Labor as a Service: The $30T Market for Work – Issue #21
Spotlight: Labor as a Service – The $30T Market for Work
For the past few years, Gen AI has transformed how we process information. Words, images, code.
But the physical world is different.
Robotics has produced impressive demonstrations for decades, yet most systems break down the moment they leave controlled environments. The gap between a perfect lab demo and a messy production floor has been the unsolved problem in robotics.
Rhoda AI is crossing that gap.
Instead of training robots on narrow lab data, Rhoda pre-trains models on internet-scale video to build a deep understanding of motion and physics, then adapts those systems inside real industrial environments. The result is closed-loop systems that continuously observe, predict, act, and adapt in real time.
Global manual labor represents roughly $30 trillion annually. Capital is abundant, but labor is finite. When factories cannot hire enough workers, production stalls. Rhoda’s “Labor as a Service” model targets the repetitive work that goes unfilled across manufacturing and logistics, unlocking more output from the humans who perform it.
The broader signals this week reinforce the same shift. NVIDIA is investing billions to expand AI infrastructure, including a $2 billion investment in Nebius and a gigawatt-scale compute partnership with Thinking Machines Lab. Equinix launched a Distributed AI Hub connecting enterprise AI infrastructure globally, while NTT DATA introduced NVIDIA-powered “AI factories” to operationalize production deployments. Microsoft is expanding Copilot into workflow execution across Microsoft 365, while OpenAI is strengthening the reliability and security of agent systems.
The shift is becoming hard to miss: AI is moving beyond chat interfaces and into real systems of work across factories, infrastructure, and enterprise systems.
Here is your Saturday guide to the signals shaping the future of AI: 👇
Signals Shaping the Future of AI:
Infrastructure
NVIDIA to invest $2 billion in Nebius to expand AI cloud capacity. The company will back large-scale data center buildouts and give Nebius early access to NVIDIA’s latest compute platforms, directly financing AI infrastructure growth. Click here
AMC Robotics partners with Hive Digital to run AI robotics workloads on GPU cloud infrastructure. AMC will use Hive’s compute capacity to develop, test, and deploy autonomous robotics systems for security, logistics, and industrial operations. Click here
Equinix launches Distributed AI Hub to connect enterprise AI infrastructure across data centers. The platform lets companies link models, GPU clouds, data platforms, and security tools through Equinix’s global network, with built-in AI security from Palo Alto Networks. Click here
NTT DATA launches NVIDIA-powered “enterprise AI factories” for large-scale AI deployment. The initiative integrates NVIDIA GPUs, networking, and AI software to help enterprises build and run production AI systems across cloud, data center, and edge environments. Click here
NVIDIA inks gigawatt-scale compute deal with Thinking Machines Lab. The multiyear partnership will deploy at least 1gigawatt of Vera Rubin systems for frontier AI training, alongside a strategic investment from NVIDIA. Click here
Broadcom unveils Taurus optical DSP for AI data centers. The new chip supports 400G per lane and 1.6T pluggable modules, targeting bandwidth constraints inside large AI clusters. Click here
Enterprise
Anthropic’s Claude chatbot experienced a temporary outage with elevated error rates. Users reported login failures and slow responses across Claude.ai and Claude Code, though the API remained operational as the company deployed a fix. Click here
Microsoft Wave 3 expands Copilot into a full agent platform. New capabilities, including Copilot Cowork, deepen agentic functionality across Microsoft 365 for enterprise deployment. Click here
Anthropic survey finds 57% of organizations already deploying AI agents for multi-step workflows. The report shows agents moving beyond simple automation into cross-team processes, with 80% of organizations reporting measurable economic returns from agent deployments. Click here
NVIDIA prepares open-source enterprise agent platform “NemoClaw.” WIRED reports the company is launching a platform to help enterprises deploy AI agents with built-in security and privacy tooling, moving further up the stack. Click here
Microsoft launches Copilot Cowork to turn AI from chat into task execution across Microsoft 365. The system can plan and run workflows across apps such as Outlook, Teams, and Excel, with human approvals and enterprise controls. Click here
OpenAI acquires Promptfoo to bolster agent testing and security. The AI security platform will integrate into OpenAI Frontier, strengthening evaluation, red-teaming, and compliance workflows. Click here
OpenAI tops $25 billion in annualized revenue, signaling growing independence from Microsoft. The milestone intensifies speculation around a future mega-IPO and shifting power dynamics in their partnership. Click here
Capital Flows
Rivian spin-out Mind Robotics raises $500 million at a ~$2 billion valuation to build AI-powered factory robots. The startup aims to use Rivian’s manufacturing data to train more dexterous industrial robots, focusing on real production workflows rather than humanoid demos. Click here
Yann LeCun’s AMI Labs raises $1.03 billion to develop “world model” AI. The startup will focus on systems that learn from real-world data rather than text alone, aiming to advance a new class of AI beyond traditional LLMs. Click here
Meta is acquiring Moltbook, a Reddit-style social network for AI agents. The Moltbook team will join Meta’s AI division, with plans to build more secure agent experiences as Meta deepens its push into autonomous agents. Click here
Research
Anthropic finds limited evidence that AI has raised unemployment so far. Its new labor-market study shows the most AI-exposed jobs are growing more slowly, with early signs that hiring may be softening for younger workers. Click here
Humanity’s Last Exam sets a new benchmark for expert-level AI performance. The 2,500-question test was designed to remove anything current models could already solve, and early results show even the strongest systems still struggle. Click here
OpenAI shares guidance on resisting prompt injection attacks. The update outlines deployment patterns to secure agents that browse, use tools, and take actions. Click here
OPCW flags AI as a growing factor in chemical weapons risk. In a March 11 report, the Organisation for the Prohibition of Chemical Weapons warned AI could reshape chemistry research, security, and weapons verification frameworks. Click here
OpenAI publishes research on instruction hierarchy in LLMs. The work focuses on improving model reliability when handling competing directives in complex agent scenarios. Click here
Policy
U.S. proposes “American AI” clause for federal contractors. The GSA proposed new rules requiring contractors on federal Schedule contracts to use U.S.-developed AI systems and disclose AI used in government work. Click here
Anthropic sues the Trump administration over the AI blacklist. The company is challenging its “supply chain risk” designation and federal ban on Claude, arguing the move is unlawful retaliation over limits on military use. Click here
EU lawmakers reach a deal to amend the AI Act, extending high-risk compliance deadlines and backing a ban on non-consensual deepfakes. Committee votes are scheduled for March 18 as the bloc refines its AI rulebook. Click here
Global AI Strategy
China’s new five-year plan elevates AI and robotics. National policy priorities emphasize accelerating AI capability and advanced manufacturing across the economy. Click here
India plans a $10.8 billion semiconductor fund to scale domestic chip production. The move expands state support for chip design and manufacturing as New Delhi pushes to compete in advanced semiconductors. Click here
Ukraine to share battlefield drone data with allies to train military AI. Kyiv will provide anonymized combat datasets to improve autonomous systems while accelerating its own AI deployment. Click here
Microsoft plans to train 3 million Africans in AI as competition with China’s DeepSeek intensifies. The push includes new cloud investments and a Copilot distribution deal with MTN to expand adoption across the continent. Click here
Talent Signals
Each week, we spotlight key roles tied to the themes shaping this week’s AI headlines, connecting talent to the companies driving the news.
@Wonderful develops AI-native software designed to simplify and automate creative and business workflows. As SaaS platforms integrate AI directly into user experiences, Wonderful is growing across engineering, design, and go-to-market roles. Open roles are listed on its careers page. Click here
@Nscale builds large-scale AI cloud infrastructure designed for training and serving advanced models. The company recently raised $2 billion at a $14.6 billion valuation as demand surges for GPU compute and AI data centers powering enterprise and frontier AI systems. Open roles are listed on its careers page. Click here
@RhodaAI builds AI systems designed to automate complex enterprise workflows and help teams turn operational data into real-time insights. As organizations embed AI more deeply into day-to-day software systems, Rhoda AI is expanding across engineering, product, and go-to-market roles. Open roles are listed on its careers page. Click here
You can see all the opportunities at Mayfield-backed AI companies here, and across the broader ecosystem here.
Social Signals
The most important conversations in AI are unfolding across social media, where top voices are shaping the next wave of signals and strategy. Here are some of the top social signals and their takes from the past week.
Sundar Pichai (Click here) — “We trained a new flood forecasting model designed to predict flash floods in urban areas up to 24 hours in advance. To help address a flash flood data gap, we created Groundsource, a new AI methodology using Gemini to identify 2.6M+ historical events across 150+ countries.” In a post viewed 564K+ times, Pichai announces a new urban flash-flood prediction system now live in Flood Hub, alongside the open-sourcing of a large global dataset to support research. The effort combines frontier models with geospatial and historical event analysis to expand early warning coverage in densely populated cities.
Nav Toor (Click here) — “Stanford proved that ChatGPT tells you you’re right even when you’re wrong. Researchers tested 11 of the most popular AI models across 11,500 real advice-seeking conversations. Every single model agreed with users 50% more than a human would.” In a post viewed 9.5M+ times, Toor points to research showing that leading models consistently validate users, even when they describe manipulation or harmful behavior. In a study of 1,604 participants, those paired with more sycophantic AI became less willing to apologize or compromise, yet rated the flattering system as higher quality. For founders and product leaders, this is the tension. Engagement and retention may be rewarded by affirmation, but long-term trust requires challenge, judgment, and alignment with human values.
Yann LeCun (Click here) — “Unveiling our new startup, Advanced Machine Intelligence (AMI Labs). We just completed our seed round: $1.03B, one of the largest seed rounds ever, probably the largest for a European company. We’re hiring.” In a post viewed 2.5M+ times, LeCun announces AMI Labs, which aims to build AI systems with world models, persistent memory, reasoning, planning, and stronger controllability. The $1.03B seed round signals significant early capital backing behind next-generation architectures focused on long-term reasoning and safety.
