem, Broadcom’s custom chip business is more “contract-driven” in terms of visibility, and less sensitive to financing conditions and high interest rates.

Valuation also provides a margin of safety. After this round of sell-offs, the S&P 500 technology sector’s forward P/E is around 17x, its first time below the broader market since 2017. Broadcom’s current share price does not carry the NVDA-style premium that assumes it “must always exceed expectations,” while JPMorgan has named it its top semiconductor pick, citing the long-term growth prospects for AI and data centers and its first-class margins. NVDA remains the most solid core holding in the computing ecosystem, and TSMC’s October revenue grew nearly 60% year over year, while advanced process capacity will effectively remain sold out through 2028, making it the supply chain’s most honest anchor. But if the question is “who will lead the next wave,” Broadcom’s custom ASIC strategy is more closely aligned with the 2027 investment theme of shifting from “expected investment” to “realized ROI.”

So my view is that the rebound will continue, but it will not be a broad-based rally. Instead, it will be a structural market shift from “computing power beta” to “computing economics alpha.” At this turning point, Broadcom has visible customer roadmaps, relatively moderate valuation, and a revenue mix that does not depend on the success or failure of any single GPU generation—all of which are the core reasons I believe it will lead the next wave.

View Original

This page contains third-party content and does not constitute any advice, nor does it represent Gate's endorsement of such views. For details, please see disclaimer.