As Apple Intelligence expands, the tech giant feels the deep sting of weak cloud power. To save its delayed AI server chip plans and rely less on NVIDIA, Apple acts swiftly. The firm is now talking with major banks and fresh tech startups. Indeed, they are looking to buy large AI chip companies. You can read how Apple hunts AI chip acquisitions to learn about these fierce new tactics. This bold move breaks their old habit of buying only small teams. Moreover, it hints at strong new plans from John Ternus, the hardware chief who will soon become CEO.
The Struggle of M2 Ultra and Baltra Delays
Right now, Apple uses a mixed strategy for smart tasks. Most light tasks run directly on the iPhone or Mac. However, heavy tasks, like the new Siri powered by Google Gemini, need a Private Cloud Compute system. Insiders say Apple still uses the M2 Ultra chip as the core of this private cloud. Originally, Apple built this chip just for Mac computers. Consequently, it struggles greatly under the massive weight of large language models.
Seeking Outside Help
To bridge this power gap, Apple must now outsource heavy AI work. They send these tasks to NVIDIA chips housed inside Google Cloud. To make matters worse, Apple’s custom AI server chip, named Baltra, has fallen far behind schedule. The company planned to launch it in 2026. Now, Apple must quickly find outside partners to boost its weak cloud power.
A Major Shift in Financial Strategy
In the past, Apple played it very safe with buyouts. They mostly spent small sums on tiny teams with special skills. Yet, the costly AI arms race demands a huge change in their money plans.
Dropping Old Money Rules
First, they are dropping their strict cash rules. Kevan Parekh, the finance chief, recently spoke to Wall Street experts. He stated Apple will no longer aim for a strict cash-neutral state. This major shift frees up vast sums of cash for massive buyouts.
Record-Breaking Buyouts
Second, their spending is reaching new heights. Just this past January, Apple spent nearly two billion dollars. They bought Q.ai, a firm that reads tiny facial cues to aid voice tools. This deal marks their second-largest buyout ever, right after buying Beats for three billion dollars in 2014. Additionally, Apple is talking with firms like PrismML. They want exclusive tech that shrinks huge AI models to run right on the iPhone.
Future Server Plans and Broadcom
Despite short-term pain, Apple’s future server map is taking shape. News reports reveal their future plans. Apple is working on a hugely powerful M7 Ultra chip. This future chip will boast up to 1.5TB of shared memory. It directly targets the NVIDIA Blackwell design, but it will not arrive until 2029. Also, Apple teamed up with Broadcom in 2024 to build stable network designs. Last week, Broadcom confirmed this vital bond will last until 2031.
Buying a New Tech Fortress
This urgent search for AI startups feels like a repeat of history. Back in 2008, Apple spent 278 million dollars to buy PA Semi. That smart move gave birth to their famous A-series and M-series chips. Today, Apple stands at a similar crossroads. For years, their chip teams focused only on devices with batteries. However, AI server chips are a totally different beast. They do not need long battery life. Instead, they demand massive data speed, strong links, and extreme cooling power.
The Push for True Independence
The M2 Ultra fails perfectly as an AI server core. This shows the fatal flaw of forcing a laptop chip into a data center. John Ternus will step up as the new CEO this September. Johny Srouji, the chip boss, also holds more power now. These leaders know that great software is simply not enough. If NVIDIA and Google control the cloud power, Apple can never truly win. Dropping their strict cash rule sounds the alarm for a fierce fight. Soon, Apple will launch its most aggressive spending spree ever to build its AI empire.
Support Our Threat Intelligence
If you find our technology report and cybersecurity news helpful, consider supporting our work.