What I Read This Week...

Micron's $100B project will generate 50,000 jobs, hold up to four semiconductor fabrication plants, and become the largest private investment in New York state's history.

Jan 10, 2026

What I Read This Week: a summary of the content that I consumed this week…

1) xAI Raises $20B and Invests $20B

On January 6th, xAI announced that their upsized Series E round finalized from the $15B target to $20B.

Approximately 600 million MAU are now reached across X and the Grok apps. Grok 5 is currently under training and a new investment of $20B to scale a data center in Southaven Mississippi was announced on January 8th. “MACROHARDRR” is now xAI’s third data center and within proximity to their recently acquired powerplant which will increase xAI’s compute power to nearly 2 gigawatts.

Mississippi’s Development Authority has approved xAI for its Data Center Incentive, providing sales and use tax exemption. This massive buildout marks the largest private investment in the state’s history. It will create thousands of jobs and there is already proactive action being taken for noise mitigation and emission modeling to take care of the nearby community.

2) Microns Massive $100B Megafab

Micron recently announced the largest semiconductor manufacturing facility in U.S. history. They will officially break ground on January 16th, for the megafab in Onondaga County, NY. This $100B project will generate 50,000 jobs, hold up to four semiconductor fabrication plants, and become the largest private investment in New York state’s history.

The company delivers a portfolio of high-performance DRAM, NAND, and NOR memory and storage products. Over 70% of Micron’s revenue is DRAM, which is where AI relevance shows up most clearly. Their high-bandwidth memory (HBM), is a specialized DRAM architecture stacked vertically and placed close to GPUs. As AI workloads scale, memory acts as the data conduit that allows GPUs to process the trillions of parameters required for modern reasoning and generative models.

Due to the massive demand spike from AI, DRAM prices have skyrocketed 171% yoy. Additionally HBM demand has grown far faster than standard DRAM, and supply is constrained because HBM is harder to manufacture, has lower yields, and only few players can make it at scale (SK Hynix, Samsung, and Micron). This has led Micron’s stock price to move up over 225% from ~$99 in Jan 2025 to ~$325 today.

3) NVIDIA: Six New Chips, One AI Supercomputer

Jensen Huang officially kicked off CES 2026 by unveiling the Vera Rubin platform. Rubin consists of six new chips designed to operate as one tightly coupled system. This includes next-generation GPUs, CPUs, networking, and interconnect silicon, all engineered together. The result is an ability to deliver up to a 10x reduction in inference token cost and 4x reduction in the number of GPUs to train MoE models when compared to their Blackwell platform.

Rubin becomes the new hardware that is specifically optimized for reasoning models. The addition of Inference Context Memory (ICM) solves the cost of long-term memory and context, one of the biggest bottlenecks in AI today. The Vera Rubin platform is in full production as of Jan 2026, with cloud partners (Microsoft, CoreWeave, Nebius) deploying in H2 2026.

Jensen also shared the breakthrough moment in physical AI leveraging NVIDIA Cosmos, an open world foundation model that creates synthetic data, prediction, and reasoning for spatial 3D inputs. This leads to their announcement of Alpamayo, the world’s first thinking, reasoning, autonomous automotive model that is trained end to end from camera to actuation. The Mercedes-Benz CLA partnership exemplifies the 8 year development that NVIDIA has put into their DRIVE AV system.