Energy Abundance, AI Power & Data Centers, NBA, & Rap posts. Background in equity & credit markets, startups, & IB. Personal views NOT employers. NOT advice.
Chicago, IL
- Tweets 30,730
- Following 1,906
- Followers 35,273
- Likes 164,432
Have to imagine the slate has been ready to go for a while now with this either or but feels like you have to be comfortable in the product pipeline and compute availability to even suggest it
Over the next 28 days, each day we’ll either ship one thing that is a clear improvement and relevant for most codex/work users or ship a full reset. Let the improvements begin.
6
2,164
Not to be a cheerleader but need to celebrate the positive news & announcements of hypers commitments to being good partners.
Amazon’s $1bn+ commitment over five years to U.S. data center communities is welcome. Alongside Microsoft’s earlier commitments, it reflects broader hyperscaler responsibility for the communities hosting this infrastructure.
Built Together includes:
>Free community college access for 300,000+ students, covering costs remaining after financial aid.
>Expansion to 25 training centers, targeting up to 100,000 learners annually by end-2028.
>Efficiency upgrades for 30,000+ homes and 300+ schools/community buildings, targeting 20–40% energy savings.
>Locally directed grants for priorities including housing, roads and emergency services.
Amazon also commits to ending new government NDAs, engaging communities earlier, publishing annual energy/water metrics and paying enough for electricity to cover required infrastructure improvements.
The limits: these are funding commitments and participation targets, not guaranteed jobs or realized savings. Amazon hasn’t pledged to forgo tax incentives or revoke existing NDAs; contractor coverage remains unclear. Protecting other ratepayers still depends on the utility tariffs and regulatory decisions implementing its power-cost commitment.
3
1
10
2,196
Amazon is exploring an ~$8B sale-and-leaseback of Nvidia chips, per FT. It would raise cash against installed hardware while keeping the compute and diversifying its funding beyond corporate bonds. Amazon would also offer investors an equity stake of up to 10% in the vehicle, while retaining no ownership itself.
The structure resembles CoreWeave’s GPU-backed financing, where contracted customer payments support repayment. FT says investors expect an investment-grade rating linked to Amazon’s credit, potentially attracting insurers and pension funds.
Nvidia has offered guarantees against falling hardware values on some financings. None has been reported here, so allocator appetite would depend on Amazon’s lease payments and how much chip-value risk investors are willing to take.
Talks remain exploratory + investors and terms are undisclosed.
5
6
29
5,952
Because it got good and better than most people or systems lol
At this point it’s more about context and specific SME (which it can get to with appropriate context + learning from you over time)
Less hallucinations and more misunderstanding your intention, overcorrecting to a prompt, or making the wrong assumption that an SME wouldn’t do or knows better to avoid
3
1,777
I think the reason people love Opus 5.5/GPT 6 Astra/Sol 6.1 is that they are both insanely intelligent but also token efficient + fast.
Prior iterations of models thought for a really long time and you got a quality answer but felt more like 'consult with this model' when dealing with a hard problem vs 'working with this model' naturally and iterating on stuff.
Opus 5.5 and Sol 6.1/Astra are the first iterations of insanely capable and intelligent models that even on non-fast mode or especially with Fast mode, they just 'get it' with limited context, get things done efficiently, know when to layer in additional effort, and output quality is always impressive.
3
3
19
3,003
People building in finance & AI… here’s a great product roadmap to aspire to get to
I wonder what's next. I suspect one interesting endpoint is that the excel model becomes the quantitative surface of a much deeper research graph / self-constructing ontology. I could imagine something like every quantitative revenue key driver being hyperlinked to a graph of the intermediate qualitative assumptions that underpin it. Maybe analysts use some kind of ambient capture during their desktop and field research process, and then there's an agent-mediated live evidence stream (e.g, expert calls, filings, transcripts, podcasts, articles, tweets, etc.) that supports or contradicts the analyst's assumptions in real time. And based on that you could calibrate your confidence about various qualitative assumptions at any given point (as new evidence comes in) and the model could compute how that flows through to quantitative KDs.
Overtime, this surface / system would learn how new evidence shifts your priors and how accurately your updates actually end up representing economic reality. It could then start to understand how to weight different pieces of qualitative evidence!
Lots to think about.
4
5
72
9,629
400MW is roughly a fifth of Japan’s ~1.8GW operating data-centre market by headline MW, before adjusting for facility overhead.
But JERA/Dell/RHAELM announced an MOU, with Apollo’s financing role described as an intention. The $140bn / 3–4GW ambition implies roughly 8–10 Chiba-sized projects before the first has been built.
BTM has potential in Asia and elsewhere, but delivery at that scale remains a huge maybe pending binding commitments and financing.
JERA, Dell, and RHAELM signed an agreement for a $15 billion, 400-megawatt behind-the-meter data center campus at JERA's Chiba plant in Japan. JERA plans to develop 3 to 4 gigawatts over five years, representing up to $140 billion in total capital outlay. Japan's installed data center capacity currently stands at 1.7 gigawatts, compared to 29.2 gigawatts in the United States.
SemiAnalysis models one utility gigawatt of NVIDIA B200 GPUs running DeepSeek-V4.1-Flash to generate $15.2 billion in annual revenue at official list rates. With fully loaded annual TCO at $8.9 billion, operating profit margins reach 42%. Offloading 196-billion-parameter Engram tables into host DRAM frees GPU capacity, boosting throughput and revenue per gigawatt by 50%.
1
1
5
2,361
One notable trend that goes against my priors is seeing a lot of long time veterans from the hypers move to Neo clouds
That doesn’t mean it’s bulletproof or anything but people with many years and in senior positions of the companies with the most flexibility are moving to levered bets on the AI buildout is notable
They’re assessing the pathways and making their bets
We’ll see how it pans out
7
6
48
5,480
Scraped together via GPT & Claude and scanned for a few of the folks I was thinking of. Just open up LinkedIn and you’ll see plenty of announcements.
Notable hyperscaler talent moves, Oct 2023–Oct 2026 generally
NEOCLOUDS
CoreWeave
• Mar 2024: Nitin Agrawal, Google Cloud VP Finance → CFO.
• Jun 2024: Chetan Kapoor, AWS accelerated computing → Chief Product Officer.
• Aug 2024: Chen Goldberg, Google Cloud VP/GM Kubernetes and Serverless → SVP Engineering.
• Aug 2024: Sachin Jain, Oracle SVP AI Infrastructure, previously Google Cloud capacity leadership → COO.
• Jan 2025: Corey Sanders, Microsoft CVP Cloud for Industry and longtime Azure leader → SVP Strategy.
• Sep 2025: Jon Jones, AWS Global VP Startups and Venture Capital → Chief Revenue Officer.
Nscale
• Dec 2025: Nidhi Chappell, Microsoft CVP Azure AI/HPC Infrastructure → President, AI Infrastructure.
• Apr 2026: Sam Huckaby, Oracle SVP OCI Data Center Infrastructure → President, Data Centers.
• Oct 2026: Justin Osofsky, Meta Chief Partnerships Officer and former Instagram COO → COO.
Crusoe
• Aug 2025: Erwan Menard, Google Cloud AI product director → SVP Product Management.
Lambda
• Mar 2024: Vijay Manyam, Microsoft engineering director, Cloud AI and Advanced Systems → VP Supply Chain and Manufacturing.
• Feb 2026: Jerry Hunter, former AWS global data-center leader, subsequently Snap COO → Vice Chairman, Compute Delivery. Exception: left AWS Oct 2016.
PE-BACKED AI INFRASTRUCTURE
Crux AI, Google/Blackstone
• Sep 2026 public launch: Benjamin Treynor Sloss, former Google engineering VP → CEO.
• Sep 2026: Alan Duong, Meta VP and Head of Data Center Engineering and Construction → Chief Development Officer.
• Oct 2026 announcement: Bikash Koley, Google VP Global Infrastructure and Capacity → Chief Technology & Product Officer.
Helix Digital Infrastructure, KKR-backed
• Jun 2026 launch: Adam Selipsky, former AWS CEO → Co-founder and CEO. Left AWS Jun 2024; became a KKR adviser in 2025.
DATA-CENTER DEVELOPMENT AND ENERGY
• Feb 2024 founding: Brian Janous, Microsoft VP Energy → Co-founder, Cloverleaf Infrastructure. Exception: left Microsoft Jun 2023.
• Jul 2024: Nur Bernhardt, Microsoft Director of Energy Strategy → VP Energy, Cloverleaf Infrastructure.
• Jun 2025 public launch: Patrick Yantz, Microsoft data-center infrastructure executive → Co-founder, GridFree AI. Left Microsoft late 2024.
• Nov 2025: Sean James, Microsoft Senior Director of Energy and Data Center Research → Distinguished Engineer, Energy Systems, NVIDIA.
• Jan 2026: Maud Texier, Google Director of Data Center Energy and Infrastructure, EMEA → Global VP Data Centers and Industrials, Octopus Energy.
• May 2026: Bobby Hollis, Microsoft VP Energy → Chief Development Officer, STACK Americas.
• Sep 2026: Jim Collins, Microsoft GM Global Energy Markets → SVP and Global Head of Energy, Digital Realty.
SPONSOR-BACKED DATA-CENTER PLATFORMS
• Jan 2024 announcement: Adam Black, Google Mid-Atlantic Data Center Services Manager → Head of Design and Construction, TA Realty’s data-center platform, subsequently TA Digital Group.
• Jan 2026 announcement: Nathanael Liu, Microsoft Cloud Operations + Innovation strategic planning → Global Product Lead, Ada Infrastructure, part of Ares.
AI-LAB INFRASTRUCTURE TEAMS
Anthropic
• 2025: Winnie Leung, Google data-center engineering/operations → Head of Data Center Infrastructure.
• Oct 2025: Adam Johnson, Google data-center electrical engineering → Data Center Electrical Lead.
• Jan 2026: Liwen Mao, Google Staff Data Center System Architect → Data Center Design Lead.
• Jan 2026: Peter Sarossy, 20-year Google veteran → Data Center Security Engineering.
• Apr 2026: Sana Ouji, Google Data Center Energy Strategic Investments and Partnerships → inaugural energy team.
• Apr 2026: Eric Boyd, Microsoft Azure AI leader → Head of Infrastructure.
• By Apr 2026: Zach Miller, 17-year Google veteran → Data Center Operations
2
2
1,190
… Manager.
• By Apr 2026: Scarlett Chen, Google Data Center Engineer, Advanced Technology and Innovation → Member of Technical Staff.
• May 2026: Mike Brinker, Google data-center services/construction/program management → data-center team.
• Jun 2026: Marcus Fontoura, Microsoft Azure Core CTO → Member of Technical Staff.
• Jun 2026: Asad Malik, Google Head of AI/ML Infrastructure and Strategy → Head of Compute Finance and Strategy.
Mistral Compute
• Sep 2026: Marc Oman, Google European data-center energy sourcing → Mistral Compute.
HYPERSCALER → HYPERSCALER
• Feb 2024: Ahmed Shihab, AWS VP Infrastructure Hardware → Microsoft Azure CVP Storage.
• Oct 2024: Hayete Gallot, Microsoft commercial leadership → Google Cloud President, Customer Experience. Returned to Microsoft as EVP Security in Feb 2026.
• Mar 2025: Umesh Shankar, Google Cloud Security Chief Technologist/Distinguished Engineer → Microsoft AI CVP Engineering.
• May 2025: Patrick Taylor, Microsoft Director of Commercial Structuring, Nuclear and Carbon-Free Energy → Google Commercial Lead, Advanced Energy Technology.
• Jul 2025: Jeesoo Lee, Google Director of Global Machine Learning Planning → Meta Director of Product Management, Infrastructure.
• Oct 2025 announcement: Richard Hage, Microsoft Director of Global Strategy, Data Center Engineering → Oracle VP Data Center Design. Subsequently moved to Nscale in 2026.
• Dec 2025 announcement: Gary Demasi, Google Global Director of Data Center Energy and Location Strategy → Meta VP Data Center Development and Strategy, via six months at MGX.
• May 2026: Shawn Bice, Microsoft CVP Security Platform and AI → AWS VP AI Services.
• Jul 2026: Rudra “Rudy” Mitra, Microsoft Purview leader → AWS VP Security Services.
• Oct 2026 scheduled start, announced Sep: Ray Fakhoury III, AWS Energy Policy Manager → Microsoft Director of Global Energy and Sustainability Policy.
REPORTED MOVE
• Jul 2026: Dave Brown, AWS SVP Compute, AI and Platform → Meta infrastructure/data-center buildout, reported by WSJ.
BOARD APPOINTMENT
• Mar 2026: Nick Clegg, former Meta President of Global Affairs → Nscale board. Left Meta Jan 2025.
DEPARTURES WITHOUT A VERIFIED NEXT DESTINATION
• Jul 2025: Kevin Miller, AWS VP Global Data Centers, after 17 years.
• Jan 2026 report: Jennifer Weitzel, Microsoft data-center development executive, following her return in 2025.
• Mar 2026: Urvi Parekh, Meta Head of Energy, after nearly 10 years.
2
928
this chart should be the death knell for the 2010s tech left whose theology compelled them to argue that Uber was some sort of labor law arbitrage that created no productivity benefits. the truth is that it created a far larger market than taxis ever had. the industrialists win
131
308
4,758
211,967
it is nice to talk to one single entity/chat but
if your projects focus on variety of things + lot of loose threads simultaneously, it gets really crammed in a single messaging interface
Before:
You manage your Chat/Codex threads
Now:
You delegate to your dot, that manages your Chat/Codex threads
Delegation isn’t for everyone or everything (I mean this earnestly) - but when you want to work this way, it’s a magical experience.
I’ll share some demos later on how this is a fundamental leap in usefulness!
5
2,029
Pre 5.5 it was everyone saying how Claude was dead in the water and that Sol and Astra had complete mindshare + Codex was the best harness by a mile
This app swings the pendulum too far at any sign of competitive pressure
OpenAI is in an extremely weird pickle.
Anthropic has better models.
SpaceX has better agents.
Meta has bigger reach.
Their saving grace is their first mover’s advantage, but their products have fallen behind.
Curious to see how they navigate this.
3
16
3,686
Citadel is impressive man, consistently putting up #s
We have some initial hedge fund returns for September. Looks like machines scored a win over humans. Traders faced a volatility month as inflation concerns pushed bond yields to multidecade highs and hit fixed-income wagers. bloomberg.com/news/articles/…
2
34
9,269
You literally could open this app and see 100 of these
“The market is asleep at the wheel. I’ve found the next HUGE AI bottleneck. Follow me for more posts like this”
*stock up 5,000% in the last 6 months already*
**half the time it was literally just pitching nvidia again**
I'm personally very happy that we're past the phase where every AI bottleneck company went up/down more than 15% per day from literally zero catalysts.
Things seem a bit more normal for now.
2
2
12
2,669
Men will literally offer infinite money from a mysterious portco with billions of capital behind it for the most precious resource in all of technology & capital markets right now instead of going to therapy and it’s gross
Looking for 500MW (minimum 250, max 2GW) of powered land for a PortCo.
Company is completely price insensitive and has billions behind it. Ideally electrified shell is already built.
I understand everyone wants this, but figured I’d throw it out there!
1
25
4,981
Brother, what will your ARR do
That answers your capex question
Anthropic's Sholto Douglas says AI capex could rise from $1 trillion this year to $4 trillion in 2028 and begin doubling humanity’s GDP by the early 2030s
“Over the course of the last four or five years, we’ve been 2x-ing or 3x-ing the amount of compute capacity devoted to AI every year."
"And so I think that’s roughly, together, the hyperscalers are spending about a trillion dollars this year on capex related to AI. A very interesting question will be, will that trend line hold, and so will it go to $2 trillion next year and then $4 trillion in 2028?"
"And if that trend line broadly holds, and you can expand that to encompass the broader robotics industry, then that means that in the early 2030s, you start to get to the point where you’re actually effectively doubling the GDP of humanity, which, again, is a little bit of a ridiculous concept, but requires a lot of things to go right.”
3
1
31
7,909
I fully agree here but I think one thing that’s under appreciated is people talking about the duration mismatch
Lot of forward credit for TAM penetration x steady state economics when in reality extremely early in usage and adoption
That means compute demand going up but it does not necessarily mean the market is deep enough *today* to sustain continued scarcity pricing on compute
Seeing this tension for the OAI and ANT IPOs hitting some friction. Many see the vision but also already bought or funding at valuations that assumed more progress
I think people don’t talk about that tension enough
This is my thesis: no one uses AI. I repeat, absolutely no one. We live in a bubble.
Even among my friends who pay for it, when I ask them to open ChatGPT and show me their queries, it’s the same handful of basic things.
Most don’t even know they can upload a photo and ask questions about it. Connecting Gmail so an agent can read and send emails blows their minds. An agent opening a browser and checking them into a flight? They’ve never even heard of it.
The massive challenge right now is adoption, and then getting people who already signed up to actually use what they’re paying for. Most have absolutely zero clue what’s possible.
Imagine the compute shortage when everyone starts using AI like the top 1% of users do today.
7
2,815