DCD Magazine #62 - The coming wave
Will compute at sea sink or swim?
News - IT Hardware & Semiconductors, Yoram Novick, Maxta
The latest news from the AI, data center, telco, chip, and cloud sector
DCD Magazine #62 out now
-
DensityAI: Chip startup founded by former Tesla Dojo execs valued at $10bn – report
Company has raised “hundreds of millions” for development of chip with 3D DRAM stacking, report claims
-
Qualcomm and AWS partner on custom silicon for large-scale AI data centers
Will also work on optical connectivity solution with up to 1.6T
-
AWS to deploy 2 million additional Nvidia GPUs
Two companies expand their partnership
-
Alchip taps AWS for semiconductor design and simulation workloads
Expands use of cloud platform
-
Pinterest signs $4bn infrastructure agreement with AWS
Will use AWS' Graviton and Trainium chips
-
Meta signs "multibillion-dollar" agreement with AWS for large-scale Graviton5 deployment
Social media giant says it will deploy “tens of millions” of Graviton5 cores in support of agentic AI workloads
-
Amazon’s Trainium product lead has left the company – report
Not known where he will be working next
-
Microsoft claims to be first hyperscaler to "power on" Nvidia Vera Rubin NVL72 system
Plus the other clouds set to adopt the latest generation of Nvidia AI infrastructure
-
-
AWS partners with big chip co. Cerebras for AI “inference disaggregation”
Will be deployed on Amazon Bedrock in AWS data centers
-
OpenAI to use 2GW of Trainium on AWS, expand Amazon cloud contract by $100bn
After Amazon pumps $50bn into the AI company
-
AWS has “never retired” an Nvidia A100 server, CEO Matt Garman claims
Demand still high from those running HPC workloads
-
AWS launches Graviton5 CPU
Now available in preview
-
OpenAI signs $38bn multi-year cloud deal with AWS for access to Nvidia GB200s and GB300s
Deal to run for seven years, and will likely see AWS build additional infrastructure to meet startup’s capacity demands
-
AWS builds separate cabinet to bypass Nvidia networking hardware
Hyperscaler couldn’t fit its custom network cards in Nvidia’s racks, so it built a sidecar
-
Arm poaches Amazon’s chip director as it doubles down on efforts to develop its own chips - report
Former Arm VP of engineering is rejoining the company after a six-year stint at Amazon
-
AWS launches UltraServers with Nvidia Grace Blackwell GPUs
Instances now generally available
-
Amazon details Project Rainier compute cluster
Hundreds of thousands of chips connected
-
IonQ's Forte Enterprise available on Amazon's Braket quantum computing service
Joins company’s Aria-1, Aria-2, and Forte on the service
-
STMicroelectronics unveils new data center photonics chip for AI workload interconnects
The PIC100 offering is the first from STMicroelectronics' new product portfolio
-
AWS Trainium's lead mechanical thermal engineer joins OpenAI
Kenny Huynh moves to generative AI company
-
Amazon's Kuiper nabs Google custom chip director to build space silicon
Maruthy Vedam goes from TPUs to satellite semiconductors
-
AWS offers instances using HPE Compute Scale-up Server 3200
Ideal for large in-memory databases
-
Marvell Technology signs five-year cloud computing deal with AWS
Chip firm seals mutually beneficial agreement with cloud provider
-
Apple uses Amazon's Graviton and Inferentia chips, also explores Trainium2
Apple's Benoit Dupin went on stage at AWS Re:Invent 2024
-
AWS announces EC2 UltraServers with Trainium2 chips and EC2 instances with Blackwell GPUs
Also announces plans for Trainium3
-
Amazon to invest another $4bn in generative AI company Anthropic
In return for a promise to use AWS Trainium and Inferentia chips
-
IBM nears deal for access to AWS GPUs
Companies reportedly negotiating $475 million agreement