ContentBuffer
AI spending is expected to hit $4.25 trillion in 2026, with 74% of enterprise IT pros planning to expand their AI budgets. However, only 45% of AI pilots make it to full production, a significant improvement from last year's 5%. This 'fast in, fast out' dynamic means AI startups must constantly prove their value, with over 50% of buyers preferring outcome-based pricing. This uncertainty is reshaping the AI startup landscape, making long-term revenue commitments less reliable.
Tesla is launching the Cybercab, a gold two-seater sedan with no steering wheel or pedals, in Austin, Texas. This vehicle is designed to use AI and cameras to navigate the world, changing the landscape of autonomous driving. For developers and tech enthusiasts, the Cybercab represents a significant step towards a future where autonomous vehicles are commonplace, potentially impacting urban planning and transportation infrastructure. As of the article's publication, 420 autonomous vehicles were registered in Texas, and Tesla's fleet has consisted of Model Y SUVs, which are harder to distinguish on the road than other vehicles. The Cybercab aims to make it harder to miss failures when they happen, ensuring transparency and accountability in the robotaxi network. Let's dive deeper…
In today's ContentBuffer update:
Thinking Machines Eyes $1B Fundraise at $40B Valuation
Fusion Power Plant Deal in Wisconsin
Abliteration.ai Launches Guardrail-Free AI Model Service
OpenAI Unveils Astra: The Most Capable AI Model Yet
5 new AI tools & 5 new AI jobs
More tech news
Latest Development
Image source: techcrunch.com
Summary: AI budgets are growing, but the fast-in, fast-out dynamic means startups must constantly prove their worth. Fewer than half of pilots make it to full production, and pricing models are evolving to tie fees to outcomes.
Key Points:
AI spending to hit $4.25 trillion in 2026, with 74% of enterprise IT pros planning to expand AI budgets.
Only 45% of AI pilots make it to full production, up from 5% last year, indicating improving success rates.
Over 50% of technical buyers want AI fees tied to work produced or outcomes, not just usage.
Enterprise trial budgets fueled the initial AI boom, but this year was supposed to see long-term commitments.
AI startups face uncertain revenue even after products graduate from pilot phases and get adopted by companies.
Why it matters: If you're an AI startup, the fast-in, fast-out dynamic means you must constantly prove your product's value. Over 50% of technical buyers want fees tied to outcomes, not just usage. This shifts the focus from SaaS-era usage models to outcome-based pricing, making long-term revenue commitments less reliable.
Image source: techcrunch.com
Summary: Tesla launches the Cybercab in Austin, a gold two-seater sedan designed for autonomous driving. This could be a game-changer for the robotaxi market.
Key Points:
Tesla is launching the Cybercab, a gold two-seater sedan with no steering wheel or pedals, in Austin, Texas.
The Cybercab is designed to use AI and cameras to navigate the world, aiming to revolutionize the robotaxi market.
As of the article's publication, 420 autonomous vehicles were registered in Texas, according to the state's automated vehicle tracker.
Tesla's fleet has consisted of Model Y SUVs, which are harder to distinguish on the road than other vehicles.
The Cybercab aims to make it harder to miss failures when they happen, ensuring transparency and accountability in the robotaxi network.
Why it matters: If you're working on autonomous vehicle projects, the Cybercab's launch in Austin could be a pivotal moment. The vehicle's design and advanced AI capabilities could set new standards for the industry. However, the success of the Cybercab will depend on its ability to handle real-world scenarios and maintain transparency in its operations.
Image source: techcrunch.com
Summary: Thinking Machines is in talks to raise $1B at a $40B valuation, down from its $50B target. Investors are backing the round despite high-profile departures, valuing the company based on its $100M revenue run rate.
Key Points:
Thinking Machines is in discussions to raise $1 billion at a $40 billion valuation, down from its previous $50 billion target.
The company's annual revenue run rate stands at over $100 million, reflecting an extraordinarily high revenue multiple.
Investors backing the round include Andreessen Horowitz, Nvidia, GV, Lightspeed, and Conviction Partners.
The company's revenue run rate is over $100 million, indicating strong financial performance despite market volatility.
Thinking Machines introduced Inkling, an open-weight model generating revenue through usage-based compute fees.
Why it matters: If you're an AI startup founder or investor, this valuation drop is a reality check. The $40 billion valuation, while still massive, shows that even top-tier AI companies are adjusting to market realities. For investors, this means a more conservative approach to funding, focusing on revenue multiples and financial performance rather than hype. For startups, it's a reminder that high-profile backing and pedigree are crucial but not enough without solid financials.
Image source: techcrunch.com
Summary: Realta Fusion and MGE are building a 200MW fusion power plant in Wisconsin. This deal could make fusion power a reality for utilities by 2035.
Key Points:
Realta Fusion and MGE will build a 200MW grid-connected fusion power plant in Wisconsin, starting construction in the mid-2030s.
Madison Gas & Electric made an equity investment in Realta Fusion as part of the deal.
Realta Fusion is converting an old Oscar Mayer factory in Madison into its R&D facility.
Commonwealth Fusion Systems (CFS) partners with Dominion Energy to build a 400MW power plant, Arc, near Richmond, Virginia.
Google and Eni have agreed to buy electricity from Arc, expected to come online in the early 2030s.
Why it matters: If you're a utility looking to diversify your power sources, this deal could be a game-changer. Realta Fusion's access to MGE's expertise and infrastructure could pave the way for more fusion power plants. However, the long timelines and significant risks involved mean this is not for the faint of heart.
Image source: techcrunch.com
Summary: Abliteration.ai launches a service to remove guardrails from open-weight AI models, enabling users to perform tasks other models refuse. This democratizes access to uncensored frontier models, but raises concerns about potential harm.
Key Points:
Abliteration.ai hosts modified versions of open-weight models, including Z.ai's GLM-5.3, with guardrails removed.
Users can access the models via web browser or API, with free trials available.
The platform aims to enable offensive cyber, red-teaming, and agent testing work.
Critics warn that making abliterated models available at scale could lead to real harm.
Abliteration.ai is in talks with venture capital firms to raise funds.
Why it matters: If you're involved in red-teaming or offensive cyber work, Abliteration.ai's service democratizes access to uncensored frontier models. This can speed up testing and defense strategies, but it also raises serious ethical and security concerns.
Image source: techcrunch.com
Summary: OpenAI released Astra, the most powerful AI model yet, aimed at cybersecurity and software engineering. It's a game-changer for teams dealing with complex security tasks and software development.
Key Points:
Astra, OpenAI's latest AI model, is available to customers on Thursday, marking a new frontier in AI capabilities.
Astra is being rolled out to OpenAI's Pro, Plus, Enterprise, and Business accounts over the next week, expanding its reach.
Astra's unique opaque recurrence technique obscures the model-monitoring process chain of thought, a critical form of oversight.
Astra's ability to identify and develop zero-day exploits can help defenders find and patch weaknesses, enhancing cybersecurity.
Astra's capabilities in handling tasks with fewer language tokens or no language tokens make it a standout in software engineering.
Why it matters: If you're working on advanced cybersecurity or software engineering tasks, Astra's unique capabilities can significantly enhance your workflow. Its ability to handle tasks with unmatched speed, accuracy, and safety makes it a critical tool for teams dealing with complex security challenges and software development tasks. However, the opaque recurrence technique and reduced ability to monitor certain tasks raise important questions about oversight and transparency.
New Tools & Job
MachineTranslation.com - MachineTranslation.com is an AI-powered translation platform that translates text and documents by comparing outputs from multiple leading AI translation models. Its consensus technology helps users choose the most accurate translation, reducing errors and improving quality across more than 330 languages. What it does Translates text and documents using AI. Compares translations from multiple AI models. Provides a consensus translation to improve accuracy. Supports 330+ languages. Helps users evaluate and refine translations. What makes it good Higher accuracy: Compares multiple AI translations instead of relying on a single model. Better quality: Consensus technology helps identify the strongest translation. Fast and easy: Delivers results in seconds with a simple interface. Broad language support: Covers over 330 languages. Useful for professionals and travelers: Suitable for business, marketing, legal, technical, and everyday communication.
Format Factory - Format Factory is a browser-based workbench for common audio and video conversion jobs. It supports batch processing for video and audio conversion, compression, merging, audio extraction, and audio removal, with transparent credits and temporary results. Try Format Factory for practical media conversion without installer bundles or per-file setup.
HappyHorse.AI - HappyHorse.AI is a multimodal AI video generator for creators, marketers, and social media teams. Users can create videos from text prompts, animate images, or generate videos from reference characters. The platform focuses on browser-based video creation and native audio support.
NeatScribe - NeatScribe is designed for anyone who needs to reuse information from recorded speech. It creates timestamped transcripts, supports translation, and exports results as text files, documents, subtitles, captions, or timed lyrics.
Sakana Marlin - Sakana Marlin is Sakana AI's first commercial product: an autonomous 'Ultra Deep Research' agent positioned as a Virtual Chief Strategy Officer. Given a research topic, Marlin works autonomously for up to roughly eight hours, forming hypotheses, gathering and reconciling information from online sources, and synthesizing a detailed strategy report up to ~100 pages plus executive-summary slides. It builds on Sakana AI research including AB-MCTS (NeurIPS 2025 Spotlight) and The AI Scientist (published in Nature). A closed beta ran from April 2026 with ~300 professionals across finance and consulting.
Senior Machine Learning Engineer, Agentic - Senior Machine Learning Engineer, Agentic at Robinhood — sourced first-party from greenhouse:robinhood on 2026-05-25.
Staff Machine Learning Engineer, Ads Conversion - Staff Machine Learning Engineer, Ads Conversion at Pinterest — sourced first-party from greenhouse:pinterest on 2026-05-25.
Sr. Staff Machine Learning Engineer, Content Ecosystem - Sr. Staff Machine Learning Engineer, Content Ecosystem at Pinterest — sourced first-party from greenhouse:pinterest on 2026-05-25.
Technical Program Manager, Gen AI Operations Planning - Technical Program Manager, Gen AI Operations Planning at Scale AI — sourced first-party from greenhouse:scaleai on 2026-05-25.
Frontier Agent Engineering Manager - Frontier Agent Engineering Manager at Scale AI — sourced first-party from greenhouse:scaleai on 2026-05-25.
QUICK HITS
Ultrahuman Raises $70M for AI Ring Expansion - Ultrahuman raised $70M to expand its AI ring, which can run software locally and reduce cloud reliance. Qualcomm Ventures is backing the round, with Ultrahuman aiming to triple U.S. sales over the next year.
Circular Unveils Ring 3 With FDA-Cleared ECG and 12 Days Battery Life - Circular unveiled the Ring 3 series, featuring FDA-cleared ECG and up to 12 days battery life. Big for health-conscious wearables fans.
Ollie Achieves SOC 2 Compliance, Ensuring Secure Data Handling - Ollie achieved SOC 2 compliance, ensuring secure data handling without sharing user info. Privacy-conscious users, take note.
Google Launches Docs Live, Gmail Live, and Keep Live - Google launches Docs Live, Gmail Live, and Keep Live with AI assistants. Available for Google AI Plus, Pro, and Ultra customers, these features will soon hit Workplace Business too.
Google's WeatherNext 3 Boosts Forecast Accuracy by 60% - Google's WeatherNext 3 AI model outperforms all competitors, offering 60% better rain prediction accuracy. Developers building weather apps should take note.
TikTok Adds Voice, Poll, and Photo Carousel Comments - TikTok is rolling out voice comments, polls, and photo carousels to its comment sections, aiming to boost user engagement. These features are a big deal for creators looking to interact with their audience in new ways.
THAT'S A WRAP
Our newsletter reaches a diverse audience of tech professionals, investors, engineers, managers, and business leaders worldwide. Get in touch today.
Be the first to comment
Get the daily edition delivered free every morning.
Join 3,463 builders reading daily.
Tesla launches the Cybercab, a gold two-seater sedan designed for autonomous driving, in Austin. The vehicle aims to revolutionize the robotaxi market with its advanced AI and camera system.
PsiQuantum and Brookhaven Lab are developing fault-tolerant quantum algorithms, Pasqal and True Nexus are applying quantum computing to protein gelation, and QuSecure demonstrated post-quantum security for the U.S. Army. All projects announced on September 2, 2026.
OpenAI's Astra model employs opaque recurrence, a technique that complicates monitoring AI behavior. This shift has sparked debate among AI safety experts, who fear it may hinder efforts to ensure model alignment.