Topic
inference
26 stories related to this topic, newest first.
TechCrunchai2 hrs ago
ZML Releases LLM Inference Server Compatible With Multiple Chip Types
French startup ZML launched ZML/LLMD, a free inference server that runs open-source models on Nvidia, AMD, Google TPU, Apple Metal and Intel Arc hardware. The product is not open source. TechCrunch reported the details.
airedale.futurecdn.nettechnology14 hrs ago
DeepSeek Develops Inference AI Chip to Reduce Nvidia Reliance
China's DeepSeek has spent the past year building its own inference-focused AI chip and hiring design engineers while holding talks with foundries. Nvidia shares slipped 1.6 percent in premarket trading after the reports surfaced. Huawei already holds half of China's $50 billion…
SE
Rapplerai22 hrs ago
Chinese Startup DeepSeek Developing Custom AI Inference Chip
DeepSeek is building an inference-focused chip to reduce reliance on Nvidia and Huawei hardware. The effort began about a year ago and remains at an early stage.
forbes.comtechnology6 days ago
Confidential Computing Protects Data During AI Processing
Confidential computing uses hardware-based trusted execution environments to secure data while models perform calculations. The approach addresses gaps in conventional encryption that leave information exposed in memory during training and inference.
wccftech.comai8 days ago
Anthropic Claude Models Now Run on Nvidia GB300 Blackwell Ultra in Azure
Anthropic's Claude models are now available on Nvidia's GB300 Blackwell Ultra platform hosted in Microsoft Azure. The integration expands access to advanced AI inference hardware for enterprise users.
DE
FI
servethehome.comai10 days ago
OpenAI Develops Custom Jalapeño Inference Chip With Broadcom
OpenAI unveiled its first custom inference chip, Jalapeño, built in partnership with Broadcom. The development was discussed in a TechCrunch Equity podcast episode released June 26, 2026.
Semaforai11 days ago
OpenAI, Google, Apple and SpaceX Develop Custom AI Chips
OpenAI announced plans for Jalapeño, a custom inference chip built with Broadcom. The move joins similar efforts by Google, Apple and SpaceX to reduce reliance on a single supplier.
Fox Newstechnology12 days ago
OpenAI unveils custom inference chip developed with Broadcom
OpenAI introduced its first custom-built inference chip on June 24. The chip, developed with Broadcom and known internally as Jalapeño, is part of broader U.S.-China competition over computing infrastructure.
TechCrunchai13 days ago
OpenAI Unveils Custom AI Inference Chip Built With Broadcom
OpenAI introduced its first custom inference processor, developed with Broadcom and named Jalapeño. The chip targets lower operating costs for running AI models and is still undergoing testing.
servethehome.comai13 days ago
OpenAI and Broadcom unveil Jalapeño chip for ChatGPT inference, with 2026 deployment planned
OpenAI and Broadcom introduced the Jalapeño chip Wednesday as the company's first custom AI silicon. The chips will be manufactured by Broadcom and deployed for inference workloads in ChatGPT and other services starting late 2026.
Cnbctechnology28 days ago
D-Matrix Begins Shipping SRAM-Based Corsair Inference Chip for Small Models, Claims Speed and Efficiency Gains vs Standalone Nvidia GPUs
D-Matrix begins customer shipments of its Corsair inference chip in June. The startup says the chip runs small workloads 10 times faster and uses five times less energy than a standalone Nvidia GPU.
technology30 days ago
Intel CEO outlines CPU focus as AI demand shifts
Intel's new leadership has cut 34 percent of staff and secured an $8.9 billion U.S. government investment. The company now sees rising demand for central processing units to run AI inference tasks.
yahoo.combusiness39 days ago
Delaware Judge Applies Adverse Inference in WWE Shareholder Merger Lawsuit Over Signal App Settings
Judge J. Travis Laster ruled Tuesday that WWE executives destroyed evidence by using Signal's auto-delete feature. The court will assume five statements are true unless defendants rebut them.
rediff.compolitics45 days ago
Judge Orders Adverse Inference Against CoreCivic in ICE Detainee Death Case
A federal judge sanctioned CoreCivic for destroying video evidence in a wrongful death lawsuit filed by the family of an ICE detainee. The company later reached an undisclosed settlement before trial.
France 24ai47 days ago
AI Inference Growth Spurs Interest in Smaller Data Centers
Demand for running AI models is shifting focus from training to inference. Startups and cities are exploring compact data centers to manage costs and local impacts.
FR
Forbesfinance47 days ago
AI Inference Costs Can Rise Sharply With Production Query Patterns
Production AI systems often see costs increase when traffic moves from narrow pilot patterns to wide, spiky distributions. A small share of complex queries can drive most latency and expense. Teams that model cost by query class rather than volume can preserve product options.
New York Postai54 days ago
Cerebras Shares Rise 89 Percent in Nasdaq Debut
Cerebras Systems opened at $350 per share on its first day of trading, above the $185 IPO price that raised $5.55 billion. The Sunnyvale, California-based chip designer reached a fully diluted valuation of $106.75 billion. The company, founded in 2015, designs wafer-scale process…
Substrate placeholder — needs reviewai76 days ago
Google Cloud Announces Eighth-Generation TPUs Split into Training and Inference Chips
Google Cloud unveiled its latest tensor processing units, divided into models for AI training and inference, with significant performance gains. The company also committed to offering Nvidia's Vera Rubin chip later in 2026 and is collaborating on networking technology. TechCrunch…
TE
techjuice.pktechnology76 days ago
Google Cloud Launches Eighth-Generation TPUs Split into Training and Inference Chips
Google Cloud announced its eighth generation of tensor processing units, divided into TPU 8t for model training and TPU 8i for inference. The new chips offer significant performance improvements, including up to 3x faster training and 80% better performance per dollar. The compan…
TE
/TeslaModelS-55c1b0e1349049f7922082f677a41901.jpg)
investopedia.comfinance76 days ago
Tesla Reports Strong Q1 2026 Earnings as Google Unveils New AI Chips
Tesla beat expectations on Q1 2026 revenue, earnings, and subscriptions, driving a 5% stock surge. Google introduced specialized chips for AI training and inference. Other developments include IBM and Kinder Morgan earnings, a fatal shooting at Mexico's Teotihuacán pyramids, and…
UN
FI
LI
FI
+91Substrate placeholder — needs reviewfinance78 days ago
DataBank Secures $2 Billion Financing for Red Oak Data Center Campus Near Dallas
Dallas-based DataBank has obtained $2 billion in construction financing for three data center buildings in Red Oak, Texas, with plans for a fourth building funded by an additional $600 million. The campus, leased to a major hyperscaler, targets growing demand for AI inference com…
Substrate placeholder — needs reviewfinance81 days ago
BofA Analyst Reiterates Buy Rating on Apple, Sets $325 Price Target
BofA Securities analyst Wamsi Mohan issued a note reiterating a Buy rating on Apple Inc. and setting a $325 price forecast. The note highlights Apple's development of the M5 generation hardware focused on local inference. Benzinga reported the details of the analyst's assessment.
finance.yahoo.comtechnology83 days ago
Parasail Raises $32 Million in Series A Funding for AI Inference Cloud Services
Parasail, a startup providing cloud computing for AI model inference, has raised $32 million in a Series A funding round. The company processes 500 billion tokens daily by renting capacity from data centers worldwide. This funding supports scaling operations amid growing demand f…
Substrate placeholder — needs reviewfinance84 days ago
Meta Platforms Agrees to Deploy 1 Gigawatt of Custom AI Chips with Broadcom Through 2029
Meta Platforms announced a multi-gigawatt deal with Broadcom for custom AI chips, including an initial 1 gigawatt deployment of Training and Inference Accelerators. The partnership extends through 2029 for chip design and related technologies. Broadcom CEO Hock Tan will leave Met…
Substrate placeholder — needs reviewfinance89 days ago
Google Expands Partnership with Intel for AI Data Center Chips
Google has agreed to use multiple generations of Intel central processing units in its artificial intelligence data centers, expanding an existing partnership. The deal includes Intel's Xeon 6 CPUs for AI training and inference workloads. No financial terms or timeline were discl…
Substrate placeholder — needs reviewtechnology90 days ago
Anthropic Releases First AI Model After Nine Months of Development
Anthropic has released its first AI model following nine months of work on a complete technology stack. The development began with initial discussions on research directions and basic scripting for Llama model inference. @_jasonwei, associated with the project, shared these detai…
_J