AI data centers

The AI boom is driving massive data center expansion, creating an urgent need for reliable power. Bloom Energy has emerged as a critical provider, supplying onsite, always-on energy solutions through its solid oxide fuel cells. This shift from clean energy focus to powering AI infrastructure has led to significant stock growth. Key partnerships and major data center projects are bolstering Bloom’s position as a preferred power provider, though profitability and production scaling remain areas of focus.

Himax Technologies reaffirms its strong partnership with FOCI on Co-Packaged Optics (CPO). Their joint development leverages Himax’s WLO nano imprinting technology with FOCI’s silicon photonics expertise. The first-generation CPO solution is undergoing customer validation for mass production in 2026. Future efforts focus on next-generation high-speed optical transmission for AI data centers and HPC. This collaboration highlights Himax’s strategic expansion into advanced optical technologies.

Semiconductor company Marvell announced a deal to acquire AI‑focused startup Celestial for an initial cash‑and‑stock price of at least $3.25 billion, potentially rising to $5.5 billion if revenue targets are met. The acquisition adds Celestial’s silicon‑photonic “photonic fabric” to Marvell’s networking portfolio, aiming to address optical‑interconnect bottlenecks in large‑scale AI training and inference. Marvell’s shares rose 13 % after a strong Q3 earnings report and a forecasted 25 % increase in data‑center revenue. The deal is expected to close early next year.

OpenAI and Foxconn are partnering to co-develop and manufacture AI data center components in the US. This aims to accelerate infrastructure deployment, secure domestic manufacturing, and address hardware bottlenecks. Foxconn will leverage its US facilities to manufacture power systems, networking, and cooling solutions, aligning with supply chain security goals. OpenAI’s CEO highlights this as a “generational opportunity” for US reindustrialization. This partnership is part of OpenAI’s broader strategy, involving significant investments and alliances to build AI infrastructure. However, Foxconn’s past US manufacturing ventures present cautionary lessons.

Qnity Electronics, a recent spin-off from DuPont focused on semiconductor solutions, debuted strongly on the NYSE, drawing positive analyst attention. CNBC’s Investing Club received Qnity shares representing 2.04% of its Charitable Trust’s holdings. Freed from DuPont’s broader portfolio, Qnity aims to capitalize on the growing semiconductor market driven by AI, with 65% of its offerings linked to the sector. The company projects substantial market expansion, fueled by demand for AI data centers, and has partnerships with industry leaders like Nvidia, TSMC, and Samsung. Analysts are bullish, anticipating Qnity’s business update later this week.

Amazon’s $11 billion “Project Rainier” in Indiana is a massive, operational AI data center training cutting-edge AI models for Anthropic using Amazon’s Trainium chips. This facility, part of a trillion-dollar industry-wide investment, highlights the competition to build supercomputing infrastructure for AI. While raising questions about energy and community impact, Amazon leverages its logistics expertise to deploy this unprecedented AI infrastructure. Amazon and Anthropic are collaborating on future chips to improve performance for their AI models.

CleanSpark (CLSK), a Bitcoin miner, is diversifying into AI data centers, leveraging its rapid infrastructure deployment capabilities. CleanSpark secured a 100MW site in Cheyenne, Wyoming, outpacing Microsoft, thanks to its “speed to market.” This move addresses power supply competition and halving’s impact on Bitcoin mining profitability. CleanSpark partners with Submer for advanced AI infrastructure. Its ability to quickly scale power and redirect energy to the grid provides a unique competitive advantage, highlighting the strategic importance of energy-efficient solutions. CleanSpark’s strategy maximizes asset utilization and long-term profitability.

Meta and Oracle are upgrading their AI data centers with NVIDIA’s Spectrum-X Ethernet, designed for the demands of large-scale AI. This shift to open networking aims to improve AI training efficiency across massive compute clusters. NVIDIA’s MGX system offers flexibility and scalability, while innovations like 800-volt DC power delivery and power-smoothing technology address power efficiency. Spectrum-X supports scalable connectivity between data centers and integrates with various NOSs, enhancing AI infrastructure accessibility. The upcoming Vera Rubin architecture will further support next-gen AI factories.

Sivers Semiconductors announced an $8.3 million directed share issue to Swedish and international institutional investors, managed by Pareto Securities AB. Funds will accelerate R&D (60%), support strategic investments like acquisitions and talent (20%), and optimize capital structure (20%). The company is focusing on high-growth markets in SATCOM and AI data centers and aims to strengthen its position through strategic partnerships, expanded product offerings, and a growing customer pipeline. Board members and the CEO intend to participate in the offering.

NVIDIA’s Spectrum-XGS Ethernet aims to link geographically dispersed AI data centers, addressing the capacity limitations of single-site facilities. This “scale-across” approach complements “scale-up” and “scale-out” strategies, using distance-adaptive algorithms and advanced congestion control to minimize latency and optimize network performance. Cloud provider CoreWeave will be an early adopter. The technology seeks to reshape AI data center planning, potentially reducing costs and improving performance by distributing workloads across multiple sites. Its success will depend on real-world effectiveness and navigating complexities beyond networking.