Cloud & Capital · Market Tape

Data Sources

Every number on Market Tape comes from one of three layers: live market data refreshed every 30 minutes, manually-maintained benchmarks updated on a scheduled cadence, or public announcements compiled by hand. This page documents all of them.

Live Market Data

Prices, momentum scores, sector rotation, and macro indicators are fetched from yFinance via a data pipeline that runs every 30 minutes and publishes to github.com/cloudandcapital/market-tape. The frontend revalidates against that source every 30 minutes.

Macro & Market Internals

SPY, ^VIX, TLT, HYG, GLD, USO, IWM, QQQ, DXY — prices, 1M relative strength, trend grades, breadth metrics

US Sector ETFs

XLK, XLF, XLV, XLE, XLI, XLY, XLP, XLU, XLB, XLRE, XLC — ranked by 1-month relative strength vs SPY

Valuation Multiples (live approximations)

Approximate NTM P/S computed as: TTM P/S ÷ (1 + trailing revenue growth). Equal-weighted median across each basket. Labeled with ~ to communicate approximation. Sourced from Yahoo Finance with the same 30-minute cache as the rest of the pipeline.

Public Cloud hyperscaler basket

AMZN · MSFT · GOOGL · ORCL

Blended corporate P/S — AWS, Azure, and GCP are segments within these companies, not isolated. Treat as a proxy.

SaaS basket (representative BVP Cloud Index names)

CRM · NOW · SNOW · DDOG · ZS · HUBS · WDAY · VEEV

Eight liquid names from the BVP Nasdaq Emerging Cloud Index. Not the full index; WCLD ETF is the closest tradeable equivalent.

AI infrastructure semiconductor basket

NVDA · AVGO · AMD · MU · MRVL

Semiconductor and AI hardware names. Dominated by NVDA; results are sensitive to NVDA revenue forecasts.


AI Compute Commitments

The $1.5T+ commitment table is compiled manually from public announcements and press releases. Each row is sourced individually. The data is static until a new deal is announced — source code at lib/aiCompute.ts.

AnthropicAWS (Trainium / Inferentia)

$100B+ · 5 GW · 10 years · announced Apr 2026

Anthropic blog, Apr 2026

source ↗

AnthropicGoogle Cloud / TPU (Broadcom)

$200B · 3.5 GW TPU · 5 years (starts 2027) · announced Apr 2026

Financing closed June 2026: Apollo Global Management + Blackstone closed private credit package via SPV that purchases Google TPUs and leases back to Anthropic. Infrastructure online starting 2027. Anthropic run-rate revenue ~$30B as of Jun 2026.

source ↗

AnthropicxAI / SpaceX (Colossus 1)

~$40B ($1.25B/mo thru May 2029) · 300 MW / 220K+ NVIDIA GPUs · 3 years · announced May 20, 2026

TechCrunch + xAI blog, May 20 2026

source ↗

OpenAIStargate (Oracle, SoftBank, partners)

$500B total / $400B+ committed · 10 GW target / 7 GW planned · 2025–2029 · announced Jan 2025 (updated May 2026)

OpenAI blog, May 2026

source ↗

OpenAIMicrosoft Azure

ongoing (overflow + existing traffic) · · ongoing · announced 2024+

OpenAI / Microsoft partnership

source ↗

OpenAINVIDIA (equity + Vera Rubin compute)

$30B equity + compute commitment · 5 GW (3 inference + 2 training) · · announced Feb 2026

Yahoo Finance, Feb 2026

source ↗

MetaNVIDIA (Blackwell + Vera Rubin + Grace)

~$50B · · multi-year · announced Feb 2026

Tom's Hardware, Feb 2026

source ↗

MetaAMD (Instinct MI450+)

~$60B · 6 GW · 5 years · announced Feb 24, 2026

TECHi / Tom's Hardware, Feb 24 2026

source ↗

MetaAmazon (Graviton5 chips)

multi-billion · · · announced 2026

The Next Web, 2026

source ↗

Jane StreetCoreWeave

$6B · · · announced Jun 2026

Financial trading firm as major AI cloud buyer — signals enterprise/financial sector emerging as serious compute demand category alongside AI labs. CoreWeave IR, June 2026.

source ↗

GoogleSpaceX / xAI

$920M/month (~$30B total) · · Oct 2026 – Jun 2029 (32 months) · announced Jun 5, 2026

~110,000 NVIDIA GPUs hosted at xAI data centers. Bridge capacity for unexpected Gemini Enterprise demand surge while Google builds its own infrastructure. Terminable with 90 days notice after Dec 31, 2026. Notable inversion: a hyperscaler (Google) as the compute buyer, not the supplier — complicates the standard "AI lab rents from hyperscaler" framing.

source ↗

Industry Benchmarks

Slow-moving external statistics maintained manually in lib/industryBenchmarks.ts. Each entry has a review cadence. Run npm run check-benchmarks to see current freshness status.

gpuSupplyStatusH100 broadly available ($0.99/hr spot · $12–18K secondhand) · H200 4–8 weeks / $31–40K · Blackwell B200/GB200 heavily constrained (~3.6M unit backlog, HBM3e bottleneck)
Thunder Compute / Cyfuture GPU pricing guides (June 2026)

Updated Jun 9, 2026 · Next review Sep 1, 2026 · quarterly

As of June 2026: H100 (Hopper) has flipped to broadly available — spot rates as low as $0.99/hr on some clouds, secondhand hardware $12–18K (down from $40K peak). H200 standard lead time 4–8 weeks, $31–40K/unit hardware; could extend to 6–9 months if Chinese demand materializes, cloud provider allocations pre-secured. B200/GB200 (Blackwell) still heavily constrained — ~3.6M unit backlog as of April 2026, HBM3e supply is the structural bottleneck. Vera Rubin (GB300) first deployments targeting Q1 2027. Refresh after each NVIDIA earnings call or major SemiAnalysis update.

dataCenterConstructionYoYDemand +38% / construction -5.6% YoY
CBRE North America Data Center Trends H2 2025

Updated May 1, 2026 · Next review Nov 1, 2026 · semi-annual

CBRE H2 2025 reported absorption of 2,497.6 MW vs 1,809.5 MW in 2024 (+38%) and construction pipeline of 5,994.4 MW vs 6,350.1 MW in 2024 (-5.6%, first decline since 2020). Story = demand outpacing supply, not construction boom.

saas2021PeakMultiple20x+
Bessemer Venture Partners Cloud Index — historical

Updated Apr 24, 2026 · Next review Nov 1, 2026 · semi-annual

Peak NTM revenue multiple reached by the SaaS cohort during the 2021 bull market. Used as a compressed-from reference point. This is a historical data point and will not change. Review cadence is semi-annual only to confirm the narrative framing is still accurate.

hyperscalerCapexTrendExpanding
AWS / Azure / GCP / Meta Q1 2026 earnings results + Jun 2026 guidance updates (Tom's Hardware)

Updated Jun 16, 2026 · Next review Aug 1, 2026 · quarterly

Q1 2026 actuals: Amazon CapEx $44.2B; Google + Microsoft + Amazon combined Q1 $112B. Full-year 2026 guidance: Big 4 (Amazon, Google, Microsoft, Meta) tracking ~$725B (+77% vs 2025 record of ~$410B). Per-company full-year: Amazon ~$200B; Google up to $190B (raised ~$5B from prior guidance, Jun 2026); Microsoft ~$190B (includes DRAM/flash cost increases from tariffs); Meta $115–135B. All four flagged AI demand exceeding current supply — Azure and GCP both noted compute constraints. Values: Expanding / Stable / Contracting. Next update due after Q2 2026 earnings (typically July–August).

nvidiaDcRevenue$75.2B Q1 FY2027 (+92% YoY, +21% QoQ) · networking $14.8B (+199% YoY) · Q2 guidance ~$91B
NVIDIA Q1 FY2027 earnings (reported May 2026)

Updated May 25, 2026 · Next review Aug 1, 2026 · quarterly

NVIDIA Q1 FY2027 (quarter ended April 2026, reported May 20–22 2026). Total revenue $81.6B (+85% YoY). Data center segment $75.2B (+92% YoY, +21% QoQ) — the primary demand signal for hyperscaler AI CapEx flowing through to silicon. Data center networking (InfiniBand + Ethernet for AI) $14.8B (+199% YoY). Q2 FY2027 guidance ~$91B total. Refresh after each NVIDIA earnings call (next: ~Aug 2026).