Good morning,
Visit our vibe coding or guides section for more applied Gen AI posts.
Today’s article is going to be AI news focused, like a brief - because frankly this is one of the biggest AI news weeks of the 2025, with some Earnings surprises, layoffs and insights on the future of BigTech.
Technology, business and stock market news are an important context for understanding the bigger picture around AI in 2025. I want to take this moment to recognize Nvidia’s key role in AI. Let’s try to process this. 🧠
Nvidia Makes History 🌊
"Nvidia CEO Jensen Huang said “the company projects global spending on AI infrastructure will be $600 billion this year, and could reach $3 trillion to $4 trillion by 2030. That would be a growth rate of 400% to 566% at the top range.”
While Nvidia’s own earnings are on November 19th, this week we learned Nvidia have new deals with Nokia, Samsung, Hyundai, Palantir, Eli Lilly, Uber among many others. Watch the GTC Washington Keynote here. It’s now a $5 Trillion dollar company, just four month since surpassing $4 Trillion. This is fairly historic and outside the norm.
Nvidia’s time to market caps of $3 Trillion and $5 Trillion look highly unnatural compared to historical norms, 66 and 78 days respectively.
This is what would occur in an AI boom with semiconductor and AI Infrastructure characteristics, the start of an AI bubble - though we’re “not in an AI bubble” according to Jensen Huang.
Jump to Sections you might like: GTC Keynote:
I’m bolding the sections I think are the most interesting:
00:00:00 America, the Land of Innovation
00:04:29 Intro
00:06:03 GPU to CPU Accelerated Computing Shift
00:08:09 CUDA-X Libraries
00:11:05 CUDA Opens New Markets
00:15:34 Telecommunications
00:17:34 Nokia to Build AI-Native 6G on New NVIDIA Arc
00:21:37 Quantum Computing
00:24:00 The Future of Quantum Computing
00:26:17 Announcing NVIDIA QLink Quantum-GPU Interconnect
00:28:40 Department of Energy Partnering with NVIDIA to Build 7 New AI Supercomputers
00:30:32 AI - The New Industrial Revolution
00:42:13 Three Scaling Laws
00:50:03 Extreme Co-Design
00:51:57 Grace Blackwell NVL72 - “A Thinking Machine”
00:53:30 Extreme Co-Designed Blackwell NVL72
00:58:03 CSP Capex Spend
01:01:10 Exceptionally Strong Demand for Grace Blackwell NVL72
01:03:00 Blackwell: Made in America, Made for the World
01:06:48 NVIDIA Extreme Co-Design Delivers X-Factors on One-Year Rhythm
01:08:20 Vera Rubin Superchip, Vera Rubin Compute Tray, Vera Rubin CPX Compute Tray, BlueField-4, NVLink Switch Tray
01:14:42 AI Factories
01:15:22 Omniverse DSX
01:19:08 Open Models
01:21:20 NVIDIA is Everywhere
01:26:30 Physical AI
01:28:51 America is Reindustrializing
01:34:00 Newton Simulation Platform
01:34:37 Humanoid Robotics
01:35:55 NVIDIA Drive Hyperion and Uber Partnership
01:37:42 Global Robotaxi Ecosystem
01:39:23 Close
Quote: 💬
Jensen spent 2 min on InferenceMAX at the GTC keynote
“InferenceMAX is completely open source and completely impartial, with the support of AMD, OpenAI, Nvidia, Microsoft, Oracle, Pytorch, SGLang, vLLM, Cursoe, Nebius, Tensorwave, etc” - , Semianalysis on X.
Nvidia is also important given the rumor that Blackwell chips might be on the table to sell to China (though this was not on the table when the two leaders met a few hours ago). Hard to know the accuracy of this but Trump said he may discuss Nvidia’s Blackwell AI chips with Chinese President Xi Jinping. It turns out this did not actually materialize.
Nvidia has invested in a host of Open Source AI startups among others, while having most of its stock market investments in CoreWeave and ARM.
With Nvidia’s 70% gross profit margins, Bernstein calculates that the company captures nearly 30% of total AI data center spending as profit.
There are dozens to hundreds of new datacenters being built all over the world and they are getting a lot bigger in the 2020s. GPUs make up 39% of the cost of the biggest datacenters and will need to be continually upgraded. AI datacenters will not only get more numerous but far larger in the years to come. Nvidia has a 94% market share in the best AI chips.
Age of GPU Accelerators? 📊
CEO Jensen Huang said Nvidia projects global spending on AI infrastructure will be $600 billion this year, and could reach $3 trillion to $4 trillion by 2030. That would be a growth rate of 400% to 566% at the top range, and a compound annual growth rate between 38% and 46%.
China U.S. Trade Deal Update 🌏
U.S. President Donald Trump said he has reached a 1-year agreement with China on rare earth (metals) supplies.
Trump said the two sides had discussed “a lot of chips” but not the most advanced Blackwell chips. Taiwan was not part of the discussion. The meeting was shorter than expected and inspite of the upbeat tone, little seems to have been accomplished outside of Rare Earth Metals for lower tariffs. China seems to have won this round.
October 30th, 2025 in South Korea.
AI Product updates ✨
Two product AI updates that really caught my attention:
Grammarly has changed its name to Superhuman in a huge AI upgrade. Grammarly is adopting its new name from AI email app Superhuman Mail, which Grammarly acquired in June.
Cursor announces Cursor 2.0 and Composer. The vibe coding tool Cursor, from startup Anysphere, has introduced Composer, its first in-house, proprietary coding large language model (LLM). Anysphere in October said it was considering funding at a $30 Billion valuation. Watch the promo video for Cursor 2.0 here.
By TSMC, though most are still made in Taiwan.
Short video (source): 3 min, 45 seconds:
Nvidia has become an Emerging Tech Nexus in 2024-2025
The GTC conference once more about GPUs is now also about robotics, physical AI and quantum computing partnerships.
Nvidia is driving the AI bubble and AI Boom 🔔
NVIDIA hosted its GTC conference in Washington, D.C. from October 27-29, 2025, focusing on AI advancements, accelerated computing, and national infrastructure. Coinciding with Trump’s visit to Asia there was always going to be a lot of partnerships announced, but this is even better than we expected.
The Blackwell processor, Nvidia’s flagship artificial intelligence accelerator, and the newer Rubin model are fueling an unprecedented surge of sales growth through 2026, Huang said Tuesday at the ‘GTC’ company presentation in Washington.
Nvidia’s stock is up nearly 50% YTD (49.69%), which is unbelievable given the $5 Trillion market cap. Partnering with the Trump Administration has done it well. NVDA 0.00%↑
AI data centers are measured in gigawatts of power these days.
Mega AI Datacenters are coming Online 🗺️
AI datacenters are getting noticeably bigger in 2025. GPUs are the main cost, which explains Nvidia’s incredible rise.
TD Cowen analysts put this in context, writing in a research note this week that 1 gigawatt is roughly the output of a nuclear reactor. That’s the new baseline for next-generation AI data centers, such as xAI’s Colossus 2 in Memphis, Meta’s Prometheus in Ohio and Hyperion in Louisiana, OpenAI’s Stargate, and Amazon’s Mount Rainier project in Indiana. The truth is even bigger AI datacenters are coming.
Colossus 2 (xAI)
Prometheus (Meta)
Hyperion (Meta)
Stargate (OpenAI, Oracle)
Mount Rainier (AWS)
Future of work: 👀
BigTech is Flattening Management again in 2025 with More Layoffs
The mega Amazon layoffs really puts into question the future of BigTech jobs. It’s also a trend to flatten management as capex will continue to rise significantly heading into 2026. More than 78% of the roles eliminated on Tuesday were held by managers assigned L5 to L7 designations, the internal data obtained by BI showed.
But Amazon is not alone, Google, Broadcom, Meta and others are doing the same though not at the scale of the Amazon layoffs which nears 4-5% initially scaling up to potentially 8-10% of its entire corporate workforce. Retail managers were hit the worst, but this could spread to AWS too.
While 14,000 is the official count, it could impact as many as 30,000 insiders say. As BigTech increase capex substantially in 2026, more layoffs could be coming. Google reportedly fired 35% of its managers in August, affecting those who ran smaller teams. Google recently began cracking down on remote work in its company culture. A tighter labor market, a mismatch between talent supply and demand, more net retirements, less immigration, more BigTech capex are all major factors in recent layoffs. Not especially AI. These are the largest layoffs in Amazon’s history.
Check most recent layoffs in Tech here.
Generative AI is Accelerating Hyperscaler Cloud Growth 🚀
Cloud growth and Capex surge ahead:
Google’s flagship AI app Gemini now has more than 650 million monthly active users. Google Cloud reported 32% revenue growth. They also updated 2025 Capex to a range of $91 billion to $93 billion.
Microsoft Azure Cloud posted 40% revenue growth. Revenue increased 18% in the fiscal first quarter from $65.6 billion a year ago. Microsoft will hold a 27% stake worth about $135 billion in OpenAI as they will likely do an IPO with a valuation of well over $1 Trillion. Microsoft reported capital expenditure of nearly $34.9 billion for its fiscal first quarter alone, concerning some investors.
Amazon’s AWS isn’t expected to report growth this fast yet again. Meanwhile Oracle is likely to make huge strides in 2026 although with a lot of debt. While earnings is reported after this article goes live, AWS is expected to bring in $32.4 billion, up 18% from the $27.5 billion the segment generated in the third quarter of last year. So while Microsoft and Google are growing closer to the 35% level, the leader AWS is growing at about only half that pace.
Google’s stock should be higher today, while Meta and Microsoft are likely to be down. GOOG 0.00%↑, MSFT 0.00%↑, META 0.00%↑.
Quotes on AI Infrastructure demand for compute acceleration:
“To date, we keep on seeing this pattern where we build some amount of infrastructure to what we think is an aggressive assumption, and then we keep on having more demand to be able to use more compute..” - Meta CEO (via The Transcript).
““We’ve been building TPUs for 10 years, so we now have seven generations in production for internal and external use. Our seven- and eight-year-old TPUs have a 100% utilization. And that just shows what the demand is. Everyone, of course, prefers to be on the latest generation, but they’ll take whatever they can get.” - Alphabet VP (via The Transcript).
OpenAI completes its for-profit recapitalization
OpenAI has completed its transition to a for-profit company, after court battles and public criticism from one of its founders, Elon Musk. It’s deal with Microsoft is coming to light this week.
Under the new structure, the non-profit OpenAI Foundation will have legal control over a public benefit corporation called OpenAI Group, which is free to raise funding or acquire companies without legal restraint.
The Foundation will hold a significant stake in OpenAI Group and will appoint its board of directors.
Microsoft said it’s invested a total of $13 billion into OpenAI, with $11.6 billion of that funded as of the end of September. That stake is now worth in the area of $135 Billion or 27% equity (likely to go up when OpenAI goes public).
Microsoft said that under the agreement, “OpenAI has contracted to purchase an incremental $250 billion of Azure services,” and that Microsoft will no longer have the right of first refusal as a compute provider.
Analysts say that Microsoft and OpenAI are still “key partners” in cloud and artificial intelligence, but they’re increasingly becoming competitors as well.
The deal will also extend Microsoft’s IP rights to OpenAI models through 2032.
OpenAI wants to build out a chip network that would consume 250 gigawatts (GW) of energy by 2033. That is equivalent to a fifth of America’s entire electric generation capacity. According to one of Scott Galloway’s Newsletters.
Alphabet’s Earnings were Better than Expected
Alphabet’s Monster Quarter
Alphabet’s crossed $𝟭𝟬𝟬𝗕+ in revenue in Q3. But it was the validation of its full-stack AI infrastructure bet.
Since Alphabet and Amazon have significant equity in Anthropic they are giving it a huge boost. 𝗧𝗣𝗨 𝗺𝗲𝗴𝗮𝗱𝗲𝗮𝗹: The Anthropic pact (”tens of billions”) means according to some analysts that this elevates Google’s custom TPUs from an internal tool to a major commercial alternative to NVIDIA’s GPUs. I’m not sure that is correct, but it is notable. In 2026 there will be significantly more AI chip makers.
Google Cloud is Accelerating Growth
Jobs in the AI Era
Future of work:
There’s not much real evidence Gen AI is impact jobs, however:
New graduates are struggling to find entry level jobs.
AI coding is making it harder for Software Engineers to start their careers.
BigTech companies are doing more layoffs, in part to invest more in AI Infrastructure.
Other companies doing massive layoffs in 2025: Amazon, Intel, UPS, Salesforce, Ford, Target and Paramount.
Peter Thiel Backed ASML Disruptor Substrate Comes out of Stealth ⚗️
Substrate intend to utilize X-rays over EUV for lithography, claiming it to be a cheaper alternative. Trying to disrupt the likes of ASML and TSMC in the process.
They have recently raised $100 million in a funding round, which valued it at $1 billion valuation. The round saw participation from Peter Thiel’s Founders Fund, General Catalyst, Allen & Co., Long Journey Ventures, Valor Equity Partners, and the CIA-backed not-for-profit firm In-Q-Tel.
In an interview with the Wall Street Journal, Substrate co-founder and CEO James Proud – described by the outlet as a “protégé of investor Peter Thiel” (these protégés are multiplying rapidly!) – explained that the startup’s tool uses x-ray light to etch microscopic patterns onto silicon wafers. Substrate is nearly four years old. Substate’s machine uses a particle accelerator to create a light source from shorter-wavelength X-rays — something that Proud said allows for the creation of narrower beams.
“Our light source begins with radio-frequency cavities accelerating pulses of electrons using powerful electric fields. The electrons ride each successive wave, gaining energy and increasing their velocity to near the speed of light,” Substrate says on its website.
Tesla’s Share of the EV Market is in Free-Fall in 2025
Can Tesla scale robotaxis and robots fast enough to make up for the fall in EV sales and saturated EV global marketshare? It’s a tough sell, I’m bearish on this.
China E-commerce Dominates South-East Asia
According to a CNBC article they noted that in Indonesia, Thailand and the Philippines, Chinese e-commerce players have quickly come to dominate around half of the online shopping market, Bain and Company said in a report Thursday.
The U.S. lags in Renewable Energy in a huge Industrial Policy Mistake as Energy Bills will Skyrocket in the AI Datacenter Era
Microsoft increased its AI Infrastructure capacity by 80% this year in 2025, and nearly double their datacenter footprint over the next two (2026 to 2028) according to a post by Satya Nadella. AI Infrastructure, the exponential demand for compute, and Nvidia carrying a Semiconductor boom continues to be the main story of Earnings.
2025 has been an incredible year for AI
Thanks for reading! Next week we’ll look at some of the top AI books and continue to explore Anthropic’s Agent skills among other stories I’ll be watching.
What really gets me about NVIDIA's trajectory is how they've positioned themselves at the convergance point of every major compute trend - not just AI datacenter GPUs but now quantum interconnects, 6G telecoms infrastructure, and physical AI. The 66-day jump from $3T to $5T is absolutly wild, and while Jensen keeps saying there's no bubble, those numbers feel more like a market reflecting future infrastructure build-out rather than current fundamentals. The projection that AI infrastracture spending could hit $3-4 trillion by 2030 basically validates NVIDIA's premium valuation if they can maintain that 94% market share in AI chips. But I'm curious how sustainable that moat really is once we see more competition from custom silicon and alternative architectures in 2026-2027.