AI news, 17 August: Stripe seals over-$7B deal for OpenRouter AI gateway

Stripe finalizes OpenRouter acquisition. Payments giant Stripe has finalized a deal to acquire OpenRouter, the popular AI model gateway and marketplace, for more than $7 billion, according to Bloomberg and TechCrunch reports on 16 August. OpenRouter lets developers route requests across more than 400 models from providers including OpenAI, Anthropic, Google, Meta and DeepSeek; it serves roughly 8 million users and has processed vast volumes of tokens. The price marks a sharp jump from the startup’s $1.3 billion valuation in a May 2026 funding round. The move, following Stripe’s earlier purchase of metering firm Metronome, positions the company to control key layers of model selection, usage tracking and billing as AI agents become more common in business software. Stripe declined to comment on the reports.

Anthropic CEO frames AI backlash as trust crisis. In a series of posts on 16 August, Anthropic chief executive Dario Amodei argued that public skepticism toward AI is “fundamentally a crisis of trust” rather than mainly the result of industry warnings about risks. Ordinary people, he said, often suspect companies and governments of finding new ways to take advantage of them. Amodei acknowledged that AI firms, including his own, “haven’t yet delivered on our big promises to benefit the world” and that marketing claims such as curing cancer ring hollow; real progress on major problems like cancer treatment is what would rebuild confidence. He defended carefully designed regulation that can address safety risks while constraining the power of the largest players and still leaving room for open-weight models.

Google begins shutting down Imagen 4 image APIs. Google is retiring three Imagen 4 model endpoints (standard, ultra and fast) with an earliest shutdown date of 17 August 2026, directing developers to migrate to Gemini image generation instead. The change is not a simple drop-in replacement: image creation now runs through a different method, and teams must re-test prompts, aspect ratios, watermarks, latency and quotas. The shift continues Google’s consolidation of generative tools under the Gemini family.

Young Americans largely distrust AI tech leaders. A CNBC/Generation Labs poll of more than 1,000 Americans aged 18–34 found high levels of distrust toward prominent AI executives. Large majorities said they do not trust figures including Palantir’s Alex Karp, Peter Thiel, Mark Zuckerberg, Elon Musk and Sam Altman to act responsibly on AI; Microsoft’s Satya Nadella was the relative exception. Nearly half worried AI would harm their careers, and 60 percent wanted slower data-center expansion.

Nvidia–OpenAI data-center talks continue in background. Reports from mid-August indicate Nvidia remains in discussions around a possible multi-billion-dollar investment linked to a large planned Ohio data-center project for OpenAI involving SoftBank’s SB Energy, alongside broader credit-support arrangements. No final agreement has been confirmed.