Ricursive Intelligence is betting on AI-designed chips to help solve the compute crisis

techcrunch.com Oct 15, 2026

Ricursive Intelligence is betting on AI-designed chips to help solve the compute crisis

Curated by Shubham · TechCapsules · Oct 15, 2026

AI-summarised brief · reviewed before publication

Ricursive Intelligence’s founders, Anna Goldie and Azalia Mirhoseini, announced at TechCrunch Disrupt that their startup is using AI to design computer chips, aiming to address the projected 10,000‑fold increase in compute demand over five years. Their AI system, AlphaChip, learns from each iteration, producing unconventional layouts such as donut‑shaped and wave‑like designs that outperform traditional block‑based approaches. The company argues that rapid, automated chip design will reduce reliance on expanding data centers, improve wafer utilization, and enable custom chips for emerging applications.

💡 Why It Matters

· By automating chip design, Ricursive Intelligence could dramatically cut development time, allowing faster deployment of specialized hardware for fields like AR, VR, and bio‑computing, and potentially easing the strain on global manufacturing resources.

AI Capsules Semiconductors Startups

Read full article on techcrunch.com

Gudea leverages AI to help organizations prevent social media crises before they…

Gudea, a 2023‑founded startup, uses AI to map how rumors and misinformation spread across more than 500 online platforms, offering companies real‑time insight into who drives conversations and how quickly stories gain traction. Unlike traditional social‑listening tools that report only metrics, Gudea’s system analyzes behavior, relationships, and information pathways to help firms proactively protect reputations. The company presented its platform at TechCrunch Disrupt 2026, boasting nearly 40 clients, including Fortune 50 firms, major AI labs, and government agencies.

💡 By revealing the hidden structure of online narratives, Gudea equips decision‑makers with the context needed to intervene before crises erupt, a capability increasingly vital as generative AI lowers barriers to coordinated misinformation.

9to5mac.com · Oct 15

Apple Intelligence reportedly coming to China in December, but without Siri AI

Apple plans to launch Apple Intelligence in China in December as part of an iOS 27.3 update, but the rollout will omit the new Siri AI assistant and several other features. The move follows Chinese regulators’ approval of Alibaba’s AI models to power the service, after Apple secured permits in September. The Chinese version will include Genmoji, AI‑summarized notifications, writing tools, and a photo cleanup feature, while image‑generation tools remain unavailable. Apple’s partnership with Alibaba and Baidu reflects concessions to meet local AI content and security regulations.

💡 The delayed, pared‑down launch underscores how regulatory hurdles can stall even established tech giants, forcing Apple to tailor its AI offerings to fit China’s strict oversight. This strategy may influence the competitive balance in a market where domestic brands already dominate with advanced AI features.

techcrunch.com · Oct 15

Meet the five companies competing in the Startup Battlefield finals at Disrupt…

TechCrunch Disrupt 2026’s Startup Battlefield finals feature five finalists across four categories: Apate, Flox Intelligence, Intero Biosystems, Industrial Next, and Satlyt. Apate deploys AI agents to lure scammers, Flox Intelligence uses edge devices to deter wildlife, Intero Biosystems develops multi‑cell intestinal organoids for drug testing, Industrial Next offers AI‑driven robotic stations that retool production lines in minutes, and Satlyt creates satellite‑based shared computing for space missions. The finals begin Thursday at 11:30 am PT, with the winner receiving a $100,000 equity‑free prize.

💡 The finalists showcase breakthroughs that could reshape cybersecurity, wildlife management, pharmaceutical research, manufacturing automation, and space operations, illustrating the breadth of innovation driving the next wave of tech entrepreneurship.

quantumcomputingreport.com · Oct 15

Bloq Quantum Launches Bloq Maya AI Agent to Streamline Enterprise Quantum Algorithm…

Bloq Quantum has rolled out Bloq Maya, an AI‑driven assistant that automates quantum circuit design, algorithm formulation and code generation for enterprise users. Integrated into Bloq’s unified low‑code workspace, Maya translates high‑level business requirements into executable quantum programs, parameterized quantum circuits and QUBO models. The tool also configures development environments, aiming to shorten R&D cycles for sectors such as financial risk modeling, portfolio optimization, logistics and drug discovery. Bloq Maya expands the company’s enterprise software suite, which already includes a low‑code algorithm platform and partnerships with workforce and education providers. The launch was announced by Sreekuttan L S. on LinkedIn and is documented in the GQI Portal’s quantum industry database.

💡 By turning abstract use cases into ready‑to‑run quantum code, Maya lowers the expertise barrier, accelerating corporate adoption of quantum computing for high‑value problems.

semiconductor-digest.com · Oct 15

Eliyan Introduces NuLight Optical Interconnect Chiplets

Eliyan unveiled its Nexus™ and Oscar™ NuLight™ chiplets at the OCP Summit 2026, marking the first 2D microVCSEL interconnect platform that surpasses copper’s bandwidth and reach limits for AI systems. The Nexus chiplet offers a 1.6 T reverse gearbox and 212 G PAM4 SerDes, while Oscar provides laser‑driver and TIA interfaces for microVCSEL arrays. Together they deliver 50‑meter optical reach, low power, high reliability, and on‑the‑fly lane repair, enabling scalable AI compute connectivity across racks.

💡 By breaking the copper wall, Eliyan’s microVCSEL solution unlocks higher bandwidth and longer distances without sacrificing reliability, directly addressing the growing demand for efficient AI data‑center interconnects.

techcrunch.com · Oct 15

Figma announces new features to define how agents work on its platform

Figma unveiled a Markdown-based editor, Figma Specs, at its inaugural Config event in India, enabling teams to codify design languages and reference design, code, and library assets. The editor can exist as a standalone file or within the canvas, linking to FigJam diagrams and code layers. The company also released a public beta of its code layers feature, allowing AI agents to convert designs into code or modify code for prototypes. These tools aim to standardize brand rules and steer AI toward predictable outcomes.

💡 By centralizing design rules and automating code generation, Figma tightens consistency across large teams, reducing manual hand‑offs and accelerating iteration cycles.