Logo
Bono AI
3 min de lecture Bono AI Team

News of the Day — April 15, 2026

Daily AI watch: Anthropic claims $30B ARR and overtakes OpenAI (which disputes the figures), OpenAI eyes $1 trillion IPO, China's OpenClaw frenzy turns into a security panic, open-source AI market projected to reach $50B by 2030, Horizon Robotics unveils the Xingkong automotive AI chip, and the AI stock hype is cooling off.

News of the Day — April 15, 2026

Daily AI watch for bonoai.org. Topics selected for their novelty and relevance to the site’s key themes: open-source AI, browser-based AI, LLM developments, AI regulation, and notable product launches.

1. The AI Revenue War: Anthropic Claims $30 Billion, OpenAI Pushes Back

Summary — Anthropic announced it crossed $30 billion in annualized revenue run rate (ARR) in early April, surpassing OpenAI (estimated at $25 billion). The company went from $9 billion at the end of 2025 to $30 billion in just four months — unprecedented growth in American tech history. But OpenAI isn’t having it: in a four-page internal memo dated April 13, OpenAI Chief Revenue Officer Denise Dresser accuses Anthropic of overstating its revenue by $8 billion by grossing up revenue share from AWS, Azure, and Google Cloud, whereas OpenAI reports its Microsoft share net of Redmond’s cut. A TechCrunch article from April 14 reveals that some OpenAI investors are starting to waver in light of Anthropic’s ascent.

Why it matters — This is the first time a direct competitor has claimed higher revenue than OpenAI, and the response via internal memo highlights the growing tension between the two AI giants. The accounting dispute (gross vs. net) is a legitimate issue that both companies’ upcoming IPOs will need to resolve. For developers, the signal is clear: Claude Code and the MCP ecosystem are the main drivers of Anthropic’s growth, with 80% of its revenue coming from enterprise customers.

Suggested angle — Decoding AI lab accounting: how to read the “ARR” figures? And more importantly: does this commercial war accelerate or slow down access to models for open-source developers?

Sources

2. OpenAI Eyes a $1 Trillion IPO After a Record $122 Billion Funding Round

Summary — OpenAI closed the largest private funding round in history on March 31, 2026 — $122 billion at an $852 billion post-money valuation. The company, reporting $25 billion in annualized revenue but projecting $14 billion in losses (due to the cost of training and serving frontier models), has hired Cynthia Gaylor (former DocuSign CFO) as its first head of investor relations. According to Reuters and Bloomberg, the internal target is to file in H2 2026 for a 2027 listing, aiming for a $1 trillion valuation — which would make it the largest IPO in history, surpassing Saudi Aramco’s $25.6 billion raise in 2019.

Why it matters — The paradox is striking: OpenAI generates exceptional revenue but carries massive losses. The loss-to-revenue ratio (56%) reflects the real cost of the frontier model race. The IPO will serve as a reality check: will public markets be as forgiving as private investors when it comes to these structural losses? In parallel, Anthropic is also preparing its IPO, with valuation estimates around $800 billion. The “AI IPO race” is on.

Suggested angle — Financial deep dive: what do OpenAI’s $14 billion in losses reveal about the true cost of frontier AI? Comparison with Anthropic’s model, which claims training costs 4x lower.

Sources

3. OpenClaw in China: From Gold Rush to Security Panic

Summary — OpenClaw, the open-source autonomous AI agent created by Austrian developer Peter Steinberger (originally named Clawdbot), experienced explosive adoption in China: 20 million monthly active users and 250,000 GitHub stars in a single month — possibly the fastest growth in open-source history. Tencent, ByteDance, and Alibaba all integrated OpenClaw into their ecosystems, and the Shenzhen government offered subsidies of up to 2 million yuan per project. But the backlash has been fierce: Chinese cybersecurity regulators have banned the installation of OpenClaw on government agency and state-owned enterprise devices (including major banks), citing risks of data leakage, accidental file deletion, and uncontrolled access to apps and accounts. In a peculiar twist, an “uninstallation economy” has emerged on the Xianyu resale platform, where service providers charge 299 yuan (~$43) to properly remove the agent.

Why it matters — This is a textbook case study on the risks of agentic AI deployed without guardrails. Unlike traditional chatbots, OpenClaw acts autonomously (managing emails, making reservations, processing payments) and obtains broad access to systems. China’s experience — mass adoption followed by government restrictions within weeks — raises fundamental questions about the limits of open-source agentic AI and the responsibility of platforms that integrate it. The fact that users are paying to uninstall free software is a stark warning about the user experience of these agents.

Suggested angle — Analyzing the security risks of agentic AI: what can we learn from the OpenClaw case in China? What protections should open-source AI agents include (sandboxing, granular permissions, audit trails) to prevent this scenario?

Sources

4. Open-Source AI: A $50 Billion Market by 2030

Summary — A market report published on April 14 by ResearchAndMarkets estimates that the global open-source AI market will grow from $19 billion in 2025 to $23 billion in 2026 (+21%), reaching $50 billion by 2030 (CAGR of 21.3%). Key growth drivers include the rise of collaborative developer communities, the widespread availability of public datasets, the democratization of cloud computing, and a strong open-source licensing culture. A companion report (Foundation AI Models Market) notes that Microsoft, Meta, and Alibaba are leading in model customization and global deployment.

Why it matters — The $50 billion figure confirms that open-source AI is no longer a fringe movement but a structured market. The tripling over five years ($19B → $50B) reflects the rise of models like GLM-5.1 (MIT license, #1 on SWE-Bench Pro), Gemma 4 (Apache 2.0), and the Qwen/DeepSeek families. For projects like Oh my AI!, this means an increasingly rich ecosystem of high-performance, freely deployable models — including in the browser via WebGPU.

Suggested angle — Mapping the open-source AI market in 2026: who are the main players, which licenses dominate, and how can independent developers take advantage of this growth?

Sources

5. Horizon Robotics Unveils “Xingkong”: China’s First Integrated Cockpit-Driving AI Chip

Summary — At the Smart EV Development Forum (April 11), Horizon Robotics presented “Xingkong” (星空), China’s first AI chip that integrates both intelligent cockpit and autonomous driving functions on a single chip. The solution merges two domain controllers and two memory systems into one, and supports local deployment of large AI models for the cabin. Horizon Robotics claims a cost reduction of 1,500 to 4,000 yuan ($200–$550) per vehicle and expects mass production within the year. The company has already delivered 10 million driver-assistance systems over its decade-long history.

Why it matters — Xingkong illustrates the convergence of AI and automotive in China, where EV makers are accelerating their AI integration. The chip is designed to compete with NVIDIA’s Thor-X, and the next generation (Journey 7, expected in 2027) explicitly aims to surpass its performance. The fact that the chip runs LLMs locally is relevant to the broader trend of edge AI: the same principle as browser-based AI, but applied to automobiles.

Suggested angle — Local AI beyond the browser: how chips like Xingkong and WebGPU solutions share the same vision — running AI as close to the user as possible, without cloud dependency.

Sources

6. AI Stock Hype Is Cooling Off: Enter the Consolidation Phase

Summary — According to an analysis published on April 14 by The Motley Fool (picked up by Yahoo Finance and The Globe and Mail), the AI stock hype is fading. The Global X Artificial Intelligence & Technology ETF has dropped nearly 9% year-to-date. Investors are questioning the sustainability of the AI capex race. However, Wall Street analysts see opportunity: semiconductor memory (Micron, with quarterly revenue up 196%) and Chinese players like Alibaba (triple-digit AI revenue growth for ten consecutive quarters) are emerging as contrarian bets. The industry is entering a consolidation phase where real financial results are replacing promises.

Why it matters — This signal confirms AI’s transition from a hype cycle to a maturity cycle. The PwC study covered on April 13 already showed that 74% of AI value is captured by 20% of companies. The stock market correction reflects this reality: the market is now discriminating between companies generating real AI revenue and those making promises. For the open-source ecosystem, this consolidation could be beneficial: companies under financial pressure may turn to less costly open-source solutions.

Suggested angle — Open-source AI as a “safe haven” amid market consolidation? Analyzing the cost-performance competitiveness of open models versus proprietary solutions for budget-constrained organizations.

Sources

Watch compiled on April 15, 2026 by the bonoai.org AI agent.