Last Friday marked a historic day for Broadcom as its stock surged by 24%, setting an unprecedented record and pushing its market capitalization past the $1 trillion mark, reaching an impressive $1.05 trillionThis leap translated to a staggering increase of $206 billion in market value within a single day (roughly equivalent to 1.5 trillion yuan). With this substantial rise, Broadcom became the third semiconductor company in the world following Nvidia and TSMC to join the ranks of those valued at over $1 trillion.

The extraordinary performance in the capital market can be attributed to Broadcom's strategic acquisition of VMware for $69 billion, with a key focus on the demand for high-performance AI chips amid the growing generative AI eraFinancial results showcased the growth clearly, as the revenue for the fourth quarter of the 2024 fiscal year reached $14.054 billion, marking a 51% year-over-year increase

Advertisements

Additionally, the net profit stood at $4.324 billion, while the adjusted EBITDA culminated in a net profit of $9.089 billion, showing a remarkable 65% increaseFor the entire fiscal year, revenues hit $51.6 billion, reflecting a 44% growth compared to previous levels.

At the heart of Broadcom's success lies its AI segment, which emerged as the primary engine for revenue growthIn the 2024 fiscal year, the revenue attributable to AI operations reached $12.2 billion, a jaw-dropping increase of 220% year-over-yearThis surge is largely driven by the innovations in artificial intelligence XPU technology, which propelled the semiconductor division to break records with $30.1 billion in revenue, a 58% increase from the previous yearSuch stellar results have caused Broadcom’s stock to skyrocket and thus become the ninth corporation globally to surpass a $1 trillion market valuation.

Since the advent of ChatGPT, the realm of artificial intelligence has entered a new generative phase, igniting a fresh wave of innovation

Advertisements

Numerous large language models have emerged, resembling a tidal wave that underscores the insatiable demand for effective AI solutionsHowever, it is crucial to point out that training these colossal models necessitates creating potent AI infrastructures, a feat that businesses like Nvidia and Broadcom have adeptly seized upon to reap the benefits of this AI joltNvidia, in particular, has become synonymous with high-performance chips, which are now a pivotal resource for firms looking to tap into the potential of AI.

The demand for Nvidia’s chips has escalated as companies scramble to enhance their AI capabilitiesThe primary tech giants—Meta, Google, Microsoft, and Amazon—have poured a staggering $200 billion into AI, with a considerable portion allocated towards purchasing Nvidia hardwareThe company’s performance has skyrocketed, yielding surging revenue numbers accompanied by a market value that has reached an astonishing $3.29 trillion

Advertisements

This meteoric rise is deeply rooted in investor optimism regarding the commercial prospects of AIYet, amidst Nvidia’s growth, there lies a pervasive concern about its monopolistic grip on the AI chip market.

Tech leaders are increasingly recognizing the need for diversification in chip supply chains to mitigate reliance on NvidiaTo this end, there is a shift towards pursuing custom-made chips that cater to varied application scenarios—an effort aimed at reducing dependency on Nvidia’s offerings while meeting diverse demands effectively.

Broadcom has disclosed its collaborative efforts with three major cloud clients to develop customized AI chips, predicting that by 2027 each client will deploy around 1 million AI chipsIn a recent earnings call, Broadcom CEO Hock Tan indicated that the custom AI chip sector could generate revenues between $60 billion and $90 billion by 2027. Such ambitions demonstrate the escalating demand for tailored AI solutions that not only fulfill these companies' requirements but also lessen their reliance on Nvidia.

This anticipated proliferation of custom AI solutions suggests that Broadcom could potentially disrupt Nvidia’s longstanding dominance

• Financial Services Go Digital: Trends and Challenges
• Fed Cuts Rates, Dow Plunges 1,000 Points
• Can AI Drive Future Economic Growth?
• NVIDIA Market Valuation Surpasses $1 Trillion
• Is $1 Trillion About to Flow Back to A-Share Markets?

Wall Street is also keeping a close watch on the demand for ASICs, particularly from large cloud computing firms like Google, which signal a shift in the competitive landscape.

AI ASICs, specifically tailored for distinct task requirements, are predominantly utilized for AI computationsThis has contributed to their designation as customized AI chips within the industryCompanies such as Marvell also find themselves benefiting from the robust demand for AI solutions, notably in relation to sales of new types of customized AI chips dedicated to Amazon and other large data centersThe significant increase in AI business revenues has offset declines in performance related to telecommunications, automotive, and other sectors.

By offering customized ASIC solutions, both Broadcom and Marvell have managed to carve out competitive advantages in the market, showcased in their impressive financial reports which affirm the immense strength of ASIC demand

The backdrop of fierce competition among major tech companies vying for dominance in the AI large model arena has positioned Broadcom as a critical player in providing necessary chips to giants like Google, Apple, and Meta, indicating a strategic pivot towards customization.

Interestingly, many companies that provide large models have struggled to monetize their offerings effectively; even OpenAI has found itself in a situation where expenditures outweigh revenues, continuously exacerbated by the soaring costs of training AI models and maintaining server infrastructures.

The competitive edge of AI large models hinges substantially on possessing a robust AI computational infrastructureHowever, for many teams behind these large models, the steep operational costs pose substantial barriers to achieving profitabilityThe ones who actually cash in from this burgeoning AI model sector are predominantly chip manufacturers, particularly Nvidia, which is thriving amidst this demand.

AI chip manufacturers are often perceived as the "shovel sellers" in the AI large model mining landscape