The technology sector closed the trading week under significant pressure, with the NASDAQ 100 dropping 2%, a selloff driven by execution concerns and geopolitical tensions, as discussed by Caroline Hyde in New York and Ed Ludlow in San Francisco on Bloomberg Technology. The hosts opened the segment highlighting the juxtaposition of a lower U.S. market, weighed down by tech stocks like Nvidia, against Chinese names trading higher, a dynamic shaped by major domestic policy news.
The immediate source of the market's "AI angst" was Broadcom, which was headed for its biggest stock decline since January. Despite reporting a massive $73 billion backlog in AI-specific sales, the company's stock slid because CEO Hock Tan "held off" from giving a clear, forward-looking revenue number, which Ed Ludlow noted was what the market "wanted". Analysts pointed out on Bloomberg Technology that AI sales were actually "above expectations," with the next quarter's outlook of $8.2 billion far exceeding the $6.8 billion consensus. The true "hiccup" concerned the margin structure of its deal with Anthropic, where $21 billion of the backlog will go toward a lower-margin TPU system sale. However, Karl Ackerman, a guest on Bloomberg Technology, argued that investors are being "shortsighted," overlooking the fact that Broadcom is moving toward a "full solution stack," providing both AI compute and networking, akin to Nvidia, and offering the entire IP stack in-house.
Further compounding the market’s unease was breaking news from Bloomberg Technology regarding Oracle. Ed Ludlow reported that Oracle's shares dropped almost 5% after the story broke that the company will be delaying the completion dates for several gigawatt-scale data centers being developed for OpenAI, pushing the timeline to 2028 from an initial plan of 2027. The delay is reportedly due to labor and material shortages. Caroline Hyde questioned whether this simply meant returns on AI investment would be "jam tomorrow, not today," raising fears about the ability to execute on the necessary infrastructure. The scale of the project is unprecedented, with Oracle attempting to build five gigawatt-scale data centers simultaneously, an ambition so high that delays due to labor and material constraints were "not a huge shocker".
Meanwhile, the geopolitical tech conflict intensified, with Caroline Hyde reporting that China is planning to allocate a monumental sum to its domestic semiconductor sector. Bloomberg Technology reported that the plan is to pour up to $70 billions of state money into the sector, which would be China’s largest ever semiconductor-specific incentive package. This funding is roughly on a similar scale to the U.S.'s $52 billion industrial policy effort. China is seeking to support its AI chipmakers, such as Huawei, and manufacturers like SMIC. Caroline Hyde highlighted that the Chinese government's commitment to self-sufficiency is so strong that there are indications they may reject the U.S. strategy of selling less advanced chips, like the H-200, confirming their goal is to prop up national champions rather than rely on U.S. supply.
Despite the market jitters, one expert, Margie Patel, maintained a positive outlook on Bloomberg Technology. She asserted that the "long-term trends are still in place" and that the current selloff, with the NASDAQ 100 down 2%, represents an end-of-year buying opportunity. She argued that the pressure on the sector is due to short-term traders preserving gains, not a change in fundamentals. The overall message relayed by Bloomberg Technology was that while the AI boom is undeniable, the current moment is defined by infrastructure bottlenecks and the necessity for companies to demonstrate innovation and flawless execution to satisfy highly demanding investors.
