Samwise Tech/AI/Robotics Newsletter
Thursday, April 30, 2026
Microsoft’s AI Revenue Hits $37B Run Rate as Nadella Moves to “Exploit” Revised OpenAI Deal
Microsoft CEO Satya Nadella said Wednesday the company is “ready to exploit” its newly renegotiated deal with OpenAI, calling the updated terms a clear win for both parties. The revised agreement grants Microsoft royalty-free, nonexclusive access to OpenAI’s models and products through 2032, while ending the previous exclusivity arrangement that had drawn legal scrutiny following OpenAI’s $50 billion Amazon deal. Microsoft’s AI business now runs at a $37 billion annual rate, a 123 percent year-over-year gain. The company retains its 27 percent stake in OpenAI’s for-profit entity and holds commitments for OpenAI to purchase more than $250 billion in Microsoft cloud services over the agreement’s term.
Sources: TechCrunch
Google Cloud Tops $20B in One Quarter for First Time, but CEO Says AI Demand Is Still Outrunning Capacity
Google Cloud topped $20 billion in quarterly revenue for the first time, driven by surging enterprise demand for Gemini AI models and infrastructure services, CEO Sundar Pichai announced Wednesday. The figure represents a 63 percent year-over-year increase, yet Pichai warned the company is “compute constrained in the near-term” — indicating growth would have been higher still without supply limitations. Cloud backlog doubled to $462 billion, new customer acquisition doubled year-over-year, and customers outpaced initial commitments by 45 percent quarter-over-quarter. Alphabet updated its 2026 capital expenditure guidance to $175–185 billion, reflecting the scale of AI infrastructure investment needed to meet persistent demand.
Sources: TechCrunch
Amazon Launches Three OpenAI Products on AWS Bedrock, Including Exclusive Agent-Development Tool
One day after Microsoft and OpenAI formalized their restructured partnership, Amazon Web Services moved quickly, launching three new OpenAI offerings on its Bedrock platform. AWS customers now have access to OpenAI’s latest frontier models and its Codex code-writing service. The two companies also co-developed a stateful runtime technology for building AI agents on Bedrock, with AWS securing exclusive rights to distribute OpenAI’s new agent-development tool, Frontier. The arrangement gives Amazon a strong foothold in enterprise AI agent infrastructure as cloud providers compete aggressively for a share of growing corporate spending on AI services, particularly agentic applications requiring persistent context and multi-step workflows.
Sources: TechCrunch
SoftBank Forms Roze AI to Build Data Centers with Robots, Eyes $100B IPO
SoftBank is assembling a new company called Roze AI that will use robotic construction systems to build data centers faster and more efficiently than conventional methods. The venture, disclosed Tuesday, combines SoftBank’s robotics portfolio — including assets from its 2025 acquisition of ABB’s robotics division — with its broader infrastructure ambitions. The company is already planning a $100 billion IPO, according to sources familiar with the matter. Roze AI targets the bottleneck created by surging AI compute demand outpacing construction capacity. SoftBank has identified robotics, AI chips, AI data centers, and energy as its four core investment priorities, and Roze AI sits at the intersection of the first three.
Sources: TechCrunch
Apptronik Hires Waymo’s Former CPO to Lead $935M Humanoid Robot Company into Commercial Scale
Humanoid robot startup Apptronik appointed Daniel Chu, former chief product officer at Waymo, as its new CPO as the Austin-based company shifts from engineering research toward large-scale commercial deployment. Chu led Waymo’s product organization through the launch of the world’s first fully driverless ride-hailing service and brings proven experience scaling autonomous-vehicle hardware and software to millions of customer trips. Apptronik, which has raised $935 million at a $5 billion-plus valuation, is targeting industrial warehouses and eldercare facilities with its Apollo humanoid robot. The company said Chu’s hire represents the most significant leadership addition since completing its Series A, signaling a decisive push toward volume manufacturing.
Sources: The Robot Report
IBM Releases Bob AI Coding Platform Globally After Internal Pilot Reaches 80,000 Employees
IBM launched its AI-powered software development platform Bob to enterprise customers globally on Wednesday, expanding a product that began as a 100-person internal pilot in summer 2025 and now serves more than 80,000 IBM employees. Bob orchestrates multiple AI models through a unified interface with human-in-the-loop checkpoints designed to flag insecure or incorrect code before it reaches production systems, directly addressing concerns that AI-generated code frequently introduces security vulnerabilities. IBM claims Bob delivers a 45 percent productivity gain across software development workflows. The platform competes with GitHub Copilot, Cursor, and a growing field of AI coding assistants as enterprises seek production-safe alternatives to consumer-grade tools.
Sources: VentureBeat
Meta Posts $4 Billion Reality Labs Loss in Q1 as Hardware Bets Continue to Drain Cash
Meta reported a $4 billion operating loss from its Reality Labs division in the first quarter of 2026, maintaining the steep pace of hardware investment that has cost the company roughly $60 billion since 2020. The losses span its AR glasses, VR headsets, and VR software portfolio. Meta’s overall capital expenditure guidance for 2026 stands at $125–145 billion, exceeding earlier analyst estimates. CFO Susan Li acknowledged the company repeatedly underestimates its compute requirements. Meta stock fell more than 5 percent in after-hours trading following the earnings announcement. Despite the losses, Meta frames Reality Labs spending as foundational to its long-term position in spatial computing and next-generation AI hardware platforms.
Sources: TechCrunch
Tech Pulse
Top Frontier Models: Claude Mythos Preview (93.9%) | Claude Opus 4.7 (87.6%) | GPT-5.3 Codex (85.0%)
Top Open Source Models: MiniMax M2.5 (80.2%) | GLM-5 / Zhipu AI (77.8%) | DeepSeek V3.2 (73.0%)
Top Small Models (15–50B): Mistral Small 4 24B (72.1%) | Qwen3.6-27B (70.3%) | Gemma 3 27B (67.5%)
Top Edge Models (0–15B): Phi-4 mini 3.8B (62.4%) | Qwen 2.5 7B (58.1%) | Llama 3.2 3B (54.7%)
AI Leaders: NVIDIA $4.33T | Alphabet $4.05T | Microsoft $3.83T
Robotics Leaders: Intuitive Surgical ~$160B | Fanuc ~$38B | Teradyne ~$18B
Curated by JD · samwise.agency
Leave a Reply
You must be logged in to post a comment.