Intel has had a busy few weeks. The chipmaker is up nearly 70% so far in 2026, fueled by two major headlines: a deepened partnership with Google and a surprise entry into Elon Musk’s Terafab semiconductor project.
Intel Corporation, INTC
The Google deal centers on future Intel Xeon chip generations. Google Cloud already runs Intel Xeon 6 processors across its C4 and N4 instances, handling both AI workloads and general computing tasks. The expanded agreement keeps Intel embedded in Google’s infrastructure as AI demand continues to climb.
The two companies will also expand their work on custom infrastructure processing units, or IPUs. These chips offload network, storage, and security tasks away from main CPUs, freeing up compute capacity and cutting power use. Google’s Amin Vahdat called CPUs and infrastructure acceleration “a cornerstone of AI systems.”
A week after the Google update, Intel confirmed it was joining Terafab — Musk’s project to produce semiconductors for Tesla, SpaceX, and xAI. The goal is to hit one terawatt of compute per year. For context, all leading-edge AI chips being produced globally today add up to roughly 20 gigawatts — just 2% of that target.
Intel’s exact role in Terafab remains unclear. Whether the company will license technology, co-invest, or operate part of the fab is still unknown.
But on the same day as the Terafab announcement, Intel Foundry researchers published something telling: a breakthrough on ultrathin gallium nitride, or GaN, chiplets.
GaN is more resilient than silicon in high-voltage and high-radiation environments — two conditions that are very common in space. Intel’s team found a way to grow GaN directly on a standard 300mm wafer using existing production equipment, cutting manufacturing costs. They also developed a thinning method that brought the silicon base down to just 19 microns — roughly one-fifth the width of a human hair.
The bigger technical feat: Intel managed to combine GaN power electronics and silicon logic on the same chiplet. Normally, these two types of transistors have to be kept apart because power chips generate heat and electrical noise that can disrupt logic circuits. Keeping them separate means bigger, heavier systems.
Intel’s layer transfer process puts them together in one compact package. Testing confirmed the chiplets hold up under high-stress conditions.
For SpaceX, lighter and more radiation-tolerant chips translate directly to cost savings. Launch prices currently range from $1,000 to $10,000 per pound of payload. Shaving weight anywhere on the rocket matters.
Musk has said most of Terafab’s chip output is earmarked for SpaceX, both for satellite-based AI data centers and a broader space industrial economy.
Despite the stock’s sharp run, Wall Street is cautious. Of 35 analysts covering Intel, the consensus is a Hold. The average price target sits at $50.83, roughly 19% below where the stock trades today.
The post Intel (INTC) Stock Rallies 70% After Google Deal and Surprise Terafab Move appeared first on CoinCentral.


