OpenAI is running out of computing capacity after its new image generator went viral, CEO Sam Altman announced. Altman warned that upcoming product launches will be delayed and services may break or slow down. To manage the crush, OpenAI pushed back access for free ChatGPT users and paused Sora video generation for new signups.
Nvidia is reportedly in advanced talks to acquire server rental startup Lepton AI for several hundred million dollars. The two-year-old startup rents out compute powered by Nvidia’s own chips, putting the chip giant on track to compete directly in the server rental market. Why it matters: Nvidia is no longer just selling picks and shovels—it is steadily moving into infrastructure operations.
Groq just partnered with PlayAI to deploy its Dialog text-to-speech model on GroqCloud, featuring English and native Arabic support. Using PlayAI's adaptive contextualizer alongside Groq's LPUs, the model responds with conversation-aware emotion at up to 140 characters per second—roughly 10x faster than real time. Why it matters: developers can now build voice agents with near-zero latency and convincing prosody directly on a single inference platform.
Boston Dynamics just expanded its partnership with NVIDIA to bring the Isaac GR00T framework and Jetson Thor compute to its electric Atlas humanoid. The setup lets Atlas run heavy multimodal models alongside its native whole-body controllers, while training complex locomotion in Isaac Lab. Why it matters: The world's most agile robot just picked up the most popular AI compute stack.
Nvidia and Elon Musk's xAI just officially joined the AI Infrastructure Partnership (AIP), a coalition led by Microsoft, BlackRock, GIP, and Abu Dhabi's MGX. The alliance aims to mobilize $30 billion in private capital to build massive data centers and energy infrastructure, primarily in the US. Why it matters: Surging compute and power demands are pushing fierce tech rivals into unprecedented infrastructure alliances.
Nvidia's biggest event of the year is underway, with CEO Jensen Huang scheduled to deliver his keynote on Tuesday at 10 a.m. PT. Viewers can stream the presentation for free on Nvidia's website or YouTube without registering, with updates expected on Blackwell Ultra and the next-gen Rubin architecture.
Meta just kicked off small-scale testing of its first custom AI training silicon, aiming to rein in its gargantuan infrastructure spending and ease its reliance on Nvidia. The MTIA chip is designed to be more power-efficient than general-purpose GPUs, and Meta hopes to deploy it for production training by 2026.
OpenAI just inked a massive five-year deal worth $11.9 billion with CoreWeave for AI data centers and infrastructure services. As part of the agreement, OpenAI will also receive around $350 million in equity tied to CoreWeave's upcoming IPO. The commitment secures vast GPU capacity for OpenAI while supercharging its partner's public debut.
xAI just purchased a 1-million-square-foot property in Southwest Memphis to expand its massive Colossus supercomputer cluster. The facility complements an existing site that features an $80 million wastewater recycling plant and the world's largest Tesla Megapack deployment to reduce strain on local grids. Why it matters: Elon Musk's race for raw computing scale isn't slowing down, and xAI is locking down the physical footprint needed to power next-gen frontier models.
Apple just unveiled the new MacBook Air powered by the M4 chip, now starting at $999 with 16GB of unified memory by default. The upgraded Neural Engine accelerates on-device AI tasks up to 3x faster than the M1 model, built specifically to power Apple Intelligence. Pre-orders are live today, with shipments arriving March 12.
Apple just introduced the updated Mac Studio powered by M4 Max and the new M3 Ultra chips. The compact desktop supports up to 512GB of unified memory and Thunderbolt 5, allowing developers to run LLMs with over 600 billion parameters entirely in memory. It basically turns a desktop Mac into a serious local AI powerhouse.
Apple just refreshed the iPad Air with its M3 chip, making the midrange tablet fully ready for Apple Intelligence. It packs a 60% faster Neural Engine over the M1 model to speed up on-device AI tasks, landing in 11-inch and 13-inch sizes at the same $599 starting price alongside a cheaper Magic Keyboard.
TSMC just committed at least $100 billion to expand its US semiconductor manufacturing footprint, funding two new factories in Phoenix alongside its existing three. CEO C.C. Wei confirmed the new sites will build advanced AI chips. It is a massive bet to onshore supply chains ahead of planned semiconductor tariffs.
California state Senator Scott Wiener just introduced SB 53, a fresh AI bill reviving the least controversial parts of the vetoed SB 1047. It shields whistleblowers at frontier AI labs from retaliation if they report catastrophic risks, while proposing a public cloud computing cluster called CalCompute for researchers.
AWS just unveiled Ocelot, a prototype quantum computing chip that cuts error correction costs by up to 90%. Built around 'cat qubits' that naturally suppress environmental noise, the microchip integrates correction hardware from the ground up. AWS researchers say the architecture could speed up the timeline to practical fault-tolerant quantum computing by up to five years.
Nvidia CEO Jensen Huang just pushed back on fears of slowing AI demand, revealing that next-gen reasoning models require 100 times more computation than when ChatGPT debuted. Addressing market concerns after DeepSeek's rapid rise, Huang argued that step-by-step reasoning actually multiplies the need for chips. Why it matters: Despite worries that efficiency gains would cut spending, Nvidia insists smarter AI only makes the appetite for hardware grow faster.
Qualcomm just unveiled Dragonwing, a dedicated brand targeting industrial IoT, networking, and enterprise infrastructure. The lineup includes the IQ series packing 100 INT8 TOPS for rugged environments and the Q series optimized for edge AI and Wi-Fi 7. Qualcomm is keeping Snapdragon for consumer gadgets while sending Dragonwing to challenge Nvidia and AMD in factories.
Chinese tech giants including Tencent, Alibaba, and ByteDance are reportedly ramping up orders for Nvidia H20 chips to keep up with DeepSeek's low-cost models. While investors initially feared efficient models would gut hardware demand, heavy inference usage across sectors like healthcare and education is driving massive server orders instead.
Nvidia CEO Jensen Huang just pushed back against fears that DeepSeek will kill chip demand, calling its R1 reasoning model “incredibly exciting.” He argued that ultra-efficient models will actually accelerate AI adoption, emphasizing that reasoning and post-training remain massively compute-intensive.
OpenAI is forecasting a major shift in its computing power over the next five years, expecting to get three-quarters of its data center capacity by 2030 from Stargate, a project heavily financed by SoftBank. The transition away from Microsoft won't happen overnight, but OpenAI projects burning $20 billion in cash in 2027 as inference costs eventually overtake training.
