The global technology sector was abuzz this week as NVIDIA opened its highly anticipated GTC 2026 conference in San Jose, California, drawing the attention of investors, engineers, and industry leaders from around the world. From March 16 to 19, the event became the epicenter of breakthroughs in artificial intelligence (AI) hardware, with a particular focus on advanced memory technologies and the relentless expansion of AI-driven applications.
At the heart of the excitement was NVIDIA CEO Jensen Huang’s keynote, where he delivered a bold forecast that set the tone for the entire conference. According to Global Economic, Huang projected that by 2027, combined orders for NVIDIA’s Blackwell and Vera Rubin AI platforms would reach at least $1 trillion—a dramatic leap from the company’s previous $500 billion target and a figure that sent ripples through the industry. Huang’s optimism wasn’t just about numbers. He emphasized the transformative impact of autonomous agent applications, asserting, “Tokens are the new raw material,” as the exponential growth in token processing continues to fuel demand for ever-faster, more efficient computing.
Investors responded with enthusiasm. On March 17, the US stock market posted robust gains, led by big tech companies riding the wave of optimism from the GTC conference. NVIDIA’s stock rose 1.65%, as reported by SBS Biz, with the broader S&P 500 and Nasdaq indices both jumping more than 1%. Apple, too, saw its stock climb 1.08% on news of a new AirPods Max release, while Meta surged over 2% after announcing plans to cut up to 20% of its workforce in a bid to offset the heavy costs of AI infrastructure investments. The rally extended to other semiconductor and AI-related stocks, with companies like Micron and Broadcom benefiting from the sector’s buoyancy.
But the real action was on the show floor, where the world’s leading memory makers unveiled their latest innovations. Samsung Electronics made headlines by publicly displaying its 7th generation high-bandwidth memory (HBM4E) for the first time. As detailed by The Elec, the HBM4E chip boasts a blistering 16 gigabits per second (Gbps) per pin and 4 terabytes per second (TB/s) of bandwidth. Samsung’s exhibition also introduced a cutting-edge hybrid copper bonding (HCB) technology, which enables higher-layer stacking for memory modules while improving electrical performance and thermal resistance by 20% compared to older methods. This leap in packaging is crucial for supporting the massive data flows required by next-generation AI systems.
Samsung’s booth at GTC 2026 was a showcase of synergy with NVIDIA. The company presented not just HBM4E, but also HBM4 for GPUs, SOCAMM2 server memory modules based on LPDDR5X, and the PM1763 PCIe Gen6 enterprise SSD—all tailored for NVIDIA’s Vera Rubin platform. A live demonstration of SCADA technology allowed GPUs to access SSD storage directly, bypassing traditional CPU and DRAM bottlenecks to maximize input/output performance. The display also featured next-generation memory architectures like GDDR7, LPDDR6, and PM9E1 SSDs, all designed to improve AI inference capabilities and energy efficiency. Samsung AI Center head Song Yong-ho was scheduled to deliver a keynote on March 17, outlining the company’s vision for supporting NVIDIA’s next-gen AI systems with comprehensive memory solutions.
SK Hynix, another titan of the memory world, also took center stage at GTC 2026. From March 16 to 19, the company’s “Spotlight on AI Memory” exhibition offered hands-on experiences with their HBM4, HBM3E, and SOCAMM2 memory modules, all integrated into NVIDIA AI platforms. According to Shin-A Ilbo, the display included interactive zones where visitors could explore memory technology through games and simulations, such as the HBM 16-layer stacking challenge designed to explain the intricacies of through-silicon via (TSV) processes and advanced packaging. SK Hynix also showcased its liquid-cooled eSSD and LPDDR5X memory, which power NVIDIA’s DGX Spark AI supercomputer.
SK Hynix executives—including Chairman Chey Tae-won and CEO Kwak No-jung—used the conference as an opportunity to meet with global tech leaders, discussing the future of AI manufacturing and the critical role of memory in high-performance AI. Technical sessions delved into how memory technology is becoming a defining factor in AI infrastructure, with SK Hynix vowing to leverage its expertise from data centers to edge devices in partnership with global players like NVIDIA.
Beyond the exhibition halls, the conference’s impact was felt across global markets. The surge in AI optimism helped push Bitcoin to $74,000, its highest level since February, as investors sought alternatives amid ongoing geopolitical uncertainties. Meanwhile, international oil prices, which had spiked on fears of conflict in the Middle East, eased after US Treasury Secretary Bessent reassured markets that oil shipments through the Hormuz Strait would continue unimpeded. The International Energy Agency also pledged to release strategic reserves if needed, calming fears of supply disruptions and helping oil prices retreat by more than 5% for WTI and over 3% for Brent, according to SBS Biz.
Back in the tech sector, NVIDIA’s ambitions extended far beyond chips and memory. Huang unveiled the company’s next-generation Vera Rubin platform, which offers double the power efficiency of its predecessor and is already running on Microsoft Azure’s cloud infrastructure. He also introduced Groq, a language processing unit developed by a startup founded by former Google engineers. When paired with Vera Rubin, Groq can boost token processing performance per watt by up to 2.5 times—a significant leap for AI inference workloads. NVIDIA’s roadmap stretches even further, with plans to deploy fully autonomous robo-taxi fleets in partnership with Uber, starting in Los Angeles and San Francisco before expanding to Europe, Asia, and Australia by 2027. Major automakers like Geely, Nissan, and Isuzu have adopted NVIDIA’s Drive Hyperion level 4 autonomous driving platform, while Hyundai and Kia are collaborating on next-generation autonomous technologies.
NVIDIA’s aggressive push into AI hardware, memory, and autonomous vehicles signals a new era of collaboration and competition. The company’s success, however, hinges on the stability of its supply chain and the ability of partners like Samsung and SK Hynix to deliver ever-more advanced memory solutions. As one semiconductor analyst told Global Economic, “If the $1 trillion order number is to be realized, supply chain stability and advanced packaging technology are prerequisites.”
The GTC 2026 conference made it clear: the age of AI is accelerating, and the world’s leading tech firms are racing to build the infrastructure needed to power it. From innovative memory chips to autonomous vehicles and beyond, the future looks both fast and fiercely competitive.