Broadcom Shares Fall Despite Better-Than-Expected Earnings

The Silicon Valley chipmaker provided a better-than-expected revenue outlook, though it fell short of some higher analyst expectations.

A smartphone with a displayed Broadcom logo is placed on a computer motherboard in an illustration taken on March 6, 2023. Dado Ruvic/Illustration/Reuters

Broadcom Inc., whose shares are trading just below a new all-time high, narrowly beat Wall Street expectations on June 5 after the Silicon Valley chipmaker and AI infrastructure giant reported its fiscal second-quarter earnings following the market’s close in New York.

For the period ended May 4, the Palo Alto, California-based company reported earnings of $4.9 billion, or $1.03 per share, up 134 percent from $2.1 billion, or 44 cents per share, in the same quarter of fiscal 2024. Revenue rose 20 percent to $15 billion, compared with $12.4 billion a year earlier.

On an adjusted basis, Broadcom reported net income of $7.78 billion, or $1.58 per share, narrowly topping the Wall Street consensus estimate of $1.57, according to FactSet. Revenue also slightly exceeded analysts’ expectations of $14.98 billion.

Although second-quarter earnings and revenue beat Wall Street expectations, its shares fell 4.01 percent in after-hours trading following a third-quarter revenue outlook of $15.8 billion. The guidance exceeded analysts’ estimate of $15.75 billion, according to FactSet, but fell short of some higher expectations. The company also projected adjusted earnings to be at least 66 percent of anticipated revenue.

Broadcom shares closed at $259.93 on the Nasdaq, down 0.44 percent for the day. Over the past month, the stock has jumped nearly 30 percent as the company stakes out a leading position in the AI infrastructure market, reaching an all-time high of $265.43 on June 4.

During the company’s June 5 after-market conference call, Broadcom CEO Hock Tan stated that the company expects its artificial intelligence (AI) semiconductor revenue to accelerate to $5.1 billion in the third quarter. That follows 46 percent year-over-year growth in the second quarter when AI revenue totaled $4.4 billion.

Tan specifically mentioned Broadcom’s announcement earlier this week that the company is now shipping its new Tomahawk 6, a next-generation networking switch designed to support large AI clusters at hyper-scale data centers.

According to Tan, the Tomahawk 6 will feature unprecedented scale, energy efficiency, and AI capabilities, enabling the super switch to power the next generation of AI networks with the company’s patented routing features and interconnect options. He said the Tomahawk 6 is also designed to meet the demands of AI clusters with more than 1 million XPUs, which represent the serviceable market for custom AI accelerators.

“This flattening of the AI cluster is huge because it enables much better performance in training next-generation frontier models through a lower latency, higher bandwidth, and lower power,” Tan said.

In late 2024, Tan surprised Wall Street during the company’s fourth-quarter earnings call when he said the serviceable addressable market for AI semiconductors could reach $60 billion to $90 billion by fiscal 2027, based on three large customers and additional prospects that Broadcom expects will make significant AI deployments.

During questioning from several analysts on June 5, Tan reiterated his 2024 prediction that Broadcom would see an acceleration of AI hyperscale demand into the back half of 2026 and early 2027. Without mentioning specific customers, Tan said the company is continuing to make “excellent progress on the multiyear journey of enabling our three customers and four prospects to deploy custom AI accelerators.”

“These partners are still unwavering in their plan to invest despite the uncertain economic environment. In fact, what we’ve seen recently is that they are doubling down,” he said.

Currently, Broadcom has partnerships with several Silicon Valley tech giants that are developing hyperscale AI data centers. For example, Amazon.com Inc., Microsoft Corp., Google LLC, Meta Platforms Inc., OpenAI, and Elon Musk’s xAI have all announced plans for hyperscaler data centers that require high-performance graphics processing units (GPUs) and other AI accelerators.

OpenAI, Oracle, MGX, and Softbank announced the Stargate Project on Jan. 21. This new venture intends to invest $500 billion over the next four years to build new AI infrastructure in Abilene, Texas. Arm, Microsoft, NVIDIA, and Oracle are the other key initial technology partners that will immediately deploy an initial $100 billion investment.

Musk, who submitted a $97.4 billion bid to take over OpenAI in February, is also rapidly expanding xAI’s $6 billion Colossus supercomputer in Memphis, Tenn. Meta announced in December 2024 that it is building a monumental $10 billion AI data center project in Richland Parish, Louisiana, calling it “a transformational investment that cements the state’s status as a major innovation hub and leader in the global digital revolution.”

In response to questions about the Trump administration’s decision in May to rescind the expanded export controls on AI implemented by the previous administration in January, Tan told analysts that he was unsure how Broadcom would be impacted. Under the Biden administration’s AI Diffusion Rule, issued on January 15, new licensing and quota regimes for advanced AI chips were introduced, along with a tiered framework for assessing export restrictions to various jurisdictions.

“Nobody can give anybody comfort in this environment ... Rules are changing quite dramatically as trade bilateral trade agreements continue to be negotiated in a very, very dynamic environment,” Tan said.