GoKawiil
Tech News
Qualcomm and Huawei have entered into a multiyear agreement that includes a cross-license covering 5G, AI, compute, and networking technologies. Additionally, Qualcomm will acquire a selection of Huawei's US patents related to these fields. The deal's financial terms and specific patent details remain undisclosed, and regulatory approval is pending.
Cerebras Systems co-founder and CEO Andrew Feldman is scheduled to appear at TechCrunch Disrupt 2026 for a session titled 'Can AI Keep Scaling?', addressing rising demand for compute, energy and infrastructure. Cerebras builds wafer-scale AI chips as an alternative to conventional chip architectures, raised $5.5 billion in its May IPO, signed a multiyear deal with OpenAI for 750 megawatts of systems through 2026-2028, and launched its CS-4 hardware in August.
A Show HN post introduces a tool that automatically generates, tests, benchmarks and refines CUDA GPU kernels from a workload description. It uses an LLM agent via LangGraph to propose code and launch-configuration changes, a C++/NVRTC harness to compile and run candidates, and Python to validate outputs against NumPy references, keeping only the fastest kernel that passes all correctness checks. The project was developed on Windows with an RTX 3060 Laptop GPU and requires Python 3.12+, CMake, a C++17 compiler and an OpenAI API key.
Raspberry Pi has confirmed its firmware locks boards to their factory RAM configuration, preventing users from swapping in replacement memory chips even to fix failed components. The restriction surfaced after a user couldn't get a Compute Module 5 to recognize a new 4GB chip, with company engineer PhilE explaining the firmware checks additional memory attributes beyond just capacity. Raspberry Pi says the policy exists to stop resellers from installing unauthorized higher-capacity chips and passing modified boards off as factory models.
British AI cloud provider Nscale, spun out of Arkon Energy two years ago, disclosed over $103 billion in contracts ahead of its NYSE listing, but roughly 85% of that value comes from just two customers: Microsoft and Anthropic. The Anthropic deal remains conditional on Nscale securing financing and hitting milestones the company itself calls stringent, with Anthropic free to cancel if targets are missed. Nscale is targeting a $35 billion valuation and hopes to raise $3 billion, despite posting a $1.02 billion net loss on $140.6 million in revenue for the first half of the year.
Merriam-Webster has officially added roughly 1,400 new words to its dictionary, including AI-era terms like AGI, the noun form of 'compute,' and 'uncanny valley.' The additions reflect how deeply artificial intelligence terminology has entered everyday language.
A financial presentation seen by the Financial Times shows OpenAI expects to spend $278 billion more than it earns between 2026 and 2030, driven mainly by $856 billion in compute and infrastructure costs. Even as revenue is projected to jump from $36 billion in 2026 to $350 billion in 2030, spending will still outstrip income by a wide margin. The figures were reportedly shared with investors ahead of OpenAI's anticipated IPO.
YouTuber EastMakes repurposed a 15-year-old HP ProLiant N40L micro-server by designing a custom motherboard around a Raspberry Pi Compute Module 5. Starting from the open-source Raspberry Pi I/O board schematics, he reshaped the PCB to fit the server's original chassis, added an NVMe slot, USB 3.0 ports, and Ethernet, while trimming down to a single HDMI output.
Anthropic and OpenAI are pursuing smaller data center deals in the 20-30 megawatt range, according to sources familiar with the discussions, even as both companies have already signed massive multi-hundred-megawatt and gigawatt infrastructure agreements. Anthropic has explored such deals in the UK and Nordics, while OpenAI has looked at similar opportunities in the Nordics, with talks also reportedly underway for smaller U.S. deployments. OpenAI confirmed it is building a 'diversified compute portfolio' but declined to discuss specific commercial talks; Anthropic did not comment.
Apple is reportedly building custom AI servers around future M8 Ultra chips, planned in two- and four-chip configurations, and is evaluating Nvidia's NVLink Fusion for high-speed interconnects between them. The project, said to have started roughly a year ago under former hardware chief John Ternus, aims to replace Apple's existing in-house connectivity tech, which is reportedly too slow and expensive for large-scale deployment.
Apple published a Security Research blog post explaining how its new Reference Image camera feature, launching on the iPhone 18 Pro, verifies that photos are authentic from the moment they're captured. The system creates a secure digital negative with signed metadata and timestamps inside Apple's Private Cloud Compute environment, and uses a sensor-level trust chain with revocation lists to flag compromised hardware or fraudulent images.
Prediction market platform Kalshi reportedly took down a tool tracking AI compute prices after the U.S. Commerce Department raised national security objections. Details on the exact nature of the concern have not been fully disclosed, but the removal followed direct government pushback.