TensorWave's 457 MW Is About 8% of the OpenAI-AMD Plan. The Test Is the Rest.
OpenAI's first named AMD neocloud commits 457 MW for six years, about 8% of the 6 GW program. The announcement is a milestone; the delivery schedule is the real test.
Explore the foundational technologies and hardware driving the artificial intelligence boom. From next-generation data centers and custom silicon to advanced networking and energy solutions, discover the deep tech investments building the physical and digital layers that power the AI era.
OpenAI's first named AMD neocloud commits 457 MW for six years, about 8% of the 6 GW program. The announcement is a milestone; the delivery schedule is the real test.
MountAIn raised €7.3M to run full computer vision applications on microcontrollers costing under $10. Teledyne, Roboflow and Alif are on board; revenue, pricing and shipping products are not disclosed.
Torus launched Foundry, a prefabricated 1.2 MW inference block with 576 B300 GPUs, and raised $500M led by Magnetar. The pitch: bring chips to power that is already on the grid, in 12 to 16 weeks instead of years in the interconnection queue.
Daisytuner's €6M seed and Daisyflow launch rest on an 18% inference lead over vLLM on AMD Instinct. The defensible asset is a self-learning compiler and the benchmark database it feeds — if the number survives a second chip.
Nvidia trimmed Vera Rubin's HBM4 bandwidth from 22 to 19.2 TB/s and NVLink 6 from 3.6 to 3.0 TB/s without an announcement. The eight-GPU NVL8 lost 26% of its memory bandwidth. Here is why bandwidth, not FLOPS, now sets AI infrastructure costs.
A four-person Cambridge team raised a $30M seed from Benchmark to rebuild the CPU around interaction nets. The design is novel; the benchmarks, tape-out date and software ecosystem are still missing.
Mount Thor raised $45 million across Seed and Series A to rent macOS and Apple silicon to computer-use agents. The product is really an operating-environment cloud: bare-metal Macs, two-slot VMs, 24-hour billing. Inference on Apple silicon is the funding thesis; the benchmarks are still missing.
Fractile raised $720 million at a reported $7 billion valuation, up from about $1 billion in May, before shipping a production chip. Britain's sovereign funds are buying an option on domestic AI hardware; the 2027 ship date is where the story gets tested.