@Speedata1 profile picture
Speedata @Speedata1

The Analytics Processing Unit (APU) — purpose-built silicon for AI data prep, Apache Spark SQL & batch ETL. 100x faster. 90% lower TCO. Zero code changes.

Joined May 2020
  • Tweets 86
  • Following 143
  • Followers 85
41 Photos and videos
Pinned Tweet
Speedata@Speedata1
6 Nov 2025
Analytics and AI Data is about to have its #GPU moment. Workloads like AI, video, and databases already made the leap to specialized hardware, analytics is next. 50–100× faster performance. Up to 90% lower cost. The APU is here. linkedin.com/feed/update/urn… #Analytics #SPARK #AI
1
2
5
292
AI agent token use will grow 24x by 2030, generating more #SQL queries than humans @GoldmanSachs. The foundation for enterprise AI agents is structured data, it's why #OpenAI & #Anthropic partner with #Databricks & #Snowflake. But APUs beat CPU & GPU by 1-2 orders of magnitude.
1
1
62
The limiting factor in #VLSI design isn't model strength, it's whether the agent is operating inside your workflow and your context. @IAmAdiFuchs published a field guide to making #AI coding agents actually useful in hardware design now. speedata.io/post/know-your-u…
3
1,017
Speedata is hiring a Software Engineering Manager, Performance Modeling. Purpose-built silicon. Purpose-built careers. Join us. Email recruit@speedata.io or speedata.io/careers/senior-s…

SW Engineering Manager – Performance Modeling | Speedata

speedata.io
31
Model training had scaling laws. A clear improvement trajectory. Data pipelines don't, and it's a big reason most enterprise #AI pilots quietly fail. Our CEO @gelvan_adi goes deeper on it with @DanielNenni on Semiwiki.com semiwiki.com/ceo-interviews/…

SemiWiki

The Open Forum for Semiconductor Professionals

semiwiki.com
42
The Speedata Workload Analyzer projects how much faster your real analytics or AI data prep pipelines would run on the Analytics Processing Unit (APU). Try it: tinyurl.com/SpeedataWorkload… #AnalyticsAcceleration #CloudCost #FinOps #AIDataPrep #MLOps #Semiconductors #AI #CPU #GPU
How much faster would your Spark workloads run on dedicated analytics silicon? The Speedata Workload Analyzer projects how much faster your real analytics or AI data prep pipelines would run on the Analytics Processing Unit (APU). Three ways to test: 🟧 Upload a Spark event log in the browser - results in minutes 🟧 Run our CLI locally - your logs never leave your environment 🟧 Compare against TPC-DS benchmarks Real analytics pipelines, projected against the Analytics Processing Unit (APU). Try our Workload Analyzer: https://tinyurl.com/SpeedataWorkloadAnalyzer #AnalyticsAcceleration #CloudCost #FinOps #AIDataPrep #MLOps #Semiconductors #AI #CPU #GPU

ALT How much faster would your Spark workloads run on dedicated analytics silicon? The Speedata Workload Analyzer projects how much faster your real analytics or AI data prep pipelines would run on the Analytics Processing Unit (APU). Three ways to test: 🟧 Upload a Spark event log in the browser - results in minutes 🟧 Run our CLI locally - your logs never leave your environment 🟧 Compare against TPC-DS benchmarks Real analytics pipelines, projected against the Analytics Processing Unit (APU). Try our Workload Analyzer: https://tinyurl.com/SpeedataWorkloadAnalyzer #AnalyticsAcceleration #CloudCost #FinOps #AIDataPrep #MLOps #Semiconductors #AI #CPU #GPU
22
Why does a #GPU running SQL feel like it's barely trying? What does an LPU do that a GPU can't? The architectures are different because the workloads are different, and at production scale, those differences compound into real money. Learn the difference - tinyurl.com/apugpu
Why does a GPU running SQL feel like it's barely trying? Why are TPUs cloud-only? What does an LPU actually do that a GPU can't? The architectures are different because the workloads are different, and at production scale, those differences compound into real money. A breakdown of the five processors defining modern compute, and where each one fits https://www.speedata.io/post/apu-vs-lpu-vs-tpu-vs-gpu-vs-cpu-when-to-use-each-one #semiconductors #AI #APU #CPU #TPU #LPU #GPU #Nvidia #Google

ALT Why does a GPU running SQL feel like it's barely trying? Why are TPUs cloud-only? What does an LPU actually do that a GPU can't? The architectures are different because the workloads are different, and at production scale, those differences compound into real money. A breakdown of the five processors defining modern compute, and where each one fits https://www.speedata.io/post/apu-vs-lpu-vs-tpu-vs-gpu-vs-cpu-when-to-use-each-one #semiconductors #AI #APU #CPU #TPU #LPU #GPU #Nvidia #Google
1
23
Speedata is hiring: Lead SoC Architect. Own the architecture of an ASIC built from the ground up for #analytics and AI data prep, not a repurposed processor. recruit@speedata.io or apply for open #engineering roles here: speedata.io/careers/lead-soc… #hiring #Israel #startup #AI

Lead SoC Architect | Speedata

speedata.io
62
Every Wednesday we host a session for anyone who wants to see the Analytics Processing Unit (APU) in action. We spend 20 minutes running Spark SQL or AI data prep workloads on the APU, walk through the architecture, and Q&A. Register speedata.io/live-apu-demo. #AI #SQL #Spark
The Speedata analytics processing unit (APU) is purpose-built silicon for Apache Spark analytics acceleration, and every Wednesday we host a live session for anyone who wants to see it in action. We spend 20 minutes running Spark SQL or AI data prep workloads on the APU and walking through the architecture, then open up the final 10 minutes for live Q&A. Sessions alternate between 10am PDT and 10am CEST so it works whichever side of the Atlantic you're on. Pick a Wednesday. Link to register: https://www.speedata.io/live-apu-demo #AI #SQL #Spark #gpu #semiconductor

ALT The Speedata analytics processing unit (APU) is purpose-built silicon for Apache Spark analytics acceleration, and every Wednesday we host a live session for anyone who wants to see it in action. We spend 20 minutes running Spark SQL or AI data prep workloads on the APU and walking through the architecture, then open up the final 10 minutes for live Q&A. Sessions alternate between 10am PDT and 10am CEST so it works whichever side of the Atlantic you're on. Pick a Wednesday. Link to register: https://www.speedata.io/live-apu-demo #AI #SQL #Spark #gpu #semiconductor
50
The GPU-first model made sense when AI was experimental. In production, efficiency is the priority. Running the wrong workload on the wrong chip means overpaying in power, memory, and infrastructure costs. We broke down the AI Ops pipeline: speedata.io/post/one-chip-ca…
3
3
114
Speedata is looking for a Board Designer. This role is a great fit for someone who wants to take full ownership of board development and play a key part in building our products. Apply here: lnkd.in/dNFHpyKc #semiconductors #Boarddesign #engineering #hiring #AI #gpu
2
115
AI agents ask analytics questions. But can your infrastructure answer them quickly? Agentic Analytics, executing advanced analytics queries from an LLM is only useful if the answer comes back fast. We discuss where the pipeline bottleneck lives. Recording: lnkd.in/eAJreXaM
1:06
1
21
Speedata retweeted
Adi Gelvan@gelvan_adi
Apr 17
Pointing AI at the repo isn't enough. @Speedata1's AI expert @IAmAdiFuchs breaks down how we used a UART-to-AXI bridge to test AI in VLSI verification, and what it actually takes to make it work. speedata.io/post/ai-assisted… #VLSI #ChipVerification #UVM #AIEngineering #AI
1
4
396
Speedata webinar, "One Chip Can't Do It All: The New #AI Tech Stack," is tomorrow 1PM EST. #GPUs, TPUs, LPUs, APUs - each one was built for a different job. We're breaking down where each processor fits in the AI compute pipeline - join us! linkedin.com/event/manage/74…
43
Speedata retweeted
Adi Gelvan@gelvan_adi
Apr 8
Replying to @LipBuTan1
@LipBuTan1, a @Speedata1 investor, is leading @intel's partnership w @elonmusk to reimagine chip manufacturing. At Speedata, he's backing our purpose-built silicon, the APU, to accelerate the massive #Spark, ETL, and #AI data prep workloads. Learn more tinyurl.com/sr6n9z4h

Book A Demo | Speedata

speedata.io
Intel
@intel
Apr 7
Intel is proud to join the Terafab project with @SpaceX, @xAI, and @Tesla to help refactor silicon fab technology. Our ability to design, fabricate, and package ultra-high-performance chips at scale will help accelerate Terafab’s aim to produce 1 TW/year of compute to power future advances in AI and robotics. It was fun hosting @elonmusk at Intel this past weekend!
2
4
360
Join our webinar, "One Chip Can't Do It All: The New AI Tech Stack" on April 14 - we'll break down where each processor, APUs, GPUs, LPUs and TPUs fit in your stack. Register here -linkedin.com/events/74294952…

LinkedIn Login, Sign in | LinkedIn

Login to LinkedIn to keep in touch with people you know, share ideas, and build your career.

linkedin.com
Adi Gelvan@gelvan_adi
Apr 7
If you're still running #AI workloads on general-purpose hardware like GPUs for everything, CPUs maxed out on Spark, this is the session where @Speedata1 breaks down where APUs, #GPUs, #TPUs and #LPUs fit in your tech stack. April 14, live. Register here linkedin.com/events/74294952…
43
Speedata retweeted
Liraz Blumenkrantz Menchell@LirazM
Apr 6
A jar of #Nutella just traveled farther than any human in history. Meanwhile, some #queries are still running… somewhere
Anonymous
@YourAnonOne
Apr 6
The farthest pot of Nutella in the history of humanity
0:14
1
1
112
Speedata retweeted
Adi Gelvan@gelvan_adi
Apr 6
We are hiring a VLSI Engineer in our Israel offices. Join us. #cpu #gpu #ai #ApacheSpark
Speedata@Speedata1
Apr 6
Speedata is growing at the speed of silicon! We're #hiring a new VLSI Design Engineer. Apply on LI or email your CV to recruit@speedata.io #CUP #GPU #AnalyticsAcceleration #ApacheSpark #AI linkedin.com/jobs/view/43977…
1
1
49
Speedata is growing at the speed of silicon! We're #hiring a new VLSI Design Engineer. Apply on LI or email your CV to recruit@speedata.io #CUP #GPU #AnalyticsAcceleration #ApacheSpark #AI linkedin.com/jobs/view/43977…

Speedata.io hiring VLSI Design Engineer in Netanya, Center District, Israel | LinkedIn

Posted 11:37:28 AM. Speedata is modernizing analytics infrastructure with the first purpose-built ASIC processor, the…See this and similar jobs on LinkedIn.

linkedin.com
104
Speedata retweeted
Liraz Blumenkrantz Menchell@LirazM
Mar 31
Replying to @Speedata1
@Speedata1 #Statement Regarding recent developments
KITKAT
@KITKAT
Mar 29
Regarding recent press coverage
2
4
37
Speedata retweeted
Jonathan Friedmann
@friedmannjonat1
Mar 30
Three years ago, on the 25th of March 2023 Jacob Ziv passed away. Jacob Ziv is not one of the most known Israeli scientists. He is well known in academia, and I suspect that most electrical engineers—and some computer science graduates—are familiar with his work. However, I doubt that the general public is aware of his contribution to the world. And yes—his impact is global, far beyond Israel. Jacob Ziv was an electrical engineer and information theorist. Together with Abraham Lempel, he developed the well-known Lempel–Ziv (LZ) lossless data compression algorithm. For those who haven’t heard of it—this algorithm was developed about 50 years ago, and today it is almost impossible to think of any digital device that does not rely on it. This includes cell phones, personal computers, laptops, data center servers, cameras, TVs, image formats of all kinds, PDF files, network protocols, and much more. On a personal level, I was fortunate enough to meet him during my Ph.D. After mentioning my interest in his work, my supervisor, Prof. Hagit Messer, simply took me to meet him. I had the privilege of hearing his thoughts on my research. Furthermore, Speedata's solution is simply filled with his algorithm
2
9
81
Load more