Try the free and instant web host for agents

Good morning. It's Wednesday, March 18, and we're covering Meta’s AI-driven restructuring, India’s outsourcing disruption, NVIDIA’s push into space computing, and more.

If you found this email useful, share it with someone who'd enjoy it — it's one of the best ways to support us.

NVIDIA Brings AI Computing to Space
NVIDIA unveiled new platforms bringing data-center-class AI to space, including the Space-1 Vera Rubin Module, IGX Thor and Jetson Orin. These systems enable real-time processing onboard satellites. Partners like Axiom Space and Planet are already deploying them for geospatial analysis and autonomous operations. The move supports faster insights and more independent spacecraft.

NVIDIA Launches Agentic AI Vera CPU
NVIDIA introduced the Vera CPU, designed specifically for agentic AI and reinforcement learning workloads. It delivers up to 50% faster performance and twice the efficiency of traditional CPUs, with hyperscalers including Meta and Alibaba Cloud collaborating on deployment. The chip supports large-scale AI “factories” with high concurrency and fast CPU-GPU data sharing. It signals a shift toward CPUs playing a central role in AI systems.

Senators Urge ByteDance to Shut Seedance 2.0
U.S. senators Marsha Blackburn and Peter Welch urged ByteDance to immediately shut down Seedance 2.0 over alleged copyright and likeness violations. The app enables users to generate videos featuring real people and licensed characters, prompting backlash from lawmakers and Hollywood groups. ByteDance paused its global rollout and is strengthening safeguards, as pressure mounts over AI intellectual property risks.

Anthropic Is Hiring a Weapons Expert to Curb Misuse
Anthropic is hiring a chemical weapons expert to prevent its AI from enabling harmful misuse, including potential weapons-related risks. The role reflects growing concern that advanced AI tools could be exploited despite safeguards. OpenAI has posted a similar role, signaling an industry-wide shift toward risk mitigation. Critics warn this approach may expose systems to sensitive knowledge even as firms try to control it.

here.now is web hosting for agents:

Perfect for anything you create with your agent and need a URL for: documents, websites, dashboards, tools, schoolwork, visualizations, presentations, prototypes, games, articles, galleries, portfolios, media files, etc.

Works inside your favorite agents like Claude Code, Codex, Cursor and OpenClaw.

Free, no sign-up required.

Get setup instructions for your agent at here.now

↓ Real or fake. Answer below.

Meta is considering new layoffs as it pours billions into AI infrastructure and talent. While the company still trails leaders like Google and OpenAI in top-tier models, analysts say its aggressive internal AI adoption could deliver a different kind of advantage.

Bernstein analyst Mark Shmulik argues the cuts may reflect real efficiency gains as Meta restructures around AI, not just cost pressure. If Meta can embed AI across its operations, it could widen margins and force competitors to follow. Early investor reaction was positive, with shares rising about 2%. → Read the full article here.

India’s $300bn IT outsourcing industry is facing mounting pressure as artificial intelligence tools threaten to automate core back-office tasks, triggering a roughly 20% drop in the Nifty IT index in 2026. The alarm intensified after new AI systems demonstrated capabilities in legal, compliance, and data workflows—areas central to India’s labor-driven services model.

Analysts expect a structural shift: lower revenues from routine maintenance work and greater focus on consulting and AI implementation, with growth potentially slowing to near zero after 2031 in a worst-case scenario. Still, major firms and banks argue AI will augment—not replace—the sector, creating new roles and partnerships even as entry-level hiring weakens.→ Read the full article here.

Tech companies and large enterprises are maintaining — and in some cases expanding — relationships with Anthropic despite a U.S. Pentagon designation labeling the AI firm a supply chain risk. The Trump administration ordered federal agencies to phase out Anthropic’s Claude within six months, but private-sector partners including Google and multiple Fortune 500 firms say they are not changing contracts.

Ramp data suggests Anthropic is gaining share in enterprise AI deals, alongside indications of continued revenue growth and rising app downloads. The episode highlights how deeply embedded Anthropic’s tools have become in business workflows, making them costly to replace. → Read the full article here.

NVIDIA-Backed Startup Eyes Korea Hub (Paywall): Reflection AI plans a 250-megawatt data center with Shinsegae to build advanced AI models, aligning with U.S. efforts to counter China’s tech influence.

Unsloth Launches Unsloth Studio: Open-source web UI runs locally on Mac, Windows, and Linux, training 500+ models up to 2x faster with 70% less VRAM and broad model support.

Alibaba Debuts Wukong AI Platform: New enterprise tool manages multiple AI agents for tasks like research and approvals, with planned integrations across Slack, Teams, WeChat, and DingTalk.

Pentagon Explores Anthropic Alternatives: After their dramatic falling-out with Anthropic, the U.S. Department of Defense is reportedly developing backup AI systems beyond Anthropic.

Google Expands Personal Intelligence Access: Now available to all U.S. users, it connects Gmail and Photos to tailor AI responses across Search, Gemini app, and Chrome, with opt-in controls.

The Image Is AI. Tune in next week for another Real or AI test.

Thanks for reading. See you next time!

— Matthew Berman, Nick Wentz & the Forward Future Team