Lovable in Funding Talks at a $13.2 Billion Valuation — The Rapid Rise of No-Code AI Development
Lovable, a no-code AI app development platform, is reportedly in talks to raise a funding round at a valuation of around $13.2 billion (roughly ¥1.9 trillion). The round is expected to bring in approximately $300 million, led by Menlo Ventures.
Founded in 2024, Lovable is a relatively young startup, but it has been rapidly winning users by simplifying application development with AI. If this round closes, its valuation would roughly double from its most recent one.
The no-code/low-code tool market is being reinvigorated by the rise of AI, and beyond Lovable, players such as Cursor and Replit compete in adjacent territory. The influx of capital suggests that investors see this market as a "real growth opportunity."
Modal's CTO on Infrastructure Evolution for the "Agent Experience"
Akshat Bubna, co-founder and CTA of cloud computing platform Modal... co-founder and CTO of cloud computing platform Modal, recently shared his thoughts on how infrastructure needs to evolve for the age of AI agents. The company recently closed a $355 million Series C round.
The core message: while traditional cloud infrastructure was designed "for developers," it now needs to be redesigned "for agents." Agents read documentation, call APIs, and handle errors differently than humans do, so legacy YAML-based configuration and human-oriented dashboards are increasingly falling short.
Modal has surveyed a number of providers in the "Agent Cloud" space — including Databricks, Daytona, and Railway — and argued that designing how agents autonomously consume resources, scale, and get billed will become a central infrastructure challenge. It's clear that infrastructure-side preparation is accelerating as AI agents move toward mainstream adoption.
LayerX Redefines AI Budgets as a "Second Payroll" — Every Employee's AI Usage Is Public
LayerX, a Japanese startup, has shared a distinctive approach to managing AI costs. The company publishes every employee's AI usage internally — the CEO's included — practicing radical transparency.
What stands out is that even when a team's usage exceeded its departmental budget by more than 10x, management did not reprimand the people on the ground. Behind this lies a strategy of treating AI costs not as a mere "expense" but as a "second payroll" to be converted into an organizational asset. In other words, budgets are managed against a benchmark of replacing outsourcing spend with AI usage, and the company envisions a "future org chart" in which one manager supports twenty reports.
AI cost governance is a shared challenge at many companies, but tightening the screws too hard risks discouraging AI adoption on the front lines — which would defeat the purpose. LayerX's example points to one way to combine transparency with an investment mindset.
Actenon — An Attempt to Standardize "Proof of Action" for AI Agents
Actenon, an open-source project, has been released. It provides a mechanism for proving that the decisions and actions of AI agents rest on a legitimate process.
The project aims to offer a neutral standard of proof for the behavior of consequential AI agents — those whose actions carry significant outcomes. When AI agents act autonomously in domains with major human impact, such as financial transactions or legal judgment, an auditable decision-making process is the foundation of trust.
As AI agents move into real-world use, building governance and accountability frameworks cannot wait. Standardization efforts like Actenon look likely to become a key piece of the puzzle in bringing agents into society at large.
Nika — A New Approach to Designing AI Workflows with "Intent as Code"
Nika, an open-source tool for building AI workflows under the paradigm of "Intent as Code," has been released.
Traditional workflow construction requires explicitly spelling out every step. Nika instead lets developers declare their "intent" declaratively, providing a mechanism whereby AI agents autonomously plan and execute. Just as Infrastructure as Code (IaC) revolutionized cloud management, AI workflows are now calling for a similar layer of abstraction.
Enter the Token Market — Mtok.market for Spot-Trading AI Inference Costs
Mtok.market, a platform for trading AI inference tokens on a non-custodial spot market, has launched. It aims to let supply and demand determine AI inference costs, providing an efficient price-discovery mechanism.
AI inference costs swing widely depending on use case and time of day, yet fixed-price API models remain the norm. If a spot market for tokens works, it would broaden the options for cost optimization, potentially opening up significant savings for companies that use AI at scale. That said, whether the market can actually take hold — liquidity, price volatility risk, and so on — remains to be seen.
---