I'm Chris Collard: landman, RPL, and I run Silicon & Shale. All things AI and O&G, from the lease line to the data center.
Joined January 2026
- Tweets64
- Following53
- Followers21
- Likes400
Pinned Tweet
I’m building an operating system for land and ownership across physical energy infrastructure.
Oil & gas, wind, solar, transmission, data centers.
This account documents the build.
Venezuela says it expects ~$76B in oil & gas investment over 5 years: 61 deals signed, plus a U.S. accord covering 17 fields and plans for 250 rigs. The rig count is the number to watch. That's where projections turn into barrels.
H/T @Polymarket xitter.cf/Polymarket/status/2111…
JUST IN: Venezuela announces plans for nearly $76,000,000,000.00 in oil investments over the next 5 years.
Anthropic signs its first Canadian data-center lease: 95 MW at EverWind's Point Tupper, Nova Scotia campus (late 2027), part of a C$10B+ plan with new wind. Anthropic will fund grid upgrades and shield ratepayers.
Via @Polymarket
xitter.cf/Polymarket/status/2111…
globenewswire.com/news-relea…
Chevron joined an AI subsurface collab with a signature. Its 2.67 GW gas plant for Microsoft waits on a Texas notary stamp.
Silicon & Shale Ep19 — Two Chevrons: siliconandshale.substack.com…
Chevron joined an AI subsurface collab with a signature. Its 2.67 GW gas plant for Microsoft waits on a Texas notary stamp.
Silicon & Shale Ep19 — Two Chevrons: siliconandshale.substack.com…
NVIDIA open-sourced full Run:ai GPU orchestration (DSX AI Factory)—more work from the same rack power via better utilization.
Via @NVIDIAAIInfra
xitter.cf/NVIDIAAIInfra/status/2…
developer.nvidia.com/blog/nv…
Today, we’re open sourcing the full NVIDIA Run:ai platform as part of the NVIDIA DSX AI Factory Platform.
Developers, platform teams, and partners can build on production-grade AI workloads and GPU orchestration to improve GPU utilization, increase workload throughput, and bring AI factory solutions to market faster.
Deploy the integrated NVIDIA Run:ai platform or adopt individual components to fit your needs.
See what’s included, how to get started, and how to contribute: nvda.ws/4hRTubt
ALT NVIDIA Run:ai Is Now Open Source as Part of NVIDIA DSX AI Factory Platform
FERC accepted PJM's 50 MW Large Load registry (live Oct 12) and rejected curtailing new data centers early in grid emergencies, citing state retail authority. Clearer lanes for big loads.
Via @Polymarket
xitter.cf/Polymarket/status/2109…
FERC ER26-3515: elibrary.ferc.gov/eLibrary/d…
NASA and DOE just signed an MOU on space nuclear power and propulsion, backing Space Reactor-1 (Mars, 2028) and Lunar Reactor-1. Compact, safe reactors built for space make the same tech easier to deploy here, too.
via @NASAAdmin nasa.gov/news-release/nasa-e…
Google put four TPUs in orbit Oct 1 to test whether AI compute can run in space. The draw: up to 8x the solar power of a panel on the ground. The catch: no air, so all the chip heat leaves through radiators. A research bet, not a power plan.
via @Google
xitter.cf/Google/status/21078752…
Project Suncatcher is our moonshot exploring whether we can one day host machine learning infrastructure in space 🚀
To do so, we need to know whether our AI hardware can operate in orbit. Can our TPUs handle the physical stress of spaceflight and the radiation and thermal extremes of space? These are the questions Google researchers have been exploring for the past few years.
Last week, we launched our first test satellite carrying four TPUs into orbit — putting us one step closer to getting answers.
ALT An infographic against a black background depicting the Earth below states that a rocket trip to low Earth orbit takes about 10 minutes, exposing spacecraft carrying TPUs to acceleration loads up to 10x gravity, and individual TPU chips to forces up to 50–100 g.
ALT A person wearing glasses works with specialized laboratory equipment under a red laser line below text stating that Trillium TPUs can survive a radiation dose equivalent to a five-year space mission.
Off, Manual, Auto. AI is getting a hand on the setpoint: plunger lift, refinery agents, and Copilot's new "Always allow" button.
Silicon & Shale Ep18: siliconandshale.substack.com…
~90 GW of behind-the-meter gas announced for data centers. ~2 GW running. Texas just froze data-center permits.
Silicon & Shale Ep17 — Gas-to-GPU, announced vs. running: siliconandshale.substack.com…
Anthropic widened its Cyber Verification Program to three tiers. Municipal utilities and critical-infrastructure operators are named for Defense Access; power-grid testing stays in a top tier vetted with the US government.
via @AnthropicAI
xitter.cf/AnthropicAI/status/210…
We’re expanding our Cyber Verification Program to give security professionals broader access to our most capable models.
Through this program, verified security professionals can access Claude Mythos 5.1, Opus 5.5, and Sonnet 5.5 with safeguards designed for defensive work.
We’re also opening up new tiers to allow for authorized offensive work, like penetration testing and red-teaming.
anthropic.com/news/cyber-ver…
Altman to POLITICO Decoded: "the world should accept some bad things happening" for AI's benefits. Still "a lot of daylight" vs Anthropic, even as OpenAI backs outside safety evaluators.
via @AndrewCurran_
xitter.cf/AndrewCurran_/status/2…
White House “Super Intelligence Force”: DNI Jay Clayton as AI czar, 120 days on SI risks + federal role (WSJ). Charter: lead on SI, avoid overregulation.
via @AndrewCurran_
xitter.cf/AndrewCurran_/status/2…
Epoch AI: cost of a given AI performance level down ~47%/quarter since 2023 (~13×/year) — faster than DNA sequencing, compute, batteries, or electricity ever fell.
via @EpochAIResearch xitter.cf/EpochAIResearch/status…
TotalEnergies x Mistral: frontier AI for reservoir exploration — EUR 100M+, operator-side IP.
New Silicon & Shale Ep16: siliconandshale.substack.com…
EU proposes A–G energy/water efficiency labels for data centres ≥500 kW — transparency before the grid/water fights harden.
via @Polymarket xitter.cf/Polymarket/status/2102…
US–China AI + critical minerals talks land at JPMorgan Manhattan today — the mineral line is the physical bottleneck under the model stack.
via @Polymarket xitter.cf/Polymarket/status/2101…
Anthropic skipped UK AISI pre-release tests on Mythos 5.1 — advisors in London, model access still a US call.
Verified layer is testing access (FT trail), not open weights. Via @moonshots_pod xitter.cf/moonshots_pod/status/2…
Oil patch meets algorithm: Ep15 live — this week’s E&P developments + applied AI.
siliconandshale.substack.com…
🤖 Made with AI
Sept 25. One theater in downtown LA. $3.5M for a film about a future worth living in, $2M for AI builders, decided live. I'm going.
@moonshots_pod @PeterDiamandis
🤖 Made with AI
