Tuesday, June 3, 2026

AI Intelligence Briefing

Scout + Quant · Sourced from TTL Lab Intelligence · June 3, 2026

$80B
Alphabet AI Raise
10×
Qwen3.7 Kernel Speedup
14
AI Acquisitions / Week
$20
Claude Pro / Month (Jun 15)
89%
Spec. Decode Accept

Top Stories

01 · China · Multimodal Agent
china
up

Qwen3.7-Plus: Unified Multimodal Agent

Alibaba drops a single model that perceives scenes, reads screens, operates GUIs, and writes code from visual refs — no separate vision model needed. Beats proprietary on visual agent benchmarks.

ScreenSpot Pro
79.0
AndroidWorld
81.0
MathVision
90.3
IFBench
79.1
Text: Terminal Bench 70.3 · WMT24++ 84.6 · GPQA 92.4
02 · Business · Capital
fund

Alphabet Raises $80B for AI Infrastructure

Largest AI-specific equity raise in 2026. Berkshire Hathaway participated. Confirms hyperscaler capex race has no ceiling — AI infrastructure remains the top capital allocation priority for big tech.

$80B
Single equity raise · Alphabet / Berkshire
Hyperscaler Capex Accelerating Berkshire Backed
03 · China · Coding Agent
china
up

Qwen3.7-Max: 10× Kernel Optimization

35-hour autonomous kernel optimization. Sustained 1,000+ tool calls without context regression. First model validated across Claude Code, OpenClaw, and custom frameworks — not harness-specific.

Qwen3.7
10×
GLM-5.1
7.3×
Kimi K2.6
5.0×
DS-V4-Pro
3.3×
SWE-Pro 60.6 · SciCode 53.5 · LiveCodeBench 91.6
04 · Business · M&A
fund
alert

14 AI Acquisitions in One Week

OpenAI acquired Hiro (personal finance) + a second app company. 5 of 14 deals were in Europe. Vertical AI applications are being acquired for distribution and data. Consolidation accelerating.

14
Total Deals
5
Europe
2
OpenAI
incl. Hiro
Vertical AI M&A Heating Up
05 · Policy · US Government
policy

Trump Signs AI Executive Order

Requires AI companies to provide models to federal government before full public release. Original version watered down after lobbying by Zuckerberg, Musk, and Sacks. Creates new compliance overlay for release cycles.

Compliance impact on release timelines:
Incumbents Advantage Startup Burden Release Delays
06 · Infrastructure · Local AI
up

Local LLM Agent Stack: 16× Speedup

Full stack getting agent iteration from 10–15 sec down to 1–3 sec. Key unlocks: prefix caching (16×), CUDA graphs (20–25% latency), FP8 KV cache (halves bandwidth), speculative decoding (89% accept rate).

Optimization Gain
Prefix Caching 16× 11,470ms → 706ms
CUDA Graphs 20–25% latency reduction
FP8 KV Cache memory bandwidth
Spec. Decoding 89% MTP accept rate
Qwen3.6-27B: 320k tokens in 82GB · Gemma4-31B: ~74k tokens

Market Timeline
Jun 2 · Alphabet
$80B equity raise announced
Jun 2 · White House
Trump signs AI executive order
Jun 15 · Anthropic
Claude Pro repriced to $20/mo
May 20 · Alibaba
Qwen3.7-Max released
May 28 · Nscale
$2B UK sovereign AI raise
Market Signals
Hyperscaler Capex
AI infrastructure race has no ceiling — Alphabet's $80B raise confirms big tech is still in full sprint.
Agentic AI Pricing
Claude Pro at $20/mo signals demand for agentic workflows is now material enough for dedicated billing tiers.
Sovereign AI
Nscale's $2B (UK) + geographic partitioning of infrastructure — compliance + data residency are becoming market forces.
M&A Velocity
14 acquisitions in one week. Vertical AI is the acquisition target du jour — OpenAI buying Hiro for distribution + data.

Key Numbers
Alphabet Raise
$80B
Acquisitions / Week
14
Qwen3.7 Kernel Speedup
10×
Claude Pro (Jun 15)
$20/mo
Spec. Decode Accept
89%
Prefix Cache Speedup
16×
Nscale Raise
$2B
GPQA Diamond
92.4
Qwen SVG Score
1608
Europe Acquisitions
5 / 14