The shortage of GPUs for training large language and vision models has sparked competition among companies vying for backup capacity and computing power.
NVIDIA has always been at the forefront of the generative AI wave and is one of its biggest beneficiaries, thanks to its GPUs. The chip giant predicts its performance for the next quarter to reach an astonishing $11 billion, 50% higher than Wall Street's forecast.
But this is a double-edged sword for chip manufacturers because demand far exceeds supply. "It's like toilet paper during a pandemic... it's about who you know because of the shortage," said Sharon Zhou, co-founder and CEO of Lamini, as quoted by The Wall Street Journal.
The generative AI revolution and the increase in capacity from big data and server operators are causing a shortage of GPUs. Companies are scrambling to acquire GPUs for training large language and vision models, as GPUs are well-suited for parallel processing tasks like matrix multiplication. The shortage of advanced chips for generative AI systems has led to competition for backup capacity and computing power.
As a result, NVIDIA's GPU inventory plummeted 10.6% in March, the first decline since the fourth quarter of 2019. Elon Musk is said to have acquired most of Oracle's backup server space for X.AI, a competitor of OpenAI (Elon Musk's newly established AI company), making it even more difficult for startups to find server resources. He once said, "Getting a GPU now is much harder than getting drugs."
Meanwhile, Sam Altman believes that the era of scaling up larger AI models has come to an end due to the cost limitations and diminishing returns of powerful GPUs for training and operating large language models. However, NVIDIA is seeking to salvage the situation and keep the stack flowing. The market leader, with a recent market value of a trillion dollars, is focused on building an AI-driven future that will bring it the next trillion dollars. CEO Jensen Huang outlined the company's plans at the Computex conference. Leveraging its expertise in generative AI, NVIDIA aims to diversify its AI portfolio and gain capital in the market.
CEO Huang announced that NVIDIA plans to rent out its powerful supercomputers to enterprises. Through the DGX Cloud rental service, developers can access tens of thousands of NVIDIA chips, including flagship products like the A100 or H100 chips, to accelerate the development of AI. The price for renting eight chips through this service is $37,000 per month, aiming to meet the growing demand in the AI industry. NVIDIA is collaborating with major companies such as Oracle, Microsoft, and Alphabet to provide supercomputer services, further expanding access to its powerful technology.
To make AI products more cost-effective, NVIDIA introduced new chips and software solutions at the conference. They also launched the AI Foundations, a service aimed at helping businesses train customized AI models. In addition, NVIDIA introduced technologies that accelerate semiconductor design and manufacturing processes, reducing computation time from weeks to overnight. Collaborations with industry giants such as AT&T, TSMC, and ASML Holding are underway to bring these advancements to the market.
Huang emphasized that NVIDIA's entire data center product line, including the H100, Grace CPU, Grace Hopper superchip, NVLink, Quantum 400 InfiniBand, and BlueField-3 DPU, is now in production to meet the growing demand. The company announced that the GeForce RTX 4060 Ti GPU for gamers, HGX H100 GPU server, and GH200 Grace Hopper superchip are also in full production. These developments highlight NVIDIA's commitment to advancing AI computing and providing high-performance solutions for a range of applications.
In particular, the GH200 Grace Hopper superchip, which combines the Arm-based NVIDIA Grace CPU and Hopper GPU architecture, is now in full production. GH200 offers high bandwidth and computing power, making it suitable for complex AI and high-performance computing workloads. Global mega-enterprises, supercomputing centers, and system manufacturers such as Cisco, Dell, and Lenovo will have access to GH200-based systems. NVIDIA's software stack, including NVIDIA AI, Omniverse platform, and RTX technology, will be supported on these accelerated systems and is expected to be launched later this year.
Nevertheless, there is no competition regardless of the current situation. NVIDIA holds an 88% share of the global GPU market, with competitors AMD and Intel occupying the remaining 12%. However, other companies are entering the AI accelerator market, including AMD's AMD Instinct AI accelerator. Although AMD's consumer GPUs are not as suitable for AI applications as NVIDIA's GPUs due to the lack of an equivalent API like CUDA, AMD has developed the ROCm open software platform for machine learning. However, CUDA is more mature and integrated, and superior in compatibility with AI tools like TensorFlow and PyTorch.
To reduce dependence on NVIDIA, major tech giants like Google and Amazon have developed their own custom chips for AI workloads. For example, AWS has introduced Inferentia for inference tasks, while Google has developed Tensor Processing Units (TPUs) specifically for TensorFlow. Despite the emergence of these custom chips, NVIDIA has long been a major player in the field of AI, even before using GPUs to train AI algorithms.
Source: https://analyticsindiamag.com/nvidias-game-changing-fix-for-the-gpu-crisis/
Alphabet Q1 paid subscribers reached 350M, fueled by YouTube & Google One. YouTube ad revenue slightly missed at $9.88B amid Premium growth. Total Q1 revenue hit $109.9B, beating forecasts.
Anthropic releases Claude connectors for Adobe, Blender, and Ableton, enabling AI workflow automation and app control. It also sponsors the Blender Development Foundation.
Apple introduces an App Store model splitting 12-month plans into monthly installments. On iOS 26.4+, it lowers upfront costs without reducing total price, adding billing tracking.
Google launches Aadhaar digital IDs in Wallet for India, enabling secure QR/NFC verification. A phased partner rollout targets key sectors, while passport-based credentials expand globally.
YouTube tests "Ask YouTube," an AI tool blending text and video for complex queries. Initially for US Premium users, it supports interactive follow-ups, expanding Google's AI search.
Google grants US DoD classified AI access, joining OpenAI and xAI in defense deals amid Anthropic's dispute. Contractual safeguards against surveillance and lethal weapons remain unenforceable.
Mistral AI launches Workflows, a Temporal-powered orchestration platform for secure, Python-driven enterprise AI automation featuring human-in-the-loop approvals.