Good morning. The smart money seemed to believe that 2025—at least the second half of it—would be the year for a wave of tech IPOs, following a multiyear drought. But things are suddenly looking complicated.
IPO activity for venture capital-backed companies had already slowed to a pace not seen since 2011, per PitchBook. Then DeepSeek happened, rattling markets public and private alike…and the categories for half the leading IPO candidates.
So now? In the words of the short-lived ’90s sitcom My Brother and Me: Don’t hold your breath. —Andrew Nusca
Want to send thoughts or suggestions to Data Sheet? Drop a line here.
Apple posts record revenue (with a big slide in China sales)
Apple shared its Q1 earnings on Thursday, recording the most revenue in a single quarter it ever has but logging a double-digit decline in China.
Total quarterly revenue was $124.3 billion, up 4% and beating the Wall Street consensus.
iPhone revenue specifically was $69.14 billion, a couple billion lower than estimates and Apple’s biggest miss since 2023, when it experienced iPhone 14 production issues. Apple’s fiscal Q1 is the first full period of iPhone 16 sales as well as the first quarter reflecting the release of Apple Intelligence.
But the most interesting story is Apple’s business in Greater China, representing the mainland, Hong Kong, and Taiwan. Quarterly sales in the region declined 11.1% to $18.51 billion, echoing a similar drop (12.9%) in the market during the same quarter one year ago.
In an interview with CNBC, Apple CEO Tim Cook blamed “channel inventory” (that is, the product sitting between manufacturer and consumer), missing Apple Intelligence features (they were neither available in China nor in Chinese; the latter is planned for launch in April), and the lack of a national subsidy—since issued—to support sales.
The company said it expects growth in the current quarter of “low to mid single digits” on an annual basis; analysts expect earnings per share of $1.66 on revenues of $95.46 billion. —AN
OpenAI in talks to raise $40bn at a $300bn valuation
The money just keeps flowing toward Sam Altman.
OpenAI is reportedly in talks to raise up to $40 billion in a funding round that would value the ChatGPT-maker at either $260 billion, $300 billion, or $340 billion, depending on the report.
To put that in perspective, OpenAI raised $6.6 billion in its last round, valuing it at $157 billion. That was in October 2024. At the time, it was considered one of the largest venture funding rounds in history.
But as the Microsoft-backed company turns toward a for-profit structure and prepares for stiffer competition with rivals xAI, Anthropic, and most of Big Tech, investors are keen to pile on—even if the Chinese AI startup DeepSeek has challenged assumptions about what it costs to develop and deploy cutting-edge AI.
A Wall Street Journal report notes that SoftBank would lead a round that would, in part, go toward the Stargate Project joint venture with SoftBank, Oracle, and Abu Dhabi investment firm MGX that has pledged to invest $500 billion in U.S. AI infrastructure.
A Reuters report adds that SoftBank would invest $15 billion to $25 billion in OpenAI—some of which could be used for Stargate—on top of its existing $15 billion Stargate commitment. That would help Masa Son’s conglomerate leapfrog Microsoft as OpenAI’s biggest backer. —AN
Cerebras says it’s been ‘crushed with demand’ for DeepSeek
The upstart AI chip company Cerebras has started offering China’s market-shaking DeepSeek on its U.S. servers.
Andrew Feldman, Cerebras’s CEO, told Fortune Thursday that there was huge enthusiasm among his enterprise customers for DeepSeek, which is proving disruptive because it matches the most cutting-edge “reasoning” models from companies such as OpenAI while costing far less to train and use.
“We were crushed with demand,” Feldman said of the 10 days since DeepSeek released R1.
There are significant concerns over DeepSeek’s safety and inherent biases. Certainly when using R1 through DeepSeek’s official app or web interface, the model censors itself when asked about topics deemed sensitive to Beijing. It has also proved vulnerable to manipulation, making it possible to use the model to generate things like bomb-making instructions.
Like American peers such as rival chipmaker Groq and AI “answer engine” Perplexity, which have also started offering DeepSeek to their users in the past few days, Cerebras is hoping U.S.-based hosting will alleviate some of those concerns for enterprise users.
“If you use [DeepSeek’s] app, which is now the most popular app in the world, you are sending your data to China,” Feldman said. “Don’t do that. You can use the LLM served by U.S. companies in the U.S. like us, Perplexity, and others.” —David Meyer
More data
—Intel reports Q4 earnings. Revenue was down 7% YoY to $14.26bn, beating estimates; Q1 sales forecast is below estimates.
—U.S. sues to block HPE-Juniper Networks deal. DOJ argues the $14bn acquisition would harm enterprise wireless equipment competition.
—Amazon ramps spending on X. A strategic shift for a different political climate in the U.S.
—Europol takes down two of the largest hacking forums. German authorities arrest two suspects and seize 17 servers supporting Cracked and Nulled.
—Crypto super PAC preps for 2026 midterm elections. Fairshake says $116 million raised so far.
—Samsung reports Q4 earnings. Revenue up 12% YoY to about $52.2bn, beating estimates; noted semiconductor weakness, growth plans via AI smartphones.
—Did DeepSeek buy Nvidia AI chips via Singapore? U.S. authorities are investigating.
—Amazon sues Washington state agency to block release of company records to Jeff Bezos-owned Washington Post. (Awkward!)
Endstop triggered
