Chips, clouds, data centers, training, inference, and the production AI stack. Reader coverage is updated continuously; Google Search eligibility is decided separately.
Stream archive
OpenAI presented details of its Jalapeno AI accelerator at the Hot Chips conference on August 25, according to The Register. The publication reports that a 128-chip configuration delivers 1.7 exaFLOPS and 27 TB of HBM, with OpenAI projecting higher inference throughput and lower latency than contemporary GPU systems. Volume production is reported for 2027.
Cisco said it is expanding its Secure AI Factory with NVIDIA by adding Supermicro's liquid- and air-cooled server systems to the offering. The company plans to begin offering the systems in October 2026, pairing them with Cisco networking, security and observability tooling for large enterprise, neocloud and sovereign-cloud AI deployments.
On August 25, 2026, AM Intelligence placed a binding order for about 9,000 NVIDIA Vera Rubin systems for its first AI factory in Hyderabad, Business Standard reports. Economic Times reports delivery is scheduled for the first quarter of 2027 and that the hardware will power an initial 30 MW phase. AMI has outlined more than $8 billion in capital expenditure for an initial 200 MW of compute capacity.
Apple introduced a new Mac mini with M6 and a new Mac Studio with M5 Ultra on August 25, 2026, expanding desktop options for local AI development and inference workloads. Apple's press release describes M6 as its first 2 nm chip and M5 Ultra as its first quad-die M-series architecture. CNBC reports that the Mac mini starts at $899, up from $799 for the prior model.
George Town Council approved Firmus Technologies' proposed 288-megawatt AI data centre at Long Reach in northern Tasmania on August 25, with six of seven councillors present voting in favour. ABC News reported that the approval covers the former Gunns pulp mill site in the Bell Bay Advanced Manufacturing Zone and includes conditions on air quality, noise, generators and emergency planning. People who made submissions can appeal within 14 days.
On August 21, Singapore provisionally allocated 50MW of new data-center capacity to Digital Realty under its DC-CFA2 exercise. The EDB and IMDA awarded the same amount to Equinix, Keppel Data Centres, and ST Telemedia Global Data Centres, for 200MW total; all four projects are planned for Jurong Island and carry green-energy, liquid-cooling, and efficiency commitments.
SCX.ai and DDN announced a partnership on August 24, 2026, to expand an Australian sovereign AI inferencing cloud aimed at enterprises, government agencies, and research institutions. The companies are combining SCX.ai's domestic SambaNova-based infrastructure with DDN's Infinia data platform. CFOtech reports that the announcement followed SCX.ai's AUD $40 million ASX initial public offering and that its first sovereign AI node operates in Sydney.
Intel detailed its next-generation Diamond Rapids Xeon server processor architecture at Hot Chips 2026 on August 24, presenting a design built on its 18A-P process with Foveros Direct 3D packaging and UCIe-S chiplet interconnects. Tom's Hardware reports configurations of up to 256 performance cores and 1.28 GB of last-level cache, with data-center availability targeted for 2027.
NVIDIA said Aug. 24 that its Groq 3 LPX inference system has entered full production alongside the Vera Rubin NVL72 platform. In a 100,000-token Gemma 4 31B benchmark, NVIDIA reported about 3,400 output tokens per second; the result applies to a named vendor configuration, not every deployment.
Xiaomi showed the AI Cube engineering prototype on Aug. 24, a mini PC that combines its Xring O3, O100, and D100 chips for local AI workloads. Reports based on the presentation describe a 120B-plus-3B model configuration, up to 150 W of sustained power, and a 1.22-TB/s near-memory accelerator; no consumer price or release date was announced.
At Computex in June, Intel disclosed additional details for its forthcoming Crescent Island AI inference accelerator, including a 350W air-cooled PCIe design and support for up to 480GB of LPDDR5X memory, according to Tom's Hardware and Data Center Dynamics. Intel had not disclosed raw throughput figures, leaving the capacity-focused design difficult to compare directly with HBM-based GPUs.
US datacenters used about 17 billion gallons of water in 2023, roughly triple their 2014 use, according to a Congressional Research Service estimate reported by The Register on August 24. The CRS found that the federal government does not systematically assess the industry's water consumption, while the reported figures predate much of the recent AI infrastructure construction surge.
IBM on Aug. 24 unveiled a next-generation processor design for future IBM Z and LinuxONE systems that can execute Arm and IBM instruction sets on the same cores. IBM states that the 2 nm design has 11 cores operating above 5.7 GHz, AI inference accelerators, and an on-chip data processing unit for I/O acceleration. VentureBeat reports each core can switch architectures in nanoseconds.
Some of Nvidia's largest customers have reportedly been told that prices for AI servers containing its chips could rise by more than 15% in many cases as memory costs climb. The reported increases would affect selected Vera Rubin and Grace Blackwell systems scheduled for delivery early next year, potentially raising data-center expansion costs.
Honor said its Lightning humanoid robot covered 100 meters in 9.32 seconds during a preparatory test for Beijing’s World Humanoid Robot Games on August 22. The time is faster than Usain Bolt’s 9.58-second men’s human world-record mark, but it was a robot test rather than an athletics result—an important distinction as humanoid locomotion demonstrations accelerate.
NVIDIA made a minority investment in Cloverleaf Infrastructure on August 21 as part of a partnership to develop U.S. sites for AI data centers. Cloverleaf said it will apply NVIDIA's DSX platform to weigh site, power, cooling, computing, and facility choices earlier in design; neither company disclosed the investment's financial terms.
Starcloud announced a $250 million Series A extension on August 21, 2026, at a $2.3 billion post-money valuation, with Manhattan West leading and NVIDIA participating. The round brings the company’s total funding to $450 million and will support manufacturing, engineering work with NVIDIA, and launch capacity for its orbital AI-compute plans.
Samsung Electronics raised prices for new orders of some advanced foundry services on August 19, with increases of up to 15% as AI-chip demand tightened capacity. Reuters said the reported increases covered Samsung's 4-nm SF4, 5-nm SF5, and 8-nm processes, while the SF4 line at Pyeongtaek had operated at full capacity since late 2025, citing people familiar with the matter.
NUS Medicine, DayOne and Cortical Labs announced a 20-unit CL1 biological-computing prototype at NUS in Singapore, following a 6 August 2026 showcase. The system pairs living human neurons with electrode-fitted silicon chips for joint research, but the sources reviewed do not provide workload benchmarks that would make it comparable with conventional AI infrastructure.
Waymo on August 20 disclosed technical details of the compute module in its sixth-generation autonomous driving system, including two custom ASICs built for onboard AI inference. According to Waymo's engineering post, the processors deliver more than 1,000 trillion calculations per second and are designed with redundancy so either chip can take over after a hardware failure. Bloomberg reports the ASIC is already in production in Waymo's latest robotaxi generation.
Anthropic posted Canadian roles on August 20, 2026, for a compute country lead and an Alberta-based community engagement manager, suggesting a potential expansion of data-centre capacity in Canada. BetaKit reports that Anthropic confirmed the postings, while the compute-lead advertisement assigns responsibility for bringing "gigawatts of compute online in Canada" and covers site selection, leasing, financing, and government relations.
CoreWeave announced a multi-year agreement with Hudson River Trading on August 20 for its AI cloud platform. According to CoreWeave, HRT will use NVIDIA Vera Rubin NVL72 infrastructure for AI-driven trading research and model development; TipRanks reports the agreement is worth multiple billions of dollars and expands an existing relationship.
Alibaba reported a 75% year-over-year decline in net income for the quarter ended June 30 as capital expenditure on AI infrastructure rose 75% to 67.68 billion yuan, Reuters and CNBC reported on August 20. Revenue rose 9% to 268.95 billion yuan, while AI Cloud and Compute Services revenue increased 45% to about 48.4 billion yuan. Alibaba's adjusted earnings per ADS missed the LSEG consensus estimate, according to Reuters.
Northrop Grumman introduced Raid Hunter on August 11, a short-range, gun-based air-defense system built around guided 50mm ammunition, Bushmaster Chain Gun technology and a networked controller. The company says it is designed for high-volume mixed aerial raids, including drones and cruise missiles. Published material does not provide range, interceptor cost or independent test results.
On August 19, Marvell Technology agreed to help develop Google's custom AI chips and granted Google a warrant to buy up to 58.97 million Marvell shares. Reuters calculated that the warrant could be worth about $12.18 billion if fully exercised, with most of it contingent on Google meeting purchasing targets through Marvell's fiscal 2033.
DataVita has secured a GBP 300 million debt facility to expand its DV1 data center and construct DV3 in North Lanarkshire's AI Growth Zone. The UK's National Wealth Fund has provided a GBP 202 million guarantee, while Data Centre Solutions reports that both facilities' capacity is contracted to AI cloud provider CoreWeave under a 15-year lease.
Alberta's first AI data centre town hall drew a packed crowd in Lacombe on Aug. 19, where residents pressed Technology and Innovation Minister Nate Glubish on water use, electricity, setbacks and oversight. The province's engagement effort now faces a practical test: whether it can answer those concerns with specific, enforceable details rather than broad assurances.
Mistral says its indexed Google Drive and SharePoint Knowledge Connectors will be removed at the end of August 2026 in favor of MCP Connectors. Administrators must add the replacements and users must reconnect their accounts, while Mistral's public documentation leaves the replacement retrieval architecture and the old index's cutoff-time deletion schedule unresolved.
AMD reported on August 18 that its rack-scale AI systems delivered a 4x energy-efficiency improvement from its 2024 baseline through mid-2026. The company had previously projected a 3x improvement for this point on its path to a 20x rack-scale efficiency target by 2030, according to Tom's Hardware. AMD's calculation combines measured product data with modeled estimates where final performance data were unavailable, StorageReview reports.
GitHub suffered a major outage on August 17 that disrupted web and API traffic, Actions, pull requests, authentication services and Copilot. GitHub's status record shows the incident ran from 13:40 to 21:15 UTC; one update reported roughly 20% errors for web and API traffic and about 50% errors for archive and raw-content downloads.
