# AI capex — X 热门讨论 (2026-10-01 20:48 UTC)
## @a16z (a16z) · 10-01 20:10 · ♥72 ↻9 💬9 Where does money invested into the AI buildout actually go?
For every $100 flowing into the supply chain: - $50 to chips - $20 to power - $15 to networking - $15 to cooling, buildings, and land
More charts in State of Markets II: https://t.co/MTaxKUxa2w https://t.co/f0t2IBuBaV > 引用 @DavidGeorge83: Introducing our State of Markets pt 2, along with a companion podcast where we unpack the data and discuss what comes next.
Tech is the everything cycle.
Supply: putting the buildout in context, just passed railroads as % of GDP. The wisdom of Elon is real: the factory (or the datacenter!) is the product.
Demand: diffusion is so, so early. Median AI vendor spending in the top 1% of companies is 8x that of the top 10%. Only about 30% of S&P 500 companies report a quantified AI impact, which means there’s a substantial opportunity in connecting models to a company's data and workflows. Diffusion into companies is one of the main themes of the next 5 years.
We’re entering the agent work period. Only a few million users today, but applicable to billions of internet users with massive surplus created. META/GOOG monetize US users at $200+ per year today. Agent opportunity is much higher.
Mega-trends the next 5 years: Consumer agents, Robotics, Autonomy, AI x bio, Personal health, Diffusion into enterprise, New era of American Dynamism.
Much more in our SoM report here - https://t.co/K4vOy4x3J6
@a16z @sarahdingwang @aleximm @santiago__rdz https://x.com/a16z/status/2105752116757422375
## @Ross__Hendricks (Ross Hendricks) · 10-01 20:15 · ♥45 ↻6 💬3 It's almost like the AI industry is locked into a race to the bottom in pricing, in a highly-capital intensive industry, where capital providers create a massive consumer surplus at the expense of negative ROI.
In other words, just like every other capex mania in history. > 引用 @arakharazian: NEW: AI Spend FALLS in the latest Ramp AI Index.
Price cuts at the frontier, in addition to cheaper and more efficient models at standard and lite levels, are pushing down the cost of using AI.
Open source models remain <5% of business spend. This decline in spend is driven almost exclusively by competition between OpenAI + Anthropic. https://x.com/Ross__Hendricks/status/2105753430069244215
## @SmallCapSnipa (Small Cap Snipa) · 10-01 19:24 · ♥33 ↻0 💬5 Bernstein: The visibility to near-term growth for AI infrastructure is clear
Consensus is now expecting data center capex to reach $1.3 trillion in 2027, up from ~$259B in 2024.
The investment is underpinned by the continued conviction of tech leaders to invest in AI https://t.co/2YyZe6EFoI > 引用 @SmallCapSnipa: Bernstein on AI Infrastructure: We continue to be positive on AI in both the short, medium, and long-term, and see upside to AI infrastructure exposed names
$NBIS $IREN $CRWV $CIFR $WULF $RIOT
Short-term: The AI infrastructure trade is about the willingness of AI data center operators to keep spending
Medium-term: AI infrastructure on whether AI demand is sufficient to drive a healthy ROI for data center ROI. We believe that the assumptions needed to underwrite this are not aggressive.
Long-term: it’s hard to have deep conviction, but we see model scaling as the key leading indicator, and the scaling laws look intact. Accordingly, we skew positive. https://x.com/SmallCapSnipa/status/2105740621768269984