Samwise Tech/AI/Robotics Newsletter
Saturday, May 30, 2026
Anthropic Raises $65B at $965B Valuation Ahead of IPO
Anthropic has closed a $65 billion Series H round at a $965 billion post-money valuation, positioning the AI safety company for a near-term IPO. The raise was co-led by Altimeter Capital, Sequoia Capital, and Dragoneer, with Samsung, SK Hynix, and Micron joining as new strategic investors. Amazon contributed $5 billion of a broader $15 billion hyperscaler tranche. Anthropic’s annualized revenue run rate has now crossed $47 billion, with analysts projecting a 130 percent surge and the company’s first operating profit. In a letter accompanying the announcement, Anthropic noted that Mythos-class capabilities would reach all Claude customers within “coming weeks.”
Sources: TechCrunch
XCENA Raises $135M Betting AI’s Real Bottleneck Is Memory, Not Compute
South Korean startup XCENA has raised $135 million in a Series B round at a $570 million valuation, bringing its total funding to $185 million. The company argues that memory — not compute — is artificial intelligence’s true infrastructure bottleneck. Its flagship chip, the MX1, connects to host CPUs via the CXL interconnect and claims to consolidate the workload of ten servers into one. The round was co-led by Atinum Investment and IMM Investment, with Samsung contracted to handle foundry production by the end of 2026. XCENA maintains offices in Pangyo, South Korea, and Sunnyvale, California, employs more than 90 people, and expects first revenue in 2027.
Sources: TechCrunch
Anthropic Releases Claude Opus 4.8 with Dynamic Workflows Feature
Anthropic released Claude Opus 4.8 on May 28, just 41 days after Opus 4.7, maintaining the same pricing tier. The model’s headline feature is Dynamic Workflows, currently in research preview, which enables Claude to orchestrate hundreds of parallel subagents simultaneously for complex tasks. Opus 4.8 also introduces behavioral changes: Anthropic says the model is “more likely to flag uncertainties” and “less likely to make unsupported claims,” a shift that Bridgewater Associates praised in early testing. On SWE-bench Pro, Opus 4.8 scores 69.2 percent, ranking second among commercially available models. Anthropic’s most capable Mythos model remains withheld from general availability.
Sources: TechCrunch
Groq Raising $650M After Nvidia’s $20B Chip Technology Deal
AI chip startup Groq is raising $650 million from existing investors as it pivots to its inference neocloud business, per Axios. The move follows a December agreement with Nvidia — described as a “not-an-acquisition” deal — under which Nvidia paid a reported $20 billion to license Groq’s chip technology; Groq’s investors were paid out in cash. The new round is led by interim CEO Adam Winter and CFO Matt Eng. Backers Disruptive and Infinitium have agreed to fill the round should other investors decline their pro-rata shares. Inference — the processing that occurs after an AI prompt is submitted — is more in demand than model training, the company says.
Sources: TechCrunch
Coders Won’t Work Without AI — But Research Warns It May Hurt Them
A growing body of research suggests AI coding tools may be accelerating technical debt. In February 2026, AI lab METR attempted to update its 2025 productivity study but found developers refused to participate “because they do not wish to work without AI.” A May 2026 METR survey found developers self-reporting a twofold productivity gain — but independent findings complicate the picture. Amazon shut down an internal AI-use leaderboard called Kirorank after employees gamed it by running up costs, the Financial Times reported. Uber exhausted its entire 2026 AI budget in four months. Singapore Management University researchers separately warned that “AI-generated code can introduce long-term maintenance costs into real software projects.”
Sources: TechCrunch
Apple’s Redesigned Siri in iOS 27 to Rival ChatGPT, Powered by Google Gemini
Leaked renders published by Bloomberg reveal that Apple’s upcoming iOS 27 includes a dramatically redesigned Siri that emerges from the iPhone’s Dynamic Island as a standalone app positioned to compete directly with ChatGPT, Claude, and Gemini. The new Siri leverages Google Gemini to power AI-driven search, replacing the traditional swipe-down interface with a card-style results system. Apple’s pitch: a privacy-trusted AI assistant with access to 2.5 billion installed devices, compared with ChatGPT’s 900 million weekly active users. The redesign is expected to be formally unveiled at WWDC in June, ahead of iOS 27’s fall release.
Sources: TechCrunch
Cognition’s Devin Ships 89% of Its Own Code as CEO Resists Human-Replacement Narrative
Cognition CEO Scott Wu, whose AI coding agent startup raised $1 billion at a $26 billion valuation this week, says the company’s goal has never been to replace human programmers. Cognition makes Devin, one of the first commercial AI coding agents; the company says 89 percent of the code committed by its own engineers was committed by Devin, with the remainder handled by local agents in Windsurf, the AI coding tool Cognition acquired last year. Wu describes Devin as operating “somewhere between a junior and a mid-level engineer,” depending on the task. He envisions agents freeing programmers from maintenance toil so humans can focus on software architecture and security design.
Sources: TechCrunch
Tech Pulse
Top Frontier Models (SWE-bench Pro): Claude Mythos Preview (77.8%) | Claude Opus 4.8 (69.2%) | Claude Opus 4.7 Adaptive (64.3%)
Top Open Source Models (SWE-bench Verified): DeepSeek V4-Pro (80.6%) | GLM-5 (77.8%) | Qwen 3.6-27B (77.2%)
Top Small Models (15–50B, SWE-bench Verified): Qwen 3.6-27B (77.2%) | Phi-4 14B (top reasoning) | Gemma 3-27B (production-grade)
Top Edge Models (0–15B, HumanEval): Llama 4 Scout (88%+) | Phi-4 14B (14B, on-device) | Gemma 3-9B (multilingual)
AI Leaders: NVIDIA (~$5T) | Microsoft (~$4T) | Alphabet (~$3.5T)
Robotics Leaders: ABB (~$190B) | Intuitive Surgical (~$153B) | Fanuc (~$35B est.)
Curated by JD · samwise.agency
Leave a Reply
You must be logged in to post a comment.