Funding & Deals

Europe's fast-growing "vibe coding" startup Lovable says it has crossed $500M in annualized revenue run rate, up from $400M in February, with usage now at one million new projects a week and 50M+ built to date. The company says its mostly non-technical users are increasingly building software they intend to monetize or run their businesses on - from storefronts to internal CRMs and HR tools. The open question is maintenance: whether vibe-coded software survives once the building stops.

Anthropic confidentially submitted a draft S-1 registration statement to the SEC for a proposed initial public offering of its common stock. The filing gives the company the option to go public after SEC review, with timing dependent on market conditions; share count and price aren't set. The move follows Anthropic's recent $65B raise at a roughly $965B valuation and a reported $47B revenue run-rate, positioning a potential IPO as a bellwether for the AI sector.

A historic US$65bn Series H round pushed Anthropic's post-money valuation to US$965bn, edging past OpenAI's US$852bn to make it the world's most valuable pure-play AI company. The round, led by Altimeter, Dragoneer, Greenoaks, and Sequoia, lands as both firms head toward IPOs. Anthropic credits an annualized revenue run rate above US$47bn and an enterprise-heavy customer base that it argues underpins a path to profitability.

A fresh wave of AI automation anxiety is rattling public markets, with investors questioning which software and services firms can defend margins as models become more agentic. The story here isn't 'AI is big' it's that the boundary between app and model is blurring, forcing a repricing of incumbents' moats.

Anthropic raised $30B, pushing its valuation to $380B, signaling that investors still reward frontier-model scale despite growing scrutiny over costs and governance. The move intensifies competitive pressure on model providers to prove revenue durability, enterprise adoption, and efficient inferencenot just benchmark wins.

Cohere reportedly hit about $240M in annual recurring revenue with ~70% gross margins, reinforcing that enterprise AI can scale without purely consumer economics. The company's posture suggests IPO intent, with a narrative built around capital efficiency and regulated-industry adoption.

Reuters reports AI disruption fears contributed to a sharp pullback in Indian IT stocks, erasing roughly $50B in market cap in February. For services-heavy businesses, investors are demanding proof of defensible delivery modelswhere AI boosts margins without commoditizing the work.

Anthropic appointed Chris Liddell to its board, adding heavyweight experience across public-company finance and government. The move reads like an AI lab preparing for higher scrutinythink IPO readiness, risk management, and tighter operational controls.

TechCrunch reports Anthropic raised $30B in a Series G at a reported $380B valuation, intensifying the capital race behind frontier model development. The funding signals continued pressure to lock up compute, distribution, and enterprise adoption.

JPMorgan's commercial and investment bank appointed Guy Halamish as COO to oversee data and AI strategy, signaling AI is moving from experimentation to operational control. The organizational shift suggests increased focus on governance, data quality, model risk management, and scaled deployment across business lines.

Bill Ackman's Pershing Square reportedly invested $2B in Meta, framing it as an AI-driven thesis on platform durability and future monetization. It's a reminder that markets increasingly price AI not as a feature, but as a capex-heavy moat built on compute, data, and distribution.

Blackstone increased its investment in Anthropic to ~ $1B, highlighting how large allocators view frontier AI exposure as a strategic portfolio position. For builders and buyers, it reinforces that the 'model layer' is consolidating into a small set of massively capitalized players with bargaining power on pricing, partnerships, and distribution.

Amazon CEO Andy Jassy said Amazon expects roughly US$200bn in 2026 capex, with the majority aimed at AWS to meet AI workload demand. The emphasis is clear: hyperscalers are racing to convert AI scarcitychips, power, data centre capacityinto durable platform advantage.

Alphabet CEO Sundar Pichai said 2026 capex is expected at US$175bnUS$185bn, positioning infrastructure as the bedrock of Google's AI roadmap. Google also pointed to rapid shipping across AI features and claimed major efficiency gains, including a 78% reduction in Gemini serving unit costs over 2025 via optimisations and utilisation.

EY-Parthenon's CEO Outlook framing suggests AI is pushing more leaders toward M&A, joint ventures, and alliances to acquire capabilities faster than internal build cycles allow. The survey signals a 2026 strategy pattern: treat AI as a transformation programand use transactions to buy time.

Meta outlined 2026 capex guidance of US$115bnUS$135bn, nearly doubling 2025 spend, with most funding aimed at data centres, servers, and networking for model training and deployment. Alongside the infrastructure push, Meta claims AI tooling is lifting engineering outputreporting a 30% increase in output per engineer since early 2025.

Synthesia raised a $200M Series E at a $4B valuation and is running a structured employee liquidity event tied to that same price. The company is positioning its enterprise avatar video platform as an on-ramp to AI agents for interactive training and internal knowledge retrievalan agent-shaped expansion that could deepen lock-in inside large organizations.

Harvey acquired Hexus, a startup that builds tools for producing product demos, videos, and guides, continuing Harvey's aggressive expansion amid intensifying legal-tech competition. The deal hints that legal AI vendors are investing not only in model capability, but in go-to-market enablementdocumentation, guided workflows, and adoption tooling that helps enterprises roll out AI safely.

LiveKit reached a $1B valuation, reflecting strong demand for real-time voice and communications infrastructure used by AI products. As voice agents move from demos to deployments, the infrastructure layerlow latency streaming, reliability, and developer toolingbecomes a primary battleground.

Inferact raised $150M to commercialize vLLM, underscoring how inference performance is now a front-line business problem, not a back-end optimization hobby. As model usage scales, teams are prioritizing throughput, latency, and cost predictabilityand vendors that package open tooling into enterprise-grade products can capture real budget.