WHAT’S INSIDE

👇️ Todays’ Feature: Thinking Machines Lab Rethinks Fine Tuning

• Tech News: Shutdown, Fed, TrumpRx, UK vs Apple, Apple Smart Glasses

• Company Watch: Thinking Machines, Cerebras, Periodic, Intel, Axiom Math

• Buzzy Tools: Latest Buzzy tech, AI and financial tools

• Deep Tech: The latest in deep tech, biotech, futurism and more

• Space Tech: Latest news in the space race and aerospace tech

• Crypto: Blockchain and crypto policy and startups or protocols to watch

TECH IN THE NEWS

Shutdown Showdown — The Trump administration controversially used federal websites for partisan messaging during a government shutdown, blaming "Radical Left Democrats" and praising President Trump.

Fed Firefight — The Supreme Court has ruled that Pres. Trump can’t immediately dismiss Lisa Cook from the Federal Reserve. This decision maintains her position until a January hearing, emphasizing protections for independent agency leaders.

TrumpRx Discount Portal — Launching in 2026, TrumpRx will guide Americans to discounted medications through a Pfizer deal. It includes tariff relief for Pfizer and price cuts on Medicaid drugs aimed at out-of-pocket payers.

UK Revives Encryption Backdoor Demand — The UK government renews calls for Apple to create a backdoor for iCloud data despite previous retraction. Privacy advocates warn this could weaken global encryption, leaving Apple at a crossroads.

Apple's Vision Shift — Apple has scrapped its lighter Vision Pro plans to develop two new smart glasses models. The display-free model targets a 2027 launch, while a display version is accelerated to meet Meta's market lead.

Digital Euro's AI Shield — The European Central Bank has chosen Portuguese AI startup Feedzai to combat fraud in the digital euro project with a contract worth up to 237.3M euros. Launch is anticipated in 2029, pending legislative approval.

IN PARTNERSHIP WITH ACORNS

Build wealth while you sleep

Acorns started a decade ago with a simple yet novel idea: automatically round-up every purchase you make, and invest the difference. Their philosophy has always been predicated on compounding as the key to growing wealth—true, and tested.

Now, they’re steadily compounding over $6 billion for everyday Americans. Acorns’ pre-built portfolios are designed to match your risk tolerance. All you do is deposit money, and forget about it. Or, let it deposit for you by rounding up your everday purchases.

This is a real screenshot of my Acorns Invest account, compounding over the last 4 years. I trust it to work for me, and don’t think twice. Acorns is my savings. They’re truly a phenomenal company, app, philosophy, product.

Download Acorns and try it for yourself, with $5 gifted in your account to get up and running.

How Thinking Machines Lab is Rethinking AI Model Customization

Vivien Wang

On Wednesday, AI startup Thinking Machines Lab, led by ex-OpenAI CTO Mira Murati, unveiled its first product: Tinker. Tinker enables people and organizations from researchers to individuals to businesses to experiment with advanced LLMs, removing barriers of infrastructure and select knowledge that had previously limited access to large scale distributed AI computing. 

What is Thinking Machines Lab?

Founded in early 2025 by Murati and several other former OpenAI employees including John Schulman and Barret Zoph, Thinking Machines has already raised $2B in seed funding from the likes of Nvidia and Andreessen Horowitz, bringing its valuation to an impressive $12B. Access to the first product, Tinker, is available via application, and the startup is currently not charging for its API despite future plans to do so. Thinking Machines’ founding team possesses years of expertise in reinforcement learning, robotics, and more. 

Tinker itself is a Python SDK and API that currently allows users to fine-tune open source models including Meta Llama and Alibaba Qwen with just a few lines of code to start distributed supervised or reinforcement learning runs. With Tinker, the complexity of distributed GPU cluster management is turned into an automated process that grants users detailed control over data, objectives, and algorithms while eliminating pains like resource scaling and fault tolerance. 

Tinker also utilizes Low-Rank Adaptation (LoRA), enabling efficient tweaks without retraining huge models and thus allowing cost-effective multi-user and multi-experiment backend sharing for experimentation. Tinker also comes with an open-source companion, the Tinker Cookbook. Cookbook offers ready to use techniques, letting researchers launch new projects quickly and easily. 

Early beta access groups for Tinker include those at Stanford, UC Berkeley, Princeton, Redwood Research, and more. So far, Tinker has received praise for its control, and ease of use compared to other existing solutions. 

The Market for Tinker

As individuals, businesses, researchers, and governments are scrambling to optimize LLMs for industry specific tasks, Tinker is able to address the emerging need for accessible model fine tuning. This sector has seen a 70% funding surge since the last year, with total investment in distributed computing startups adding up to around $9.8B. 

Tools like Tinker fill the gaps that are left by mainstream APIs and alternatives like VERL and SkyRL that are less developer friendly and more focused on raw infrastructure. Tinker is focused on customization and portability of AI models across platforms and clouds, whereas rivals tend to emphasize large GPU pools for high bandwidth workloads and place a lesser emphasis on customization. 

Competitors and Key Players

The infrastructure market for distributed AI compute has largely been dominated by CoreWeave, now valued at $19B thanks to recent $14B deals with Meta and OpenAI. It focuses on rapid scaling and deep GPU inventory, serving massive AI workloads for other enterprises. CoreWeave possesses a notable strategic partnership with Nvidia and differentiates itself by providing custom engineering for use cases like VFX rendering and large scale LLM training along with rapid provisioning and high availability that is unmatched by most generalist clouds. 

Another key player is Modal, recently acquired by Oracle for $850M. It is built for serverless deployment of AI and ML workloads, with a pay-as-you-go model designed for fast prototyping and unpredictable AI tasks. Modal is tightly integrated with Oracle’s cloud stack and GPUs and allows for quick deployment, but is less suitable for persistent, large-scale training. 

Thinking Machine Labs in contrast is targeting the middleware layer, providing standardized APIs and workflow orchestration that connects diverse AI models, enterprise data, and application logic. Its focus is on interoperability and composability between model endpoints, making it an attractive solution for firms that need robust AI/MLOps without becoming overly dependent on a specific vendor or platform. TML is currently using two open-source models, one in the US and one in China, broadening access globally and presumably enabling them to keep costs relatively low as these open models are well-known for their comparative ease of fine tuning. 

Conclusion

The launch of Tinker marks a readily expanding opportunity in distributed and customizable AI, as barriers to entry fall and frontier models are democratized. Challenges could be present in areas such as data governance, model misuse, and cloud cost, but there seems to be a shift occurring from monolithic, opaque AI models to more open, fine-tuned, and widely accessible AI systems. The next major step in AI could be in fine tuning, and if Thinking Machines Lab is heading in the right direction through this pursuit, it could unlock endless potential in research, commercial applications, and other undiscovered tools. 

COMPANIES TO WATCH

OpenAI — Closed a $6.6B secondary share sale at a $500B valuation, becoming the world's most valuable private company. Allows employees to cash out while staying private, highlighting strong investor confidence in the strategic direction.

Nvidia — Surpassed a $4.5 trillion market cap on growth in AI infrastructure deals, including a $100B equity stake with OpenAI and a $500B data center project. With stock prices rising and AI chip demand soaring, Nvidia cements its tech dominance.

Thinking Machines Lab — Co-founded by Mira Murati, John Schulman, Barret Zoph and other ex-OpenAI team, TML launched Tinker, a tool for automating custom frontier AI model creation via fine tuning open source models from Meta and Ali Baba.

Cerebras Systems — The AI inference services startup raised a $1.1B Series G round valued at $8.1B. This funding will enhance data center and U.S. manufacturing expansion, with an IPO planned despite regulatory delays. Total raised is ~$2B.

Periodic Labs — The AI-driven scientific discovery startup, founded by ex-OpenAI and Google DeepMind researchers, secured a record $300M seed round from Andreessen Horowitz, Nvidia, Jeff Bezos, and Eric Schmidt.

CoreWeave — The AI infrastructure company sealed a $14.2B deal with Meta, increasing its stock by 10%. Alongside a $6.5B expansion with OpenAI, Its GPU-as-a-service model is reshaping AI computing power sourcing.

Intel — Saw a 50% stock surge, increasing the U.S. government's stake to $16B. Renewed confidence in Intel's foundry strategy and potential AMD partnerships position it as a domestic alternative to TSMC amid rising chip demand.

Google — Acquired Atlantic Quantum, a quantum computing hardware startup, to enhance its superconducting qubit systems. This acquisition aims to integrate advanced modular chip technology, improving error correction and coherence times.

Tesla — Delivered 497,099 vehicles in Q3 2025 due to expiring EV tax credits. Despite a 5% production drop, challenges loom in Europe and China. Elon Musk foresees future rebounds with AI innovations like robotaxis.

Perplexity — Acquired the team from Visual Electric to create a new division focused on AI agents, transitioning from search to creative AI. This strategic move aims to integrate AI into design workflows to rival OpenAI, Google.

Axiom Math — A San Fran startup focusing on AI for solving and generating complex math problems. Raised a $64M seed from investors like B Capital, Greycroft, Madrona, and Menlo Ventures, positioning it as a leader in AI-math.

SignalPulse — Read the Top signals, Startups + Deals From Last Week

TECH BUZZ DIRECTORY
Featured Products*

SEO Bot — A fully autonomous "SEO Robot" with AI agents for busy founders.

Memelord — Meme Software for marketing memes, tech memes, & sales memes.

Otonomos — Incorporate your Delaware C-corp or tax advantaged foundation and get 5% OFF.

BUZZY TOOLS

Buzzy Tech Tools To Watch & Use

OpenAI Sora — AI video app rivals NanoBanana, TikTok - #3 on App Store.

Google Jules Tools — CLI with API for dev AI coding workflow integration.

Google Home — Gemini AI boosts speed 70% + AI-powered Home speaker.

Apple Smart Glasses — Voice-AI glasses to challenge Meta with speakers, cameras.

DoorDash — Dot robot delivers food autonomously, reducing traffic, tested in AZ.

Salesforce — Agentforce Vibes integrates with accounts for AI-driven code via NL.

BUZZY TECH

The Latest Deep Technology & Trends To Watch

Nvidia CUDA-Q — Achieves 4,000x speedup in quantum simulations.

3D Printed Glass — Binder-free method enables precise silica structures w/ light.

TSMC EUV Energy Program — Cuts power by 44%, saving 190M kWh by 2030.

Dreamer AI — Self-improving AI mines diamonds in Minecraft without training.

Amazon's Blink Arc — AI-driven accessory creates 180-degree security view.

CRYPTO WATCH

Crypto Surge Amid Turmoil — Crypto prices are climbing, led by ETH and SOL, despite a U.S. government shutdown and 32,000 job losses. The market expects a 99% chance of a 0.25% rate cut, boosting Bitcoin to $116,910.

World Liberty Financial — Plans to tokenize real-world assets like oil and gas, paired with its USD1 stablecoin, now the fifth largest globally. A debit card and mobile app are set to launch in late 2025 or early 2026.

Spacecoins — Spacecoin has successfully sent a blockchain transaction via its CTC-0 satellite, challenging centralized systems like SpaceX's Starlink. The company aims for a decentralized, censorship-resistant network with plans for more satellites.

Zcash Rally — Zcash has seen a significant rally fueled by growing interest in privacy-focused cryptocurrencies. The content stresses the need for personal research and professional consultation for investment decisions.

Stablecoins Gain Traction — Small businesses are adopting stablecoin payments to bypass high credit card fees. While they offer faster, cheaper transactions, consumer adoption challenges remain.

SPACE_RACE

Katalyst Space_Space-docking tech to boost SWIFT observatory & US space race

Firefly Alpha Rocket_Explosion in Texas delays Lockheed Martin flight, hits shares

Aetherflux_Infrared laser satellites beam solar energy to Earth for clean energy

Spacecoin_Decentralized Starlink rival sent first blockchain transaction via satellite

Sunless World Auroras_Rogue brown dwarf shows auroras without a sun

Enceladus — Cassini finds fresh organics, subsurface life on Saturn's moon?

FOR INVESTORS
Open Deals
Investors– request an intro to startups - [email protected].

Before we restart our placements of raising startups, we’re sharing an open call to send in your own company. Just shoot us an email with a deck and we may feature you in this bulletin.

For detailed pitch materials, please email [email protected] with your deck.
🔥 — hot deal!
⏰ — leaving soon.

FOR EMPLOYERS
Job Candidates

Employers– request an intro to candidates - jobs@techbuzz.ai