# AI capex — X 热门讨论 (2026-10-11 06:54 UTC)
## @0x_zoda (ZODA) · 10-11 05:06 · ♥31 ↻2 💬32 gm gVangrid!
Vangrid @vangrid_io: The Spatial Cortex for Physical AI
LLMs hit the text data wall. The next frontier is Physical AI robots, autonomous agents, world models, and sovereign systems and they are starving for real-world spatial ground truth.
Satellites see roofs. Corporate mapping fleets see roads they can drive. Neither sees a stairwell, a loading bay, a market street, or the inside of a warehouse. Vangrid collects that view.
It is a decentralized data rail that turns ordinary smartphones into edge nodes. Contributors record short multi-view videos with a normal phone.
Faces and license plates are blurred on-device before anything leaves the phone. Every capture is fingerprinted at the moment of recording content hash + coarse location + timestamp
Those fingerprints are batched and anchored on Base (via Ethereum Attestation Service / Merkle roots), so buyers can verify provenance against a public ledger instead of trusting a company.
The videos are reconstructed into 3D models and delivered as structured, provenance-stamped spatial data through an API.
Three roles: - Contributors walk the grid, record captures, and fulfill bounties.
- Requesters / organisations post and fund specific capture bounties.
- Agents can now hire humans with one HTTP call. An AI agent posts a bounty, pays in USDC on Base or Arc (via x402), a person films the place, and the 3D reconstruction comes back through the same API no account, no API key. The first physical-world service an agent can pay for.
The thesis is simple and brutal: 3 billion civilian smartphones become a zero-CAPEX sensor swarm for continuous, ground-level, multi-view ground truth.
Privacy is enforced at the edge. Integrity is enforced on-chain. Scale comes from the phones already in people’s pockets rather than expensive dedicated fleets.
Traction (as of late 2026): - 100K+ verified captures - Live bounty system with real settlements on Base - Public explorer https://t.co/TDyYHpc0HR showing captures, attestations, and grid events - $9M seed round (SAFT-structured) announced August 2026 with HashKey, Borderless, https://t.co/UwkP5CiCjH Capital, Animoca Brands, Gate Labs, Mapleblock Capital and others as key stakeholders - ~80k followers and growing attention from physical AI / DePIN / robotics circles
Strategic domains: sovereign defense, autonomous logistics, critical infrastructure, and embodied AI / robotics.
The pitch is that this becomes the spatial data layer world models and physical agents train and operate on continuously updated, verifiable, and sovereign rather than locked inside one company’s mapping cars or satellite passes.
The grid is inevitable.
Docs: https://t.co/XaOeDCzyxa App: https://t.co/Xu2TgALvFk (capture or buy data) Explorer: https://t.co/TDyYHpc0HR
Physical AI needs to know what the world actually looks like from the ground. Vangrid is building the rail that delivers it. https://x.com/0x_zoda/status/2109148496435503116
## @trevornoren (Trevor Noren) · 10-10 22:50 · ♥31 ↻9 💬3 The "crowding out" threat if AI doesn't deliver productivity gains soon: "The AI complex can go to markets and pay what it takes to get the capital they need, because they expect—and investors believe—that money poured into AI capex will deliver enormous returns in the future. Firms in other parts of the economy can’t promise those same returns, and so they are relinquishing their investment capacity to all things AI...This could be a problem for the economy. Sure, AI execs believe that future returns will be massive, but they could be wrong. And if AI doesn’t deliver on its promise then we face a double blow to growth: because AI didn’t expand our productive capacity as hoped, and because other investments which might have didn’t take place."
As I dissect in my report on "The AI Trade" (https://t.co/wQQNniS2kj), skilled construction labor exemplifies the "crowding out" impact of AI:
"The US construction industry will need ~349k net new workers in 2026 and 456k additional workers in 2027—on top of normal hiring—just to keep supply and demand in balance. It’s unlikely that balance is achieved. ~30% of construction workers in the US are foreign born and net migration was cut in half comparing 2024 to 2025. Meanwhile, ~20% of the construction workforce is over the age of 55. Workforce aging is even more problematic in the more skilled trades. For one, nearly half of US engineers are now 50 or older.
The skilled construction labor shortage is particularly problematic for data centers. According to Uptime survey results, ~60% of data center firms report that labor supply is a top reason for project delays. Data center builders face two core challenges: a limited supply of skilled labor and a mismatch between where data centers are being constructed and where that labor lives. As Bernstein noted in a report last month:
"The US MEP labor pool is smaller than the 1.8m headline labor force suggests…Not all MEP workers are qualified, geographically accessible or available. Data center work requires a specialized set of skills, so even after narrowing this pool to [non-residential] construction (790,000), the pool needs further refinement given the high training bar. There is a geographic mismatch between where the data center projects will be built versus where the labor is—only 30% of the MEP labor resides where 70% of the projects are located."
Even beyond the productivity implications of AI "crowding out", there are also rate implications. To quote Apollo: "AI's rate-insensitive boom is triggering "Dutch disease" in the US economy by draining capital, power and labor from vulnerable sectors like housing and autos. Trapped between inflation driven by the AI boom and weakness in rate-sensitive industries, the Fed will have to keep interest rates higher for longer."
FT link: https://t.co/al9SmyrX70
Apollo link: https://t.co/w8hPe1w6P4 https://x.com/trevornoren/status/2109053952583196884
## @51bodila (bodila) · 10-10 23:05 · ♥31 ↻2 💬6 Finally... my Gavin Baker GrokBot just replaced ~$350k/year AI Quant Researcher
I packed Gavin Baker’s entire AI playbook and mind into ONE agent - now 99% of AI trading/research agents look blind, slow, and expensive
Run it below (free):
→ studied 6 years of Gavin Baker’s interviews and analyze every post he makes on X, the moment he makes it
→ tracks every dollar of AI capex from Microsoft to the last chip supplier while you sleep
→ research the whole AI stack, public and private, from power plants to OpenAI - and gives profitable ideas
→ turns 6 years of Baker’s thinking into one simple answer: buy, wait, or run
→ And plenty more Baker rules in a bot that thinks like him..
Try this bot, don't forget me later - then read the article on how to build swarm of AI trading agents ↓ > 引用 @0xCodila: Grok Bot Agents: Build 24/7 AI Agent Graph That Automate Your Life (The Ultimate 5-Step Guide) https://x.com/51bodila/status/2109057759194784158
## @dons_korea (DONS) · 10-10 23:39 · ♥31 ↻1 💬4 [ 강세장은 맞지만 단기 위험 대비 수익은 불확실.. S&P500은 사실상 AI 한 종목 ]
골드만삭스의 토니 파스콰리엘로는 두 달 동안 상승 추세가 살아 있으니 주식을 계속 들고 가라고 조언해왔다.
S&P500은 이번 주 7,818로 사상 최고치를 새로 썼지만, 그는 최신 보고서에서 어느 때보다 조심스러운 입장을 보였다.
상승 추세는 그대로라고 보면서도 단기 위험 대비 기대수익은 불확실하다며 속도를 조절하라고 했다.
지수 상승은 극히 일부 종목이 이끌고 있다.
8월 말 이후 S&P500은 1.28% 올랐지만 AI 관련주를 뺀 지수는 5.19% 떨어졌다.
3분기 S&P500 이익 증가율은 28%로 예상되지만.. 중간값 기업은 9%에 그친다.
이익 증가분의 68%를 상위 10개 기업이 만들고, 이중 절반을 마이크론과 엔비디아 두 곳이 차지...
골드만은 이번 분기가 이익 증가율 정점이고 빅테크 CAPEX 증가율도 이번 분기에 정점을 찍을 것으로 본다.
헤지펀드 포지션도 한쪽에 쏠려 있다.
미국 주식 순노출은 1년, 3년, 5년 최저 수준이지만 남은 비중은 대형 기술주에 몰렸다.
M7 비중은 미국 전체 노출의 22%로 2022년 집계 이후 최고이고, 반도체 비중은 연초의 두 배인 12%.
반도체 레버리지 ETF 자산도 약 1,150억 달러에 달해 하락할 때 낙폭을 키울 수 있다.
가장 크게 달라진 부분은 금리에 대한 시각이다.
열흘 전만 해도 금리 상승을 주식에 부담 정도로 봤지만, 이번에는 국가와 기업 모두에게 더 큰 위험이 되고 있다고 평가했다.
30년물 금리는 2000년 이후 최고인 5.618%를 기록했다.
골드만은 국채시장이 AI 기업들이 금리에 부담을 느끼는 수준을 찾거나, 경제 어딘가에서 문제가 생기는 수준을 찾아갈 것으로 봤다.
파스콰리엘로는 한 투자자의 말을 인용했다.
"이건 결국 하나의 거대한 거래일 뿐이야, 알고 있지?"
상승 추세, 포지션, 이익 증가, 금리 상승 모두 결국 AI라는 하나의 주제로 연결돼 있다는 뜻이다.
그는 미국과 일본 주식을 계속 보유하되, 금리 상승에 베팅하는 포지션을 함께 가져가라고 조언했다.
주식과 채권이 함께 떨어질 수 있다는 점을 인정한 셈..
이제 시장의 다음 시험대는 10월 마지막 주 빅테크 실적 발표로 이어진다. https://x.com/dons_korea/status/2109066364350878103