China’s Chip Leap: From Zero to Global Contender in Storage and Manufacturing

In the wake of the U.S.–China tech showdown, China’s chip sector has made a startling turnaround. Ten years ago the country produced none of the memory chips—DRAM for phones and flash for storage—that power everyday gadgets. Today, home‑grown firms such as Changjiang Storage and Changxin Memory not only manufacture these components but rank among the world’s best, ensuring that smartphones, laptops and wearables can keep running even if foreign supplies are cut off. Beyond memory, Chinese fabs have pushed the cutting edge to a 7‑nanometer process, expanding capacity despite sanctions that limited access to advanced equipment. While most upstream tools—photoresists, etching machines and the like—are now sourced domestically, the only major gap remains the ultra‑high‑end lithography machines used for the tiniest chips. A private startup, Xindonglai, has already built a modest 300‑nanometer lithography system, marking the first home‑grown step in a field once completely foreign. Experts believe that by 2030 China could close the lithography gap entirely. Overall, the rapid progress in memory production, process technology and equipment localization signals a decisive stride toward self‑reliance in the global chip arena.

Read more

Positron Secures $230 Million to Take on Nvidia’s AI Chip Dominance

Semiconductor up‑start Positron has just closed a $230 million Series B round, a deal that could reshape the AI‑hardware landscape. The funding, led in part by Qatar’s sovereign wealth fund, will be poured into speeding up production of the company’s high‑speed memory chips—key ingredients for the processors that power today’s AI applications. Positron’s flagship chip, dubbed Atlas, is already being fabricated in Arizona and the startup claims it can deliver performance on par with Nvidia’s flagship H100 GPU while consuming less than a third of the power. Rather than chasing the massive compute needs of model training, Positron is zeroing in on inference—the step where trained AI models are actually run in real‑world products and services. As businesses move from building ever‑larger models to deploying them at scale, demand for efficient, low‑power inference hardware is soaring. The fresh capital will help Positron scale its manufacturing, broaden its product lineup, and push its power‑saving technology into data centers and edge devices. If the company lives up to its promises, it could give Nvidia a serious challenger in a market that’s rapidly expanding beyond pure research labs into everyday applications.

Read more