- AI Rundown
- Posts
- The AI Rundown by Lightscape Partners - 6/1/2026
The AI Rundown by Lightscape Partners - 6/1/2026
Anthropic hits a $965B valuation and ships Claude Opus 4.8 · SoftBank commits up to €75B to French AI data centers · Cognition more than doubles to a $25B pre-money on the vibe-coding wave.

Good morning and welcome back to another edition of The AI Rundown by Lightscape Partners.
A wild last week of May. Anthropic just leapfrogged OpenAI on paper, closing a $65B round at a $965B valuation and shipping Claude Opus 4.8 with a new "dynamic workflow" tool — the company also confirmed an upcoming next-gen model codenamed "Mythos." SoftBank stunned Europe with a €75B French data center pledge, while Cognition more than doubled its valuation to north of $25B pre-money even as AI coding agents become the most contested wedge in enterprise software. Beneath the headlines, hardware is back in style, hyperscalers are inking multibillion-dollar compute deals, Illinois passed the country's toughest AI safety bill, and Pope Leo XIV's first encyclical reframed AI as a moral question for the next century.
● Anthropic raises $65B at a $965B valuation, ships Opus 4.8, teases "Mythos." The company leapfrogged OpenAI on paper with the largest private financing in tech history and confirmed a new flagship model coming "in the next few weeks." Same week: a new "dynamic workflow" tool inside Claude Opus 4.8 and fresh details on its multi-year compute deal with SpaceX.
● SoftBank to deploy up to €75B into French data centers. Masayoshi Son's biggest single-country AI bet yet would build out training and inference capacity across France, anchoring Europe's response to U.S. and Chinese AI infrastructure. The deal positions France as the continent's AI gravity center and pushes the global data center capex super-cycle to a new high-water mark.
● Cognition more than doubles to $25B+ pre-money as the AI coding wars escalate. The Devin maker raised $1B+ in a round that crystallizes how much investors will pay for a credible "AI software engineer." Rivals Glean ($300M+ ARR) and GitHub Copilot are repricing — and even AI-coding skeptics like Scott Wu are warning that agents shouldn't fully replace humans.
If you haven't yet, please support the newsletter by subscribing!
Anthropic
As Anthropic launches Claude Opus 4.8, it raises $65B in new funding [Link].
● Anthropic closed a $65B round alongside the launch of Claude Opus 4.8, its most capable model yet
● The release adds long-horizon agentic capabilities aimed squarely at enterprise workflows
● Investors include a mix of strategic tech partners and sovereign capital
● The round vaults Anthropic into the top tier of private AI companies by both capital raised and product surface area
Anthropic leapfrogs OpenAI with a record $965B valuation and says its "Mythos" AI model is coming soon [Link].
● The $965B valuation is the largest ever for a private technology company and edges out OpenAI's last marked round
● Anthropic teased a next-generation model codenamed "Mythos," planned for wide release within weeks
● The financing positions Anthropic to absorb compute costs that have ballooned across the frontier-model cohort
● Pricing implies a sharp re-rating of Claude's enterprise pull-through versus OpenAI
Anthropic releases Opus 4.8 with new "dynamic workflow" tool [Link].
● The "dynamic workflow" tool lets Claude plan and re-plan multi-step tasks mid-execution
● Aimed at developers building agentic apps where the path to an outcome isn't fixed up front
● Pricing held flat versus Opus 4.5 despite material capability gains
● Early benchmarks show meaningful gains on long-tool-use evals
How long is Anthropic's lease with SpaceX? Opinions vary [Link].
● A reported multi-year compute arrangement between Anthropic and SpaceX is fueling speculation about Anthropic's infrastructure roadmap
● Estimates of lease length range from a few years to a full decade
● The deal underscores how frontier labs are locking down power and capacity outside the traditional hyperscaler stack
● It also raises fresh questions about Anthropic's reliance on Amazon as primary cloud partner
Hardware + Infrastructure
After Nvidia's $20B not-aqui-hire, AI chip startup Groq reportedly raising $650M [Link].
● Groq is raising $650M just weeks after Nvidia's $20B-effective deal for rival AI chip talent
● The round would value Groq well into the multibillion range and lengthen its runway for LPU production
● Customer demand for non-Nvidia inference silicon has surged with model latency now a competitive variable
● Expect more capital to flow into specialized inference silicon as token economics tighten
Hardware's back: AI supercharges server, PC and memory sales [Link].
● AI demand is pulling forward a multi-year refresh of servers, PCs, and high-bandwidth memory
● HBM supply remains the binding constraint at the chip layer through at least 2027
● AI-PC SKUs are now meaningful share of enterprise procurement budgets at Dell, HP, and Lenovo
● The story isn't "AI vs hardware" — it's that AI is finally lifting all hardware boats
Nvidia gives developers the tool to build secure, autonomous AI workers that scale [Link].
● Nvidia released a new developer framework for building production-grade agentic systems
● It packages identity, sandboxing, and policy controls — addressing the biggest blockers to enterprise deployment
● Tight integration with NIM microservices and CUDA enterprise stack
● Yet another move by Nvidia to climb the value chain from chips to AI software platform
The AI GPU market has belonged to hyperscalers, but AMD's MI350P is coming for the enterprise [Link].
● AMD's new MI350P targets mainstream enterprise data centers rather than hyperscalers
● Dell unveiled servers configured around the part at Dell Technologies World
● Pricing and air-cooling characteristics aim at the enterprise gap Nvidia's flagship parts left open
● Could be the first credible inroad for AMD in non-hyperscale AI infrastructure
Enterprise AI / Cloud
Snowflake's stock surges after-hours on solid earnings beat and multibillion-dollar AWS cloud deal [Link].
● Snowflake beat on the top and bottom line, with Cortex AI revenue contribution outpacing expectations
● Announced a $6B+ multi-year AWS commitment for next-gen compute
● Net revenue retention bounced after several quarters of decel
● Sridhar Ramaswamy framed AI as the company's new platform tailwind, not a pricing risk
Asana acquires StackAI to run AI agent workflows across enterprise systems [Link].
● Asana acquired no-code agent-builder StackAI to embed agentic workflows into its work-management platform
● The deal positions Asana to compete in the "agents-as-features" race against Monday, ClickUp, and Smartsheet
● StackAI's drag-and-drop builder lowers the friction for non-technical users
● Asana's stock has lagged peers, and management is betting agents reset the growth story
Glean's top line crosses $300M as AI budget-cutting becomes its major selling point [Link].
● Enterprise search/agents leader Glean hit $300M+ ARR with rapid net-new logo growth
● Customers are using Glean to consolidate spend across Microsoft, Google, and point-solution AI vendors
● Margin profile remains favorable as it leans on customer-owned LLM access
● The company is increasingly positioned as the "Workday for AI work"
IBM, Red Hat launch $5B Project Lightwell to boost open-source security [Link].
● IBM committed $5B over five years to harden the open-source supply chain underpinning enterprise AI
● The initiative funds vulnerability tooling, code-provenance research, and developer education
● It's also a play to anchor Red Hat as the enterprise distribution layer for trustworthy AI
● Expect cross-vendor adoption of the resulting tooling stack
Startup Funding & Valuations
Vibe coding startup Cognition more than doubles valuation in new $1B+ round [Link].
● Cognition, maker of Devin, raised $1B+ at a $25B pre-money valuation
● Round more than doubles the company's last valuation in roughly six months
● Customers are reportedly running Devin as a partial replacement for junior engineering hires
● One of the most heavily marked-up AI deals of the cycle
OpenRouter raises $113M to bring order to enterprise AI inference routing [Link].
● OpenRouter raised $113M to scale its model-routing layer across hundreds of LLM endpoints
● Valuation more than doubled to $1.3B over the past year
● Enterprises are increasingly multi-model and need a single billing and policy layer
● One of the cleanest "picks-and-shovels" companies in the inference economy
AI video analysis startup Airis Labs raises $60M [Link].
● Airis Labs raised $60M to expand its real-time video understanding platform for physical security and retail
● The platform combines fine-tuned vision models with rule-based alerting
● Targets a market still dominated by legacy VMS players
● Signals continued investor appetite for vertical AI in physical-world workflows
Logistics startup Stord raises $250M at $3B valuation amid rapid sales growth [Link].
● Stord raised $250M at a $3B valuation as merchants race to consolidate fragmented fulfillment networks
● AI-driven routing and inventory placement are the core differentiators
● The company is now operating at meaningful scale across the U.S. and Europe
● Validates the thesis that supply-chain software is one of the cleaner AI wedges in commerce
Exclusive: The next wave of AI drive-thrus is here — and a16z and Arc think it finally works [Link].
● A16z and Arc are betting AI drive-thru voice ordering has crossed the reliability threshold after early stumbles
● New systems handle accents, menu modifications, and upsells in low-latency conversational style
● Major QSR chains are quietly piloting at scale after McDonald's and Taco Bell's earlier missteps
● A real-world stress test for the production-grade voice stack
Data Centers + Energy
SoftBank says it will invest up to €75 billion to build French data centers [Link].
● SoftBank pledged up to €75B to build AI-focused data centers in France, with cooperation from the French government
● The investment would dwarf any prior European AI infrastructure commitment
● Anchors France's bid to be Europe's AI gravity center post-Mistral
● Pushes the global AI capex super-cycle to a new high-water mark
Utilidata raises $40M more to optimize data center power use [Link].
● Utilidata raised another $40M to scale its grid-edge software platform for data center load management
● Demand from hyperscalers is being driven by interconnect queue times now stretching past 2030 in some U.S. markets
● The company partners with Nvidia and major utilities
● Energy is now the binding constraint on the AI build-out — and software is part of the answer
Amazon Thinks the Future of Data Centers Depends on a Technical Problem It Just Solved [Link].
● AWS unveiled a new approach to data center power distribution that it says unlocks materially more compute per facility
● The design rethinks the power-substation-to-rack chain that has gated AI deployment density
● Amazon plans to deploy the architecture across new builds beginning this year
● A meaningful step toward AWS closing the AI-infrastructure gap with Microsoft and Google
Workforce + Labor
CEOs blame AI for layoffs, but an MIT professor says it fits a long-running pattern to find a cover story [Link].
● MIT economist Daron Acemoglu argues "AI layoffs" are mostly cover for cost-cutting after the ZIRP overhiring binge
● Wix, Block, Snap, and Atlassian all cited AI in recent reductions despite limited evidence of true automation
● Acemoglu warns the framing risks misallocating policy and investment attention
● The labor market data still doesn't show net job losses pinned to AI
Largest study of AI hiring algorithms to date finds "clear racial disparities" — over 25% of Black applicants tainted [Link].
● A Stanford-led audit of leading AI hiring tools found over 25% of Black applicants were rated more negatively than statistically comparable peers
● The study covered tools used at scale by Fortune 500 employers
● The findings are likely to accelerate state-level regulation in the wake of Illinois' new law
● Pressure on enterprise procurement teams to demand explainability evidence
What ClickUp's mass layoff tells us about the future of work [Link].
● ClickUp laid off a significant portion of its workforce, citing AI productivity gains
● The cuts came alongside record customer growth — a now-familiar pattern in SaaS
● Industry-wide hiring freezes for generalist software roles are deepening
● Whether AI is causally responsible or just convenient cover is still hotly debated
Policy + Regulation
Illinois Lawmakers Just Passed America's Strongest AI Safety Bill [Link].
● Illinois passed sweeping AI legislation covering hiring, healthcare, government use, and high-risk applications
● The law goes further than California's draft and includes private rights of action
● Governor Pritzker is expected to sign the bill into law in the coming weeks
● Sets a de facto national floor given Illinois' employment-law leverage on Fortune 500 HR systems
CNN sues Perplexity alleging "massive" copyright infringement [Link].
● CNN became the latest major publisher to sue Perplexity over alleged scraping and verbatim reuse of content
● The complaint accuses Perplexity of degrading attribution and harming subscription economics
● A growing roster of publishers — NYT, Wired, News Corp — are now suing AI search products
● A potential bellwether case for fair-use boundaries in the agent era
What Pope Leo XIV's First Encyclical Says About the Power of AI [Link].
● In his first papal encyclical, Pope Leo XIV called AI both transformative and a potential instrument of exclusion and harm
● The document urges global governance frameworks and labor protections
● Anthropic representatives attended the encyclical's presentation in Rome
● The cultural moment marks AI's transition from tech debate to society-wide ethical question
Lightscape Portfolio is Hiring
Our portfolio companies have 633+ open roles across 20 companies right now. A few featured opportunities this week:
● Shield AI (Series F · 323 roles) — Director - People Ex - Business Partner - APJ & Middle East — Melbourne · Browse all
● Etched (Series A · 100 roles) — Head of Manufacturing Design Program Manager — San Jose · Browse all
● Parloa (Series D · 61 roles) — Lead Agent Architect - SAP — New York Office · Browse all
● Formic (Series B · 35 roles) — Chief of Staff — San Francisco / Oakland, CA · Browse all
● Standard Bots (Series B · 32 roles) — Experiential Marketing Lead — New York City, NY · Browse all
● Articul8 (Series B · 18 roles) — Principal Applied AI Researcher - Domain-Specific Models — Dublin, CA · Browse all
● Fuse (Series A · 17 roles) — Health Physicist — Albuquerque, NM · Browse all
● Sapien (Seed · 10 roles) — Founding AE — NYC · Browse all
● Databento (Series A · 8 roles) — VP of Finance — Remote / SF / NYC / Boston / SLC · Browse all
● WearLinq (Series A · 7 roles) — Senior Fullstack Engineer — Warsaw, PL · Browse all
● Near Space Labs (Series B · 5 roles) — Director of Marketing — Brooklyn, NY · Browse all
● Reflexivity (Series B · 5 roles) — Product Manager, Data — London, UK · Browse all
● Profitmind (Series A · 3 roles) — Senior AI Agentic Engineer — Remote, US · Browse all
● Bucket Robotics (Seed · 2 roles) — Senior Full Stack Engineer — San Francisco, CA · Browse all
● BrowserOS (Seed · 1 role) — ML Research Engineer Intern — San Francisco, CA · Browse all
● IronGrid (Seed · 1 role) — Manufacturing Underwriter / Senior Underwriter — San Francisco, CA · Browse all
● Limina (Series A · 1 role) — Senior Enterprise Account Executive — Remote / Toronto, ON · Browse all
● Merlin AI (Seed · 1 role) — Product Engineer (Customer-Facing) — Bengaluru, India · Browse all
● Parallax Worlds (Seed · 1 role) — Founding Robotics Researcher — San Francisco
● Synthefy (Seed · 1 role) — Founding Product Manager — San Francisco Bay Area, CA · Browse all
Know someone who'd be a great fit? Forward this newsletter or send them straight to the role.
Thanks for reading. If you haven't yet, please support the newsletter by subscribing!