Dave Vellante (@dvellante) header
Dave Vellante (@dvellante) profile picture

Analyst, Writer, Cofounder & Co-CEO SiliconANGLE Media, co-host of @theCUBE, fortunate husband, father of four, advocate of #WiT #BreakingAnalysis

Boston
Joined January 2008
Photo by Dave Vellante (@dvellante)Photo by Dave Vellante (@dvellante)Photo by Dave Vellante (@dvellante)Photo by Dave Vellante (@dvellante)Photo by Dave Vellante (@dvellante)Photo by Dave Vellante (@dvellante)Photo by Dave Vellante (@dvellante)Photo by Dave Vellante (@dvellante)Photo by Dave Vellante (@dvellante)Photo by Dave Vellante (@dvellante)
Right on Dion - It's fundamental for AI sovereignty isn't it Dion?
I am finding that it’s very hard to beat token routers for wide, easy access to a vast number of AI models of every kind (text, coding, images, video generation, etc.) You can rapidly test different models for ability to achieve tasks without signing up and setting different keys. You can track token spend in one place, change routing rules, and lots more. If you’re not using a token router for AI access, you’re losing speed, capability, and choice. A good routing partner is a big part of the future of AI and token economics!
2
448
This week's Breaking Analysis features @Craw - we asked the question: Can enterprise AI pay for the AI boom; and how long will it take? With customer data from @qualitateio & @AnthropicAI S1 nuggets from @michaeljburry via @SiliconANGLE siliconangle.com/2026/10/31/…
2
6
490
I'm looking for $32B from Azure (+46% YoY) and $13B from GCP (+80% - subset of Google Cloud overall) - @sarbjeetjohal here's the 7 year history - these 5 will account for $400B this year
3
2
4
673
Sarb I was a little low on the Azure #. Way low on Google Cloud +103% vs my GCP target of 80%. Wow! GCP is likely up 110% or more - remember GPC grows meaningfully faster than Google Cloud overall. GCP $57B (est) in '26 - GCP now larger than AWS was in 2020.
1
153
Anthropic's S1 implies it has a $40T TAM. Global GDP is ~$117T, growing at ~3% annually. Compounded for 15 yrs that puts GDP at ~$180T, implying $ANT's TAM is 22% of 2041's GDP. Economists estimate the Internet's contribution to global GDP at about 15-20%.
3
545
Definitely a packed house at #OracleAIWorld - co-CEO's focus on serious business conversations. 1. AI needs to access your data; 2. AI needs context; 3. AI needs high performance infa; and 4. AI needs to work the way your business works -Mike Sicilia
1
4
409
Micron has become one of the most important companies in AI infrastructure. Along with SK hynix and Samsung Semiconductor the company is significantly more strategic to chip design firms, fabless designers, IDMs, hyperscaler custom-silicon teams, OEMs and enterprise customers. The question everyone's asking is have we entered a new era that structurally alters the traditional cyclicality of the memory business; or is the current trend temporary? In this Breaking Analysis we do a deep dive into how AI is changing the economics of memory. We reason through the implications of the massive divergence between memory ASPs and bit growth, we discuss the strategic role of HBM and the changing memory storage hierarchy in the context of Micron’s roadmap. Inspired by a roadmap preso from @MicronTech's Scott DeBoer this episode brings survey data in from @qualitateio with proprietary insights on how chip designers and enterprise customers are dealing with memory becoming an increasingly large piece of the bill of materials pie. siliconangle.com/2026/10/23/…
🤖 Made with AI
1
3
411
Congrats to @S_Kadakia and the team at @qualitateio on your $25M Series A, led by @insightpartners. It was great having Sagar and @CrissyBehrens in our @theCUBE @NYSEWired studio to cover this news. Here's a quick look at what Sagar is up to...It's absolutely amazing!
4
2
14
58,374
Huge day for @MarvellTech today. Deep dives into the numbers, a $400B TAM - $20B rev guide for F28 FY31 rev target $70-$90B w serious operating leverage. Investors like the story...
2
1
5
1,039
Good day with the @IBM execs - not much I can report at this time but will share thoughts on the AMA with @ArvindKrishna - one of the top AMAers in the industry!
2
2
15
1,872
In case you missed our deep dive ahead of @CoreWeave #FullyConnected26 - 15 in-depth interviews with CoreWeave customers - 7 heavy users, 6 newbies and 2 that passed. Strengths, gaps, how CoreWeave competes with @AWS & other hyperscalers, Neo-clouds & On-prem utilization hurdles - thecuberesearch.com/327-brea…
1
3
411
First big customer event for @CoreWeave - #FullyConnected26 comprised mostly of builders, operators and some partners - but mostly customers doing AI. We have some outstanding guests coming on @theCUBE today - @GoldbergChen @ia_buck fm @nvidia @cognition @VAST_Data @liquidai @CrowdStrike @IBM @CaterpillarInc and many more...
1
12
697
OpenAI announced a Pro 500 tier at its DevDay event. The details read like a mobile phone plan from the 00s. We’re introducing dots for all Pro customers, a new Pro 500 tier that includes our highest usage limits and Ultrafast, and adjusting the usage allowance included in Pro 200 after October 29th. With models like GPT-6.1 Sol, you can still get more work done with Pro than you could before. Here’s what’s changing. Changes to Pro 200 GPT-6.1 Sol brings near-Astra capabilities to a model you can use every day. It’s more intelligent than GPT-6 Sol, so you spend less time going back and forth and can take on work the previous model couldn’t handle. As our models become more capable and efficient we’re changing how we calculate the usage for Pro 200. Our commitment is that, over time, your subscription will continue to help you get more work done with an increasing level of quality. We are also committing to not reintroducing the 5hr limit so that you can fully use your weekly usage when you want. Starting October 30, 2026, your included usage in ChatGPT Work and Codex will decrease from 20 times to 10 times the ChatGPT Plus allowance. Your GPT-6 Pro limit in Chat will decrease from 200 to 100 messages per week. Your price remains unchanged and you’ll keep your current limits through October 29, 2026. As a thank you for being an existing subscriber and to help you continue to build your most ambitious ideas and transition, we’re also giving you a one-time grant of 62,500 usage credits (a $2,500 USD value). You’ll see these appear in Settings > Usage. They will expire on December 31, 2026, and use of the credits is governed by the Service Credit Terms. A new Pro 500 tier For those who want even more usage and speed, the new Pro 500 tier offers our highest usage allowance, with 25 times the usage of Plus. It also includes access to Ultrafast, with up to 8 times faster token generation. Learn more about Pro 500and compare plans. Your Pro subscription includes dots Your Pro subscription now includes dots: remarkably capable, always-on agents built to handle everything. Powered by GPT-6 Astra, dots have their own cloud computer, learn from feedback over time, and can work towards your goals 24/7. Conversations with your dot and work it does directly are included and don’t draw on your ChatGPT usage allowance. Learn more about dots.
1
2
3
1,256
I just finished reading @mattwbaker's new book "Unconventional Wisdom." Congrats Matt on writing this. Loved the closing chapter. Universally applicable advice in any field. After wading through your IT geek days I really got into the inside baseball on @DellTech, EMC & @VMware. So many people in the book are business friends and it was awesome to see them from your perspective. @furrier @sarbjeetjohal check it out - good read amazon.com/Unconventional-Wi…
3
2
1
7
673
Nice job @rafeuddin_ - excellent questions. @jpatel41 you are such a clear thinker. I was just talking to Sid Nag this morning about the merits of designing your own silicon. You mentioned that a couple times in the interview. And "never fight a mega trend" reminds me of the Jeremy Burton quip "never fight fashion." On the bubble - @johnfurrier @bjbaumann2014 I'd say Jeetu is selectively on the bubble...
1
7
743
In this week's Breaking Analysis I partnered with my friends from @qualitateio to run a custom deep dive survey on @CoreWeave adoption. We stress tested public claims by CEO Mike Intrator against customer ground truth. siliconangle.com/2026/09/26/…
2
10
7,150
Spain’s IA360 plan (21 Sep) commits to data centers that “guarantee our sovereignty over the data stored there,” plus a gigafactory, frontier oversight, and a social contract on AI. Capacity language is not a definition. Run the five pillars before you celebrate the megawatts: Territorial — Spanish floors for data at rest help. Inference in motion is the real test. Operational — who pages at 2am, and from which country? Technological — can you audit, fork, or self-host what sits on those floors? Legal — which ultimate parent and which compelled-disclosure regime sits above the Spanish SPV? Financial — if the token meter and the exit fee are unknown, residency is a brochure line. Write the disqualifier first: the one property that, if missing, makes the stack ineligible no matter how many megawatts you buy. If you cannot state it in a sentence, you are not running a sovereignty program. You are running a procurement with sovereignty in the title. Sovereignty Assessment if you want the rubric forced before the RFP → [email protected] @sanchezcastejon @oscarlopeztwit @mintradigital @mariagv @agenciaAESIA @BSC_CNS @dvellante @furrier @theCUBE
1
1
2
147