The Briefing: Chips Print Cash, Mall Talk Moves Into ChatGPT, And AI Nations Go Shopping For Power
Issue #2025W47.4 / November 20 2025
Hi There 👋
Global markets spent the last 48 hours trying to decide if AI is still a bubble or just the new electricity bill. Nvidia is printing money, Adobe is buying its way into the post-SEO world, Target is turning ChatGPT into an aisle, and Saudi Arabia plus its neighbours are quietly positioning themselves as AI engine rooms for half the planet.
Behind the noise, a handful of life science breakthroughs hinted at something more important than stock charts: AI-era biology that can switch drug resistance off like a feature flag.
The AI Value Chain Is Locking In
Nvidia’s latest quarter sounded less like a tech earnings call and more like a macro story. The company reported about 57 billion dollars in quarterly revenue, up 62 percent year on year and nearly ten times what it made in the same quarter three years ago, even as sales of advanced chips to China dropped sharply under US export controls.
If you strip China out, Nvidia’s revenue in the rest of the world roughly doubled, driven by hyperscalers like Meta, Microsoft, Google, and Amazon ramping AI capex that still mostly flows into Nvidia GPUs. Nvidia’s CFO has floated a target of more than 500 billion dollars in sales over the two years ending December 2026, compared with about 130 billion dollars in revenue in the year to January 2025.
At the same time, Nvidia keeps striking “circular” deals: investing cash into AI companies that, in return, commit to training on Nvidia hardware. Anthropic, long associated with Amazon and Google, has signed on to use Nvidia’s chips under a set of multi-billion dollar arrangements that include investments from both Nvidia and Microsoft. It is effectively vendor financing for the AI era.
Zoom out and you can see the rest of the stack starting to crystallize:
Brookfield launched a 100 billion dollar AI infrastructure program to build “AI factories” based on Nvidia’s DSX reference design, plus power and data centers across multiple countries. Nvidia and the Kuwait Investment Authority are founding partners.
AMD, Cisco, and Saudi startup Humain announced a joint venture for data centers in Saudi Arabia, starting with a 100 megawatt site fully contracted by generative video startup Luma AI, with ambitions to hit 1 gigawatt of capacity by 2030.
The US Commerce Department authorized the export of advanced Nvidia Blackwell chips to Abu Dhabi’s G42 and Saudi-backed Humain, up to the equivalent of 35,000 GPUs each, explicitly tying chip exports to “rigorous security and reporting requirements”.
For now, Nvidia’s CEO calls this a transformation, not a bubble. Skeptics point out that many of these deals involve capital recycling inside a small circle of hyperscalers, chipmakers, and well-funded AI labs. Both can be true at the same time.
Search Is Turning Into “Generative Visibility”
Adobe decided not to wait for search to reinvent itself. It is paying 1.9 billion dollars in cash to acquire Semrush, a long-time SEO and digital marketing platform, at 12 dollars per share, roughly a 77 percent premium to its last close.
On paper, Semrush is a mid-sized software firm with slowing growth. In practice, Adobe is buying the telemetry that shows how brands are discovered across classic search results and generative bots like ChatGPT and Gemini, then plugging that data into its own AI marketing tools.
The signal: visibility is shifting from “rank in search” to “be the default suggestion in conversation.”
That point was underlined by Target, which announced a “first-of-its-kind conversational shopping experience” inside ChatGPT. People will be able to tag Target in ChatGPT, get curated product suggestions, build a multi-item basket, and check out with drive-up, pickup or shipping, all without leaving the chat.
Target already uses ChatGPT Enterprise internally, and this public experience is effectively a live test of what “AI-native retail” might look like when discovery, recommendation, and checkout merge into one interface.
Quiet Frontier Breakthroughs in life science
While AI infrastructure gets the headlines, the last 48 hours also delivered several quietly important advances in health and science.
Antibody that infiltrates kidney cysts
Researchers at the University of California, Santa Barbara unveiled a specially engineered dimeric IgA antibody that can slip inside kidney cysts and block c-MET signaling, a key growth pathway in polycystic kidney disease (PKD). In mouse models it slowed or even reversed cyst expansion without damaging healthy tissue.
If this translates into humans, PKD patients could eventually get a targeted therapy that prevents progression rather than waiting for dialysis or transplant. It is a textbook example of how precision biologics plus clever targeting can transform chronic disease management.
CRISPR reverses chemotherapy resistance in lung cancer models
At ChristianaCare’s Gene Editing Institute, scientists demonstrated that knocking out the NRF2 gene with CRISPR restored sensitivity to chemotherapy in lung squamous cell carcinoma models, both in vitro and in animals.
NRF2 is implicated in drug resistance across multiple solid tumors. The work suggests a future pattern where doctors use gene editing to re-enable existing drugs, instead of continuously hunting for new ones. If clinical trials confirm the effect, it could turn CRISPR into a lever for getting more life out of current oncology pipelines, not only for new bespoke treatments.
In Other News
Larry Summers leaves OpenAI’s board and Harvard roles after the release of congressional documents detailing years of personal correspondence with Jeffrey Epstein, triggering a broader review at Harvard and reputational questions for OpenAI’s governance.
Nokia (yes, that Nokia) laid out a renewed strategy focused on AI-ready networks, cloud-friendly architectures and industrial digitalisation, trying to reposition itself as infrastructure for AI and edge workloads instead of a legacy telecoms vendor.
Zap Energy said it achieved a key high-temperature fusion milestone at Oak Ridge National Laboratory, claiming progress toward cheaper, more compact fusion devices that could complement the huge tokamak projects.
Quantum computing debates flared again, with new commentary arguing that error-corrected quantum machines could become a “general purpose industry” in their own right, while others warned of overpromising timelines.
Music labels and AI streaming edged toward détente as major labels signed licensing deals with AI music startup Klay, signalling a slow move from “shut it down” to “own a piece of it” in generative audio.
Thank you for reading The Briefing — your nightly rundown of the key stories shaping business, technology, and innovation. If you like what you see, subscribe for free to receive the next editions in your inbox daily (Mon-Fri).