Nvidia is the central bank of AI

Nvidia as “central bank of AI”

• Many find the metaphor apt: Nvidia sits at the center of AI credit and hardware flows, backstopping neoclouds and AI labs and enabling otherwise-impossible capex.
• Others argue the analogy is loose: unlike a real central bank, Nvidia can’t expand supply at will, doesn’t set “rates,” and is profit‑maximizing, not policy‑oriented.
• Comparisons are made to past vendor financing bubbles in tech; some fear this will rhyme with the dot‑com era.

Role of TSMC, ASML, and supply constraints

• TSMC is likened to a mint or platform; ASML and upstream component suppliers are cast as the deeper “foundation.”
• Capacity at TSMC (and even more so at ASML) is seen as the true hard bottleneck; Nvidia is powerful partly because it can pay most for scarce leading‑edge wafers.

Financing and “money creation” debate

• Nvidia’s ~$500B of investments/commitments are contrasted with central‑bank balance sheets.
• Large sub-thread on whether such vendor financing “creates money”:
  • One side says credit and guarantees amplify lending and thus effectively increase broad money.
  • Another insists only regulated banks/central banks expand monetary aggregates; Nvidia just reallocates existing capital or causes asset swaps on balance sheets.
• Some note these loans may be productive (deflationary via new output), unlike some central‑bank QE.

AI boom, bubbles, and systemic risk

• Multiple commenters see classic top signals and unsustainable capex; others say model capability and usage are still clearly accelerating.
• Concern about circularity: Nvidia finances customers who buy Nvidia GPUs, while those customers burn huge amounts of investor cash.
• If a major AI lab fails, views diverge:
  • Optimists think compute will be easily repurposed.
  • Skeptics argue demand is correlated, resale is slow, and pricing would collapse.

Gaming GPUs and end‑user hardware

• Debate over whether Nvidia will (or already has) deprioritized gaming in favor of data‑center AI:
  • Some expect eventual exit or major scaling back, given far higher margins in AI.
  • Others think abandoning gaming would be strategically foolish and leave a huge market to AMD/Intel or new entrants.
• Rising prices for GPUs, RAM, and SSDs are seen as making PC gaming a niche; some foresee more cloud gaming and “GPU renting.”

AI trajectory, usefulness, and calls to slow down

• Conflicting reads of frontier labs’ public calls to “pace the frontier”:
  • Cynical interpretation: they see limits to profitability or usefulness and want a face‑saving slowdown, possibly with state backstops.
  • Alternative: they think usefulness and capability are growing so fast that regulation is needed both for safety and to restrain reckless competitors.
• Some believe LLMs are approaching diminishing returns for “AGI”; others point to rapid improvements in coding agents and small specialized models.

Corporate power and governance

• Several note that Nvidia and similar firms increasingly resemble public institutions in their macro impact, without democratic constraints or public‑interest mandates.
• Worry about drift toward corporatocracy and “feudal” dynamics; counter‑voices say this is exaggerated but agree corporate power design deserves more scrutiny.

Markets, instruments, and side notes

• Jokes about a LIBOR‑style benchmark for VRAM; mention that CME is launching GPU compute futures.
• Some see Nvidia’s dominance as heavily entwined with CUDA lock‑in.
• Multiple commenters expect an eventual AI‑driven market correction but disagree on timing and severity.