RockstarMarkets
All news
Markets · Narrative··Updated 1h ago
Part of: Semiconductor Cycle

ARM and PLTR at 50-Plus Price-to-Sales Revive Late-1999 Valuation Comparisons

A $10B accounts payable discrepancy at SMCI that never appeared in NVDA's books is drawing short-seller and auditor attention, while compressed implied volatility on small-cap chip names signals one-sided positioning across ^IXIC.

R
Rocky · RockstarMarkets desk
Synthesised from 8 wires · 67 mentions in the last 24h
Sentiment
-20
Momentum
72
Mentions · 24h
67
Articles · 24h
82
Affected sectors
Related markets

Key facts

  • ARM, PLTR trade at 50+ price-to-sales; ASTS, RKLB, ALAB, CBRS similar multiples
  • SMCI $10B accounts payable discrepancy never appeared in Nvidia's own accounting
  • Samsung-union tentative deal averted strike; but manufacturing cost inflation persists
  • Options implied volatility compressed on small-cap chip names; positioning heavily long
  • Nvidia revenue $78.75B guidance, but capex sustainability questioned at higher rates

What's happening

A chorus of voices is warning that chip and semiconductor valuations have entered bubble territory, even as Nvidia reports perfect earnings. The data is stark: ARM and Palantir both trade above 50 times sales, ASTS, RKLB, ALAB, and CBRS trade at similar valuations, and smaller-cap beneficiaries of the AI boom are commanding price-to-book ratios that have no historical precedent. When asked why companies with 50 billion or higher market capitalizations trade at these multiples, the honest answer is momentum, not fundamentals. This concentration of valuation premium in a narrow band of names mirrors late-1999 dynamics.

The concerns deepened when renewed scrutiny fell on Super Micro Computer (SMCI), the server manufacturer at the heart of the AI infrastructure buildout. Earlier, a striking discrepancy emerged: SMCI had reported roughly $10 billion in accounts payable that mysteriously vanished from its books in a subsequent quarter. Nvidia, as SMCI's largest supplier, reported its own earnings and provided detailed accounting, and there was no sign of the $10 billion SMCI owed for chip purchases. This mismatch raises a red flag: either SMCI's accounting is opaque or incomplete, or there was a significant reversal unaccounted for. Auditors and short-sellers are now circling.

Investor sentiment has tilted decidedly long on small-cap chip plays, and options implied volatility has compressed. When positioning is this one-sided, any catalyst, regulatory scrutiny, accounting restatements, earnings misses, can trigger violent mean reversion. The broader semiconductor supply chain also faces margin compression if Nvidia's guidance proves conservative and capex cycles shorten. Samsung's tentative labor deal defused one strike risk, but manufacturing cost inflation remains. TSMC, the foundry backbone, also faces geopolitical risks around Taiwan and export controls on advanced nodes to China.

Defenders of current valuations argue that AI infrastructure is a multi-decade secular trend and that winners will command premium multiples for years. They point to cloud infrastructure spending curves from 2010-2020 as analogues. However, cloud beneficiaries like Amazon and Microsoft took years to reach current scale, and competition was far less intense. Today, every major tech firm is trying to build proprietary AI chips, and customer concentration risk is high. If hyperscaler capex disappoints relative to current expectations, a non-trivial risk given higher interest rates, valuation compression could be swift and brutal.

What to watch next

  • 01SMCI accounting audit results and potential restatement; SEC investigation status
  • 02Small-cap semiconductor earnings (SMCI, AVGO, ARM) through Q2; margin trends
  • 03Amazon, Google, Microsoft capex guidance; any slowdown triggers semiconductor repricing
Mention velocity · last 24 hours
Coverage from these sources
Previously on this story

Related coverage

More about $ARM

Topic hub
Semiconductor Cycle: AI Capex, Memory and the SOX Trade

Live coverage of the AI semiconductor cycle — NVDA, AVGO, AMD, ASML, memory demand, capex run rates and overbought signals.