Shares in Broadcom rose 3.7% in premarket trading on Tuesday after the chip designer announced it would produce future versions of artificial intelligence chips for Google, and signed an expanded deal with Anthropic.

The latter deal will give the AI startup access to about 3.5 gigawatts worth of computing capacity, drawing on Google’s AI processors and providing some relief for the share price.

Broadcom stock has endured a tough start to 2025, falling almost 10% year-to-date as investors have become increasingly jittery around the bull case for technology stocks. It has also been caught up in the maelstrom of the wider market caused by the U.S.-Israeli attack on Iran that began on February 28. 

The drift downwards comes despite a blowout earnings report in March, during which CEO Hock Tan touted strong future demand for the company’s chips.

He told analysts last month that he anticipates AI chip revenue in 2027 that’s “significantly in excess of $100 billion” as demand mounts for designing custom silicon.

Analysts remain bullish following Monday’s announcements. 

“The deal includes revenue commitments across that timeline, which should help ease some of the recent nervousness around TPU competition and give a clearer signal that its largest customer sees meaningful demand visibility well into the future,” said Matt Britzman, senior equity analyst at Hargreaves Lansdown. 

“We already saw upside to medium-term revenue and profit expectations off the back of recent results; these new deals help underpin that idea if deployment ramps as planned.”

Citi analysts maintained their ‘Buy’ rating for the stock, backing Broadcom to surpass its $100 billion revenue target to more than $130 billion off the back of the Google deal. 

— CNBC’s Jordan Novet also contributed to this report.

Broadcom‘s stock surged about 5% on Thursday as CEO Hock Tan touted strong demand for the company’s chips amid the artificial intelligence boom.

He told analysts on Wednesday that he anticipates AI chip revenue in 2027 that’s “significantly in excess of $100 billion” as demand mounts for designing custom silicon.

That far surpassed many bullish estimates from Wall Street analysts, who now anticipate more potential upside after Tan said the company is nearing 10 gigawatts of capacity between six customers.

Analysts at JPMorgan estimate that the company can reach between $12 billion and $15 billion in revenue per gigawatt by 2027 and lifted the firm’s AI revenue estimates “conservatively” to $120 billion or more.

The company’s “leadership in AI networking and custom silicon enables the lowest inference cost for its hyperscaler customers, and we see it delivering ongoing cost reductions on pace with market leader Nvidia,” wrote analysts at Goldman Sachs.

The comments came alongside Broadcom’s better-than-expected quarterly results, in which AI revenue more than doubled due to AI accelerator and networking demand.

Over the last few months, chipmakers have grappled with a shortage in high-bandwidth memory as AI demand snaps up supply. But Tan told analysts that the company has secured memory and leading-edge wafer supply through 2028.

Tan managed to convince investors of Broadcom’s sustainable growth trajectory, allaying fears about the company’s profitability and questions about whether Broadcom shipping more racks packed with AI chips would weigh on margins. By comparison, leading AI chipmaker Nvidia’s blowout report last month wasn’t enough to convince investors.

“We have gotten our yields, we’ve gotten our cost to the point where the model we have in AI will be fairly consistent with the models we have in the rest of the semiconductor business,” Tan told analysts.

More hyperscalers making their own chips has also sparked some fears that Broadcom could get cut out of the market. However, Tan said the difficulty of competing against a juggernaut like Nvidia should benefit the company over “many years to come.”

Large language model makers “cannot afford to have a chip that is just good enough,” he said. “You need the best chips around because you’re competing against other LLM players, and most of all, you’re also competing against Nvidia, who is by no means letting their guard down.”

Broadcom’s upbeat results benefited Credo and Amphenol on bets that customers are opting for copper connectivity — the chipmaker’s core business segment — over optical technology to connect AI servers. Shares rallied 10% and 4%, respectively.

Lumentum and Coherent, which make newer optical tech, fell over 4% each.

— CNBC’s Jordan Novet and Katie Tarasov contributed reporting

Broadcom is set to report fiscal first-quarter results after market close on Wednesday.

Here’s what analysts polled by LSEG are expecting:

• Earnings per share: $2.03 adjusted
• Revenue: $19.18 billion

At that scale, analysts are projecting nearly 29% revenue growth year over year.

The company has benefited from rising demand for artificial intelligence infrastructure, touting a 65% jump in AI revenue in fiscal 2025. Broadcom contributes to the development of Google’s tensor processing units, or TPUs.

In December, CEO Hock Tan called for fiscal first-quarter AI revenue to double to $8.2 billion.

So far in 2026, Broadcom shares have come down 9%, trailing the S&P 500 index, which is flat over the same period.

“In our view it is likely caught up in investor concerns that 2026 marks a peak” in capital expenditures on AI equipment, wrote analysts at Evercore, who recommend buying Broadcom stock, in a Monday note to clients.

Meanwhile, software stocks have been under pressure for weeks as investors ponder competitive threats from generative AI models. Broadcom bought server virtualization software company VMware in 2023.

Analysts are looking for $12.25 billion in revenue from semiconductor solutions and $7.02 billion from infrastructure software. During the quarter, Broadcom announced new Wi-Fi 8 chips.

For the fiscal second quarter, analysts polled by LSEG are expecting $2.17 in adjusted earnings per share on $20.56 billion in revenue.

Executives will discuss the results on a conference call starting at 5 p.m. ET.

WATCH: I’d pick Broadcom, Nvidia first if buying tech stocks, says Niles Investment’s Niles

In November, following Nvidia’s latest earnings beat, CEO Jensen Huang boasted to investors about his company’s position in artificial intelligence and said about the hottest startup in the space, “Everything that OpenAI does runs on Nvidia today.”

While it’s true that Nvidia maintains a dominant position in AI chips and is now the most valuable company in the world, competition is emerging, and OpenAI is doing everything it can to diversify as it pursues a historically aggressive expansion plan.

On Wednesday, OpenAI announced a $10 billion deal with chipmaker Cerebras, a relatively nascent player in the space but one that’s angling for the public market. It was the latest in a string of deals between OpenAI and the companies making the processors needed to build large language models and run increasingly sophisticated workloads.

Last year, OpenAI committed more than $1.4 trillion to infrastructure deals with companies including Nvidia, Advanced Micro Devices and Broadcom, en route to commanding a $500 billion private market valuation.

As OpenAI races to meet anticipated demand for its AI technology, it has signaled to the market that it wants as much processing power as it can find. Here are the major chip deals that OpenAI has signed as of January, and potential partners to keep an eye on in the future.

Nvidia 

Since its early days building out large language models, long before the launch of ChatGPT and the start of the generative AI boom, OpenAI has relied on Nvidia’s graphics processing units.

In 2025, that relationship went to another level. Following an investment in OpenAI in late 2024, Nvidia announced in September that it would commit $100 billion to support OpenAI as it builds and deploys at least 10 gigawatts of Nvidia systems. 

A gigawatt is a measure of power, and 10 gigawatts is roughly equivalent to the annual power consumption of 8 million U.S. households, according to a CNBC analysis of data from the Energy Information Administration. Huang said in September that 10 gigawatts will equate to between 4 million and 5 million GPUs. 

“This is a giant project,” Huang told CNBC at the time. 

OpenAI and Nvidia said the first phase of the project is expected to come online in the second half of this year on Nvidia’s Vera Rubin platform. However, during Nvidia’s quarterly earnings call in November, the company said there is “no assurance” that its agreement with OpenAI will progress beyond an announcement and to an official contract stage. 

Nvidia’s first investment of $10 billion will be deployed when the first gigawatt is completed, and investments will…

In at least one corner of the artificial intelligence market, sentiment has turned decidedly negative.

Broadcom, CoreWeave and Oracle, three companies intimately tied to the AI infrastructure buildout, all had another rough day on Wall Street on Monday after selling off sharply last week.

While the three stocks are all still solidly up for the year — CoreWeave held its market debut in March — the most recent trend suggests that investors are concerned about whether the returns on investment will ever justify the level of spending taking place.

“It definitely requires the ROI to be there to keep funding this AI investment,” Matt Witheiler, head of late-stage growth at Wellington Management, told CNBC’s “Money Movers” on Monday. “From what we’ve seen so far that ROI is there.”

Witheiler said the bullish side of the story is that, “every single AI company on the planet is saying if you give me more compute I can make more revenue.”

Still, the market was displeased last week with quarterly earnings reports from chipmaker Broadcom and cloud infrastructure supplier Oracle, even though both companies beat on revenue and issued forecasts showing that AI demand is soaring.

Oracle, which is now heavily reliant on the debt markets to fund its data center development, provided scant details about how it will continue to finance its commitments. The company said it would ramp up capital expenditures in the current fiscal year to $50 billion from an earlier forecast of $35 billion because of new contracts from the likes of Meta and Nvidia.

It’s also ratcheting up leases. As of Nov. 30, Oracle had $248 billion in lease commitments for data centers and cloud capacity commitments that will run for 15 to 19 years. That’s up 148% from the end of August.

Meanwhile, Broadcom CEO Hock Tan said he expects AI chip sales this quarter to double from a year earlier to $8.2 billion, driven by both custom chips as well as semiconductors for AI networking.

However, as the company spends heavily on more parts to produce server racks, investors are going to have to stomach a hit to profits. CFO Kirsten Spears said on Broadcom’s earnings call that “gross margins will be lower” for some of the company’s AI chip systems.

Broadcom shares fell 5.6% on Monday following an 11% slump on Friday, leaving them 18% below their record high reached on Wednesday.

Oracle dropped 2.7% on Monday and is now down 17% in the past three trading days. The company has lost 46% of its value since Sept. 10, when the stock had its best day since 1992 following disclosure of a massive AI backlog.

Venture capitalist Tomasz Tunguz, who focuses on enterprise software and AI, wrote in a Monday blog that Oracle’s recent fundraising binge has left it with a debt-to-equity ratio of 500%, “dwarfing its cloud computing peers.” Amazon, Microsoft, Meta and Google all have ratios between 7% and 23%, he wrote.

Tunguz, founder of Theory Ventures, said the other…

The sell-off in artificial intelligence stocks continued unabated Friday stateside. Broadcom shares tumbled more than 11% as investors grew concerned over lower margins and uncertain deals. Names such as Nvidia, Advanced Micro Devices and Oracle fell in sympathy, which caused major U.S. indexes to close lower.

It was a motif patterning the week. Even though the Dow Jones Industrial Average rose 1.1% week on week on the back of outperformance by financial stocks, tech names dragged down the S&P 500 and the Nasdaq Composite, which fell 0.6% and 1.6% respectively for the week.

That said, investors could have just been jittery amid the narrative of an apparent AI bubble, and were spooked by any sign of bad news. After all, Broadcom’s earnings — as well as its guidance for the current quarter — breezed past expectations.

“Frankly we aren’t sure what else one could desire as the company’s AI story continues to not only overdeliver but is doing it at an accelerating rate,” Bernstein analyst Stacy Rasgon, who has a “buy” rating on Broadcom, wrote in a Friday note.

Future prospects also look rosy, according to UBS. “We expect high profitability and the accelerating impact of the AI, power and resources, and longevity themes to drive 2026 performance,” said strategist Sagar Khandelwal.

But in the near term, investors may still be flighty, unless something concretely reassuring, such as Oracle achieving positive cash flow, reassures them the snapping sound is just a twig in the forest.

U.S. stocks dragged down by AI names. Major indexes fell Friday, a day after they hit record highs. The pan-European Stoxx 600 retreated almost 0.5%. Separately, the U.K. economy unexpectedly shrank 0.1% in the three months to October.

Oracle will finish data centers on time. The company issued its response to a Bloomberg report, which cited unnamed people, that Oracle will complete data centers for OpenAI in 2028 rather than 2027. “There have been no delays,” Oracle said.

Coinbase to have an in-house prediction market. It will be powered be Kalshi, a source close to the matter told CNBC, and is a play to expand asset classes available on the cryptocurrency exchange.

The end of the ‘Berkshire way’? Several aspects of Berkshire Hathaway’s leadership transition are signaling that the conglomerate is drifting away from the famously decentralized “Berkshire way,” CNBC’s Alex Crippen writes.

[PRO] China’s food security strategy. The spate between Beijing and Washington over soybean purchases has highlighted the evolution of China’s domestic agriculture industry. Goldman Sachs thinks this is the best way to play the sector.

Broadcom revealed during a September earnings call that it had signed a customer that had placed a $10 billion order for custom chips.

At the time, Broadcom didn’t say who it was, but on Thursday, CEO Hock Tan revealed that the mystery customer was AI lab Anthropic, which placed an order for the latest Google tensor processing units.

“We received a $10 billion order to sell the latest TPU Ironwood racks to Anthropic,” said Tan, speaking on Broadcom’s fourth-quarter earnings call on Thursday. He also said Anthropic had placed an additional $11 billion order with Broadcom in the company’s latest quarter.

While Broadcom typically doesn’t disclose its large customers, Tan’s September remark drew significant investor attention amid the AI infrastructure boom. A Broadcom official told CNBC in October that the mystery customer wasn’t OpenAI, which has its own agreement to purchase chips from the chipmaker.

Broadcom makes custom chips called ASICs, which some experts believe are more efficient for certain artificial intelligence algorithms than the market-dominating chips from Nvidia. Broadcom helps make Google’s TPUs, and last month, the search company bragged that it trained its state-of-the-art Gemini 3 model entirely on its TPUs.

The chipmaker calls its custom AI chips XPUs, and on Thursday, Tan said his company was delivering entire server racks — not just chips — to Anthropic, which is Broadcom’s fourth XPU customer.

Broadcom on Thursday also said that it has secured a fifth customer for its custom chip business. That customer placed a $1 billion order during the fourth quarter, but once again, Broadcom did not reveal the customer.

“It’s a real customer, and it will grow,” Tan said.

In late October, Anthropic and Google announced a sweeping cloud partnership that underscored just how quickly the AI infrastructure race is moving.

The deal is valued in the tens of billions of dollars. The agreement gives Anthropic access to as many 1 million Google TPUs, and is expected to bring well over a gigawatt of new AI compute capacity online in 2026.

Anthropic employs a multi-cloud, multi-chip strategy.

The startup spreads workloads across Google’s TPUs, Amazon’s custom Trainium chips and Nvidia graphics processing units. Anthropic’s various models are tuned to run on whichever platform makes the most sense for training, inference or research.

For Google, Anthropic’s bet on TPUs is exactly the kind of validation investors have been rewarding, as Wall Street increasingly ties Alphabet’s stock run-up to clear demand for its in-house AI chips.

Google Cloud CEO Thomas Kurian previously said Anthropic’s decision to dramatically expand its TPU usage reflects “the strong price-performance and efficiency” the startup has seen over several years.

After more…

Broadcom’s quarterly results and guidance sailed past Wall Street estimates. It didn’t matter.

The chipmaker’s shares plummeted 11% on Friday, their worst day since January, as investors ran for the exits on the artificial intelligence trade. Oracle dropped 4.5% a day after plunging 10% following its earnings report.

Nvidia and Advanced Micro Devices, the two leading makers of graphics processing units for AI workloads, slid about 3% and 5%, respectively.

AI has been the driver for the stock market and the broader economy this year, so any negative sentiment has potentially far-reaching consequences. The Nasdaq on Friday fell about 1.69%, and the S&P 500 declined by 1%.

The companies getting hit the hardest are the ones most closely tied to AI infrastructure, which has been booming as hyperscalers build out their data centers to try and meet what they describe as insatiable demand for compute-intensive AI services. Broadcom makes custom chips for many of the the largest tech companies, and saw its market cap about double each of the past two years before rallying again in 2025.

“This stock is up 75-80% year to date. You’re seeing a little bit of a pullback,” Vijay Rakesh, an analyst at Mizuho, told CNBC’s “Squawk on the Street” on Friday. “We would be buyers on this pullback.”

Mizuho raised its price target on the stock to $450 from $435. It closed on Friday just below $360.

“This is still where the growth is,” Rakesh said. “They are still the big supplier to Google on their entire hardware stack, to Meta, to Anthropic and even OpenAI coming down the road.”

Broadcom reported revenue growth of 28% during the quarter, largely due to a 74% increase in AI chip sales, to a total of $18.02 billion, topping the $17.49 billion average analyst estimate, according to LSEG. Adjusted earnings per share of $1.95 adjusted topped the $1.86 average estimate.

CEO Hock Tan said Broadcom expects AI chip sales this quarter to double from a year earlier to $8.2 billion, both from custom AI chips as well as semiconductors for AI networking.

One concern among investors is that margins are coming down, at least in the short term, due to higher up-front costs. CFO Kirsten Spears said on the earnings call that “gross margins will be lower” for some of Broadcom’s AI chip systems because the company will have to buy more parts to produce the server racks.

Broadcom also said it had a $73 billion backlog of AI orders over the next 18 months. Part of that is from $21 billion of orders from Anthropic, which the company revealed as a key customer on Thursday.

While OpenAI has been a highly touted customer following a multibillion-dollar agreement announced in October, Tan doused some hope for the deal, telling investors late Thursday that, “We do not expect much in ’26.”

Bernstein analyst Stacy Rasgon said in a note on Friday that “AI angst” was driving Broadcom’s shares lower.

“Frankly we aren’t sure what else one could desire as the…