Sabitlenmiş Tweet
Frank Downing
692 posts
Frank Downing
@downingARK
Director of Research, AI & Cloud @ARKinvest Disclosure: https://t.co/fB3IDVVVnO
Katılım Nisan 2021
550 Takip Edilen18.2K Takipçiler
What becomes possible when you can generate a 5 second video in 3 seconds?
Jim Keller@jimkxa
We landed on @ArtificialAnlys It’s pretty fast
English
New businesses are booming!
Patrick Collison@patrickc
Stripe Atlas just hit 100,000 all-time incorporations. Q1 2026 is +130% Y/Y.
English
Software development job postings continue to climb higher since troughing one year ago
English
AWS expansion to 37.7% from 35%
MSFT intelligent cloud down to 39.7% from 42.2%
Google Cloud expansion to 32.9% from 30.1%
Note: MSFT doens't report Azure only margins. Intelligent cloud includes more legacy products that are higher margin but slower growth so Azure growth creates headwind
English
Cloud earnings looking very strong. Flash update before the calls:
$AMZN: AWS acceleration to 28% from 24%
$MSFT: Azure acceleration to 39% from 38%
$GOOG: GCP acceleration to 63% from 48%
English
Grok 4.3 jumps above Anthropic's Sonnet 4.6 in intelligence and costs 1/10th to run. Nice update from @xai.
Artificial Analysis@ArtificialAnlys
xAI has launched Grok 4.3, achieving 53 on the Artificial Analysis Intelligence Index with improved agentic performance, ~40% lower input price, and ~60% lower output price than Grok 4.20 The release of Grok 4.3 places @xAI just above Muse Spark and Claude Sonnet 4.6 on the Intelligence Index, and a 4 points ahead of the latest version of Grok 4.20. Grok 4.3 improves its Artificial Analysis Intelligence Index score while reducing cost to run the benchmark suite. Key Takeaways: ➤ Grok 4.3 improves on cost-per-intelligence relative to Grok 4.20 0309 v2: it scores higher on the Intelligence Index while costing less to run the full benchmark suite. Grok 4.3 costs $395 to run the Artificial Analysis Intelligence Index, around 20% lower than Grok 4.20 0309 v2, despite using more output tokens. This makes it one of the lower-cost models at its intelligence level ➤ Large increase in real world agentic task performance: The largest single benchmark improvement is on GDPval-AA, where Grok 4.3 scores an ELO of 1500, up 321 points from Grok 4.20 0309 v2’s score of 1179 Grok 4.3, surpassing Gemini 3.1 Pro Preview, Muse Spark, Gpt-5.4 mini (xhigh), and Kimi K2.5. Grok 4.3 narrows the gap to the leading model on GDPval-AA, but still trails GPT-5.5 (xhigh) by 276 Elo points, with an expected win rate of ~17% against GPT-5.5 (xhigh) under the standard Elo formula ➤ Grok 4.3’s performs strongly on instruction following and agentic customer support tasks. It gains 5 points on 𝜏²-Bench Telecom to reach 98%, in line with GLM-5.1. Grok 4.3 maintains an 81% IFBench score from Grok 4.20 0309 v2 ➤ Gains 8 points on AA-Omniscience Accuracy, but at the cost of lower AA-Omniscience Non-Hallucination Rate of 8 points, so Grok 4.20 0309 v2 still leads AA-Omniscience Non-Hallucination Rate, followed by MiMo-V2.5-Pro, in line with Grok 4.3 Congratulations to @xAI and @elonmusk on the impressive release!
English
@thsottiaux Can you do a more graceful update process that doesn't require full admin on mac every time?
English
The Next Gen Internet team at ARK is firing up a new substack to share our longer form research thoughts from @GrousARK @varshikaARK @JozefARK and I. If you want to follow along, check out the first issue:
arknextgeninternetteam.substack.com/p/openai-and-t…
English
All data referenced above taken from each company's quarterly earnings releases.
English
Revenue backlog nearly doubled q/q to $460 B, and they are raising their CAPEx guide from $185 B at high end to $190 B. Notably, this CAPEX comes at a cost and will be a headwind to margins as elevated depreciation expenses flow into the income statement.
English
All the cloud numbers are incredibly impressive for businesses of this size ($80 B run rate for Google Cloud, $150 B for AWS), but Google's full stack approach to AI - being the only vertically integrated player from chips to end applications - appears paying off.
English
AMZN: ir.aboutamazon.com/news-release/n…
MSFT: microsoft.com/en-us/investor…
GOOG: s206.q4cdn.com/479360582/file…
English
All data based on @ArtificialAnlys as of today. Company level release cadence is based on time between new highest scoring model release for that company each year based on AA's Intelligence Index.
English
I think this both represents the compounding advantage of frontier models driving a real acceleration in the research process and the fierce competitive landscape across a narrowing set of firms.
English
If it feels like the pace of AI is accelerating, that's because it is!
Median days between new frontier model releases has compressed from 37 to 11 since 2023, while number of frontier releases is on track to increase for the fourth consecutive year.
English
OpenAI has been on a bit of a tear lately. Wrote about why Codex growth, GPT-5.5, and recent product launches matter in ARK’s latest newsletter.
ARK Invest@ARKInvest
English
@LorenzoARK I think same same for all intents & purposes in this case, but they used the term "acquisition" #xai-joins-spacex" target="_blank" rel="nofollow noopener">spacex.com/updates#xai-jo…
English
Largest acquisitions of private, VC-backed tech companies:
2012: Instagram → $1B
2014: WhatsApp → $19B
2022: Figma → $20B (scrapped)
2025: Wiz → $32B
2026: Cursor → $60B
A $100B private tech acquisition is coming.
With SpaceX, we'll soon have 10+ companies above $1.5T market cap. They can spend low-single-digit % of market cap to not fall behind in AI. Cheap
TechCrunch@TechCrunch
Cursor was on track to close a $2 billion fundraise this week, but chose to halt the round after SpaceX offered a $10 billion "collaboration fee" and a path to a $60 billion acquisition. spr.ly/6019BBDva9
English