Apple Inc. is currently exploring strategic acquisitions within the semiconductor chip industry. This move aims to overcome performance bottlenecks in its artificial intelligence (AI) server chip development and reduce its deep dependence on competitor NVIDIA's computing infrastructure.
Insiders reveal that Apple has been in frequent contact with investment banks and several semiconductor startups over recent months to evaluate potential acquisition targets. Historically, Apple has exercised significant restraint in its M&A strategy, favoring small-scale, technology-focused acquisitions. However, facing a critical window for capturing AI computing advantages, the company is breaking with its traditional investment patterns. This strategic shift highlights the severe technical constraints it faces in developing a self-sufficient computing supply chain.
Currently, Apple's in-house developed M2 Ultra chips, deployed in its data centers, have shown unsatisfactory performance when handling highly complex AI large language models. Earlier this year, when upgrading its Siri voice assistant, Apple's own servers were unable to meet the massive computational demands of Alphabet's large language model, Gemini. This forced Apple to compromise by renting cloud services from Alphabet that are powered by NVIDIA chips. Furthermore, Apple's next-generation, codenamed "Baltra," in-house AI server chip, originally slated for release this year, has encountered development delays. This setback means its computing power shortfall may be difficult to address through internal research and development in the short term.
Concurrently, Apple is undergoing significant shifts in its senior leadership and financial policies, which provide a mechanism for facilitating large-scale acquisitions. In September, Apple's Senior Vice President of Hardware Engineering, John Ternus, will take over as CEO from Tim Cook. Meanwhile, chip engineering head Johny Srouji will oversee all hardware engineering, including semiconductors. On the financial front, Apple's CFO, Kevan Parekh, explicitly stated during the Q2 earnings call that the company is gradually moving away from its long-standing "net cash neutral" financial policy. Analysts note this change signifies Apple will end its practice of deploying large amounts of cash for buybacks and dividends. Instead, it will retain substantial cash reserves to fund potential multi-billion dollar semiconductor acquisitions led by its new executive team.
Historically, Apple's chip design prowess originated from its 2008 acquisition of startup PA Semi, leading to significant success in the mobile device chip market. However, in the high-compute AI server chip arena, which involves vastly different power consumption profiles and technical architectures, Apple's existing mobile design expertise has reached its limits.
To rebuild its competitive edge in AI, Apple has established a clear roadmap for catching up in computing power. Reports indicate Apple is currently advancing server upgrades based on the M5 Ultra chip and plans to launch a significantly more powerful M7 Ultra chip by 2029. The latter is expected to feature memory bandwidth of up to 1.5 terabytes (TB), targeting technical specifications that compete directly with NVIDIA's Blackwell chip. Beyond seeking external acquisitions, Apple is also mitigating R&D risks by deepening its foundry partnerships with industry giants. According to a recent securities filing by Broadcom, Apple has extended its long-term technical collaboration with the company until 2031. The two will continue to work closely on cutting-edge areas, including custom AI chips.