Chips, Cash, and Construction
· The Fluency Briefing
The Fluency Briefing
Your Guide to What's Happening in AI and Why It Matters to You
Wednesday, June 24, 2026
Breaking this Wednesday: OpenAI just unveiled its own custom inference chip, Qualcomm dropped nearly $4 billion on a startup most people have never heard of, and Forbes confirmed that AI-linked companies accounted for more than half of all new market value in the Global 2000. The AI boom isn't slowing down - but it's running into some very human speed limits.
Today in AI:
- OpenAI Built Its Own Chip (and Named It After a Pepper) - OpenAI and Broadcom unveiled Jalapeno, a custom inference chip designed from scratch for large language models and delivered from concept to production in just nine months. It's OpenAI's first step toward owning the full hardware stack beneath its models. OpenAI
- Qualcomm Bets $4 Billion on a 150-Person Startup - Qualcomm announced it's acquiring Modular, the chip software company founded by ex-Google TPU engineers, for nearly $4 billion - roughly 2.5x its valuation from just nine months ago. The deal signals Qualcomm's serious push beyond smartphones and into AI data centers. Wired
- AI Companies Now Drive Over Half of Global Market Value Growth - Forbes' 2026 Global 2000 shows tech and semiconductor firms nearly doubled their combined market value to $41.4 trillion, accounting for 57% of the list's net increase. The AI boom is reshaping corporate rankings in real time. Forbes
- The AI Boom's Newest Bottleneck: Actual Humans - Saint-Gobain's CEO told Bloomberg that skilled labor shortages are now slowing data center construction across North America and Europe. Turns out you can't build a server farm with a chatbot - you still need electricians, welders, and engineers. Tom's Hardware
- Microsoft Says Its AI Data Centers Are Actually Using Less Water - Microsoft pointed to lower water consumption numbers despite its massive AI infrastructure expansion, as hyperscalers race to address growing environmental concerns about the buildout's footprint. The details are thin, but the narrative shift is notable. Axios
- The US Government Wants a Look at Meta's AI Models - Meta is reportedly the only major AI developer that hasn't voluntarily submitted its models for government safety review. OpenAI, Anthropic, Google, xAI, and Microsoft have all cooperated already. Meta says it hopes to sign on soon. Engadget
- Apple Patches the Siri AI Waitlist Hack - Developers who used a Terminal command to skip the Siri AI waitlist in macOS 27 beta 1 are finding the trick no longer works after updating to beta 2. Apple appears to have moved the access check from a local flag to server-side validation. MacRumors
- 17 Tech Workers Share How to Survive the AI Office - Sifted polled industry insiders on what skills actually matter in an AI-augmented workplace. The consensus: critical thinking, adaptability, and knowing when to trust the machine - and when not to. The era of pure technical skill as a career moat is fading fast. Sifted
Today's Takeaway:
On the surface, OpenAI designing its own chip, Qualcomm acquiring Modular for $4 billion, and NVIDIA deepening its AWS partnership look like three separate business stories. Together, they reveal something more interesting: the AI industry is frantically rebuilding its own supply chain from the ground up. OpenAI doesn't want to depend entirely on NVIDIA for inference hardware. Qualcomm doesn't want to be stuck selling phone chips while data center money flows elsewhere.
And NVIDIA is locking in cloud partnerships to make sure it stays indispensable. Everyone is vertically integrating because nobody trusts the current setup to hold. As Forbes reported Wednesday, AI-linked firms added over $17 trillion in market value in a single year - but that wealth is concentrated among companies that control their own infrastructure.
Meanwhile, as Tom's Hardware notes, the physical world isn't cooperating. You can design a chip in nine months, but you can't train an electrician in nine months. The AI boom has billions in capital and no shortage of ambition - what it's running short on is the physical capacity to turn plans into running data centers.
Think of it like ordering a race car but discovering there's a six-month wait for the road it needs to drive on.
🔧 Tool Spotlight
- Suno - AI music generation that gives anyone a producer's chair.
What it does: Type a prompt like "upbeat indie folk about Monday mornings" and Suno generates a full song with vocals, instruments, and lyrics in about 30 seconds. You can keep what works, regenerate sections, or extend the track. Who it's for: Anyone who's ever hummed a tune and wished they could hear it back - no music theory needed. Especially fun for content creators making background tracks, podcasters who want a custom intro, or parents making birthday songs. Try this first: Go to suno.com, click "Create," type a one-line description of a song about your day, and let it run. You'll have a real song you can play, share, or download in under a minute. Free or paid: Free tier gives you 10 songs/day (watermarked). Paid plans start at $10/mo for commercial rights and faster generation.
The Bottom Line
The Pattern: Every major AI player is racing to own more of the stack - chips, software, infrastructure, even the construction pipeline - because dependence on someone else's supply chain is now a strategic risk. The money is enormous, but the physical and human bottlenecks are real.
Why It Matters: If you're running a business that touches AI (and increasingly, who isn't?), the cost and availability of AI compute will shape what's possible for you. These infrastructure battles between giants will determine how fast, cheap, and accessible AI tools become over the next two years.
Your Move: Stop watching AI model releases as the main show. The real action is in who controls the pipes. This week, ask yourself: which parts of your AI workflow depend on a single provider, and what's your backup plan if they raise prices or hit capacity limits?
What We're Working On
✨ Founding Cohort Special - 60% Off! - Use code MAF20 to join for just $20/month (regularly $50). Get weekly group sessions & workshops, self-paced courses for all levels, access to tools & templates, challenges with peer feedback, and 24/7 support community. → Join Now
✨ Free 30-Minute AI Consultation - Discover how My AI Fluency can help your business unlock the potential of AI. We'll discuss your goals, explore practical AI opportunities for your industry, and outline clear next steps. → Schedule Free Call
✨ How AI-Fluent Are You? - Test your AI fluency with our interactive quiz. See how you stack up and discover what to learn next. → Take the Quiz
💬 Community | 📞 Book a Consultation | 🌐 Website
Fluently yours, The My AI Fluency Team