Sponsored by
The AI industry has a money problem hiding under all the model launches: infrastructure spending is now outrunning what regular debt markets can fund. At the same time, the tools themselves keep getting faster, cheaper in some places, and pricier in others, sometimes in the same week. Here's what actually happened, and what it means if you're running a business instead of a hyperscaler.
One Account. Every Market. No Closing Bell.
Markets don't wait for Monday. News breaks on a Saturday morning, and most traders can do nothing but watch.
Not on Liquid. Trade domestic and international equities, commodities, forex, crypto, and prediction markets — all from one account, 24 hours a day, 365 days a year. Liquid gives you access to any market, from anywhere, anytime. To us, access is arbitrage.
Getting started takes under 10 minutes: log in with Google, deposit with Apple Pay or a bank transfer, and trade from your phone or desktop — wherever you are in the world.
While everyone else is refreshing headlines and waiting for the open, you're already positioned. That's the difference between reacting to markets and actually trading them.
AI's Buildout Has a $1 Trillion Money Problem
What happened: Apollo's chief economist Torsten Slok says the AI industry could need more than $2 trillion in debt through 2030, but traditional investment-grade bond markets may only absorb under $1 trillion of that, according to Forbes. AI-related borrowing already makes up more than 40% of new long-term investment-grade corporate debt. The "Magnificent Seven" tech giants are projecting $738 billion in combined capital spending this year, and the growing shortfall is increasingly being filled by private lenders secured against data centers and chips instead of ordinary corporate bonds.
Why it matters to you: This is the clearest sign yet that the AI boom is being built on borrowed money, not just cash from profitable companies. That doesn't mean the bubble pops tomorrow, but it does mean the vendors selling you "unlimited AI for $20 a month" are themselves leaning harder on debt to keep building the infrastructure behind it.
What to do about it: Don't lock your business into a single AI vendor's roadmap promises. Keep your workflows portable enough to switch tools if pricing or reliability changes fast.
Google's Co-Founder Came Out of Retirement to Save Gemini
What happened: Sergey Brin personally urged Google DeepMind staff to "move faster" on Gemini as Anthropic and OpenAI pulled ahead, addressing hundreds of employees directly, according to Reuters. Days later, Google announced a leadership shakeup: DeepMind chief Demis Hassabis moved into a reduced chair role, his deputy Koray Kavukcuoglu took over running DeepMind, and Gemini's original technical co-leads, Jeff Dean and Oriol Vinyals, left to start their own company. The moves came after Gemini briefly led the market before falling behind on coding performance.
Why it matters to you: When the company behind Gmail, Docs, and Workspace shakes up its AI leadership because it's losing ground, expect the AI features already baked into the tools you use every day to change faster and more aggressively over the next few months, not slower.
DeepSeek Just Made Its AI Up to 14 Times More Expensive
What happened: DeepSeek, the Chinese AI lab known for undercutting Western competitors on price, raised API pricing for its V4 models effective August 17, with increases ranging from 50% up to roughly 14 times the previous rate depending on the model, token type, and time of day, according to Reuters. The hikes apply differently across its cheaper V4 Flash model and the newer, more capable V4 Pro.
Why it matters to you: "Cheap AI" pricing is a promotional phase, not a permanent feature. If any part of your business runs on a bargain AI API, budget for the day the provider decides it needs to actually make money.
OpenAI's Newest Model Just Got 14 Times Faster
What happened: OpenAI is previewing "Ultrafast," a new API tier that runs its GPT-5.6 Sol model at up to 750 output tokens per second, up to 14 times faster than standard processing, powered by chipmaker Cerebras, according to OpenAI. The company says the speed is built for time-sensitive work like real-time customer support, fraud detection, and incident response, not just faster chatbots.
Why it matters to you: If a slow, laggy AI chatbot has ever cost you a sale or frustrated a customer on hold, this is the kind of upgrade to watch for. Speed tiers like this are what make AI usable for real-time, customer-facing work instead of just backend drafting.
Apple Built Its Own AI Model, Just for China
What happened: Apple has trained its own large language model specifically for the Chinese market with support from Alibaba, according to Reuters, marking a shift away from relying entirely on outside AI providers there. Alibaba's Qwen model will power some Apple Intelligence features for Chinese users, while Apple's own proprietary model, reportedly the first foreign-built AI model approved for deployment in China, will run others.
Why it matters to you: If you sell into international markets, this is a preview of what's coming: AI features are increasingly built and regulated market by market, not shipped as one global product. Don't assume the AI tools you rely on here will work the same way, or be allowed to work at all, everywhere your customers are.
Meta Released a Free AI Agent That Works With No Internet
What happened: Meta open-sourced Muse Glimmer, a 30-billion-parameter AI agent model that runs entirely on a laptop or mid-range gaming PC, no cloud connection or per-prompt API bill required, according to Meta AI. It can read screenshots, call tools, and recover from its own errors across multi-step tasks. Meta released it under an open Apache 2.0 license with the model weights available on Hugging Face.
Why it matters to you: This is a real alternative to paying monthly for cloud AI tools for certain tasks. If your business has predictable, repeatable AI workflows like drafting, data entry, or simple research, a free model that runs locally is worth a look before you commit to another subscription.
AI Is Splitting Winners From Losers in Hospitality, and the Middle Is Getting Squeezed
What happened: Kasa CEO Roman Pedan describes a "barbell effect" playing out in hospitality, according to Fortune: giant hotel chains use AI to spread technology costs across huge volume, while tiny independent operators now rent enterprise-level AI capabilities they could never afford to build. The businesses getting squeezed are the ones in the middle, regional operators with 15 to 50 properties who carry corporate-style overhead without the scale to justify building their own AI systems. In Kasa's own case, after putting an AI-native operating system across its properties, direct bookings rose sixfold and revenue per available room grew 19.2% year over year, without changing a single building.
Why it matters to you: This barbell pattern isn't unique to hotels. If your business sits in the "too big to be scrappy, too small to build custom tech" zone in any industry, take this as a warning worth acting on, not just a hotel story.
Uber and a Chinese Robotaxi Company Are Bringing 2,000 Driverless Cars to Europe
What happened: Uber and Pony.ai are expanding their partnership to deploy more than 2,000 robotaxis across Europe, starting with a commercial launch in Zagreb, Croatia, and expanding to four more cities, according to Euronews. Pony.ai supplies the self-driving technology it already runs in Beijing and Shanghai, while Uber handles booking and payment, with local operators managing the fleets day to day.
Why it matters to you: This is a small preview of AI-driven automation moving from software into physical services most business owners assume are years away. If your business touches transportation, delivery, or logistics in any way, autonomous vehicles are no longer a "someday" story, at least in some markets.
The Bottom Line
Every story today is a version of the same tension: the money behind AI is getting shakier while the technology itself keeps getting faster, cheaper in some places, and pricier in others, often in the same week. Google panicking enough to pull its own co-founder back in, DeepSeek quietly hiking prices, an entire industry leaning on debt it hasn't figured out how to pay back yet, none of that is a reason to swear off AI tools. It's a reason to stop assuming the ground under any single vendor is stable. Build like the tools you use today might look different, or cost different, in six months. Because on this rollercoaster, they usually do.
Enjoying the Ride?
If this issue was useful, forward it to one person who needs to hear it straight, with no hype. New here? Hit subscribe so tomorrow's issue lands in your inbox too.
Talk tomorrow,
Mark Shilensky
Follow me personally on Social Media:
Facebook: https://www.facebook.com/MarkShilenskyPage
Instagram: https://www.instagram.com/markashilensky/
Threads: https://www.threads.com/@markashilensky
LinkedIn: https://www.linkedin.com/in/shilensky/
X: https://x.com/markshilensky
TikTok: https://www.tiktok.com/@mark.shilensky
BlueSky: https://bsky.app/profile/markshilensky.bsky.social