ROBOTERA hits 1,000-unit quarterly run in live logistics ops
Beijing robotics firm ROBOTERA raised $200M led by SF Group, now running 1,000+ units across 10+ logistics centers with 300% growth, marking the first humanoid robotics company to scale beyond pilots into volume commercial operations.
XBOW closes $35M Series C extension, customers now investors
Autonomous penetration-testing startup XBOW raised $35M from six strategic investors who are also customers, pushing total funding above $270M and signaling a shift in how enterprises adopt AI-driven security.
TSMC commits to five 2nm fabs in single city, doubling its expansion pace
TSMC is building five advanced-chip factories simultaneously in Kaohsiung through 2027, with every wafer already sold through 2026 and 70% annual capacity growth planned, cementing Taiwan's control of cutting-edge semiconductor production.
Tenstorrent ships Galaxy Blackhole, taking on Nvidia in inference
Tenstorrent began volume production of Galaxy Blackhole, a $110,000 AI inference server that claims 23 PFLOPS performance and 350+ tokens per second, challenging Nvidia's GPU dominance with a clean-sheet hardware design.
Lumai's optical chip runs LLMs faster on 90% less power
Oxford startup Lumai launched Iris Nova, the first optical inference server running billion-parameter models in real time, claiming up to 90% lower energy than silicon, directly challenging GPU economics as data centers face hard power ceilings.
Pentagon deploys eight frontier AI firms to classified networks, cuts Anthropic out
The DoD announced May 1 deals with OpenAI, Google, Microsoft, AWS, NVIDIA, SpaceX, Oracle, and Reflection AI for classified network deployment, formally excluding Anthropic after the company refused unrestricted military use of Claude.
Tenstorrent's $110K AI Server Challenges Nvidia's Inference Dominance
Tenstorrent launches Galaxy Blackhole inference servers at 3-5x cheaper per node than Nvidia DGX, with 16 units already deployed at Equinix and performance claims that undercut the GPU+LPU disaggregation trend.
QCi launches photonic AI chip as PCIe card, targeting GPU's edge dominance
Quantum Computing Inc. shipped NeuraWave, a room-temperature photonic inference accelerator in PCIe form, positioning photonics as a power-efficient alternative to GPUs at the edge.
IonQ and Florida LambdaRail Deploy First U.S. Statewide Quantum-Safe Network
IonQ signed a Master Service Agreement with Florida LambdaRail on April 27 to build a 100-mile quantum-safe corridor—the first statewide quantum network deployment in the U.S., connecting universities and defense-adjacent institutions across three counties.
Google Splits TPU in Two: Training and Inference Arms
Google bifurcated its eighth-generation TPU into specialized training and inference chips on April 22, 2026, marking the first architectural split in a decade and directly challenging Nvidia's grip on AI infrastructure scaling.
SK Hynix mass-produces 192GB SOCAMM2 for AI servers
SK Hynix began mass production of 192GB SOCAMM2 memory modules on April 20, 2026, targeting NVIDIA Vera Rubin AI servers — but trails Micron's 256GB samples and faces Samsung's warpage breakthrough.
Tesla AI5 chip tapes out, first stop is Optimus, not cars
Tesla taped out its AI5 self-driving chip on April 15, 2026, with 8x the compute of AI4, but it's deploying first to humanoid robots and data centers, not the vehicle fleet—a signal about where the real AI work actually is.
Chad Rigetti's Sygaldry raises $139M to shrink AI's power appetite
Chad Rigetti, the physicist behind Rigetti Computing, has launched Sygaldry Technologies with $139M in funding to build quantum-accelerated servers that reduce AI's energy consumption inside data centers rather than replace classical chips.
Meta and Broadcom Lock in 1GW Custom AI Silicon Through 2029
Meta and Broadcom announced a multi-year partnership on custom MTIA chips through 2029, committing to deploy 1 gigawatt initially with plans to scale to multiple gigawatts by 2027. The chips will be the industry's first 2nm AI accelerators, shifting the inference market away from Nvidia.
IonQ Connects Two Quantum Computers Via Photons, Stock Surges 18.3%
IonQ demonstrated the first photonic interconnection between two independent trapped-ion quantum systems on April 14, validating modular quantum networking and triggering an 18.3% stock surge amid DARPA contract awards.
SambaNova and Intel Ship Production Agentic AI Chip Stack
SambaNova and Intel announced a signed, production-ready heterogeneous inference architecture combining GPUs, Xeon 6 CPUs, and RDUs for agentic AI, deploying in standard data centers by H2 2026.
D-Robotics raises $270M Series B as chip layer consolidates
D-Robotics closed a $150M Series B2 round in 20 days, bringing total Series B to $270M. The speed and investor composition signal capital is securing foundational robotics silicon before the market consolidates.
Anvil Robotics Closes $5.5M Seed as Physical AI Hits the Assembly Line
Anvil Robotics, an eight-month-old startup selling modular robot dev-kits, closed a $5.5M seed round with Nvidia's GEAR lab as a paying customer and seven-figure revenue already in the bank—signaling that physical AI infrastructure, not just models, is now the competitive moat.
CavilinQ Raises $8.8M to Build the Missing Layer in Quantum Scaling
CavilinQ closed an $8.8M seed round on April 2, 2026, to develop photonic interconnects linking quantum processors into modular clusters—solving the bottleneck that has prevented quantum computing from scaling beyond isolated machines.
Normal Computing Raises $50M Led by Samsung Catalyst for Thermodynamic AI Chip
Normal Computing closed a $50M Series B led by Samsung Catalyst Fund on March 25, 2026, backing the world's first thermodynamic computing chip and an AI-native EDA platform already deployed at more than half of the top 10 semiconductor firms by revenue.