Cerebras missed revenue at $180.1M but grew cloud 281% YoY and carries $25.4B in remaining performance obligations. The backlog is the story: inference demand is booking faster than the vendor can recognize it, with revenue set to triple in 2027. $CBRS

Great to be featured in Business Insider discussing why I bought into the fear around $SPCX insider lockup expiration. The market spent weeks bracing for a flood of supply but the first major unlock was absorbed better than feared and the stock is now up 40% in just five days.

One of my favorite developments from $NBIS earnings was its first capacity auction because it gave us real price discovery instead of relying on management commentary around scarcity. The auction cleared ~15% above Nebius’ previous peak Blackwell pricing and ~20% above its existing pipeline as multiple buyers competed for limited GPU capacity. That effectively created a small spot market for AI compute and confirmed that immediate capacity is worth materially more than Nebius existing contracts suggest.

EXCLUSIVE INTERVIEW INCOMING We will be doing a live stream with @infleqtion tomorrow Thursday, August 13th at 2PM ET following their earnings release today We will be having Infleqtion CEO Matt Kinsella and CTO/ Co-Founder Pranav Gokhale on the stream with @StockSavvyShay You will be able to watch the stream right here on the Wolf Financial X page SAVE THE TIME IN YOUR CALENDAR

$NBIS is up nearly 30% because this quarter showed the AI compute shortage is translating into better pricing, faster paybacks, higher margins and better financing. Nebius now earns ~$25M per MW on normal contracts and up to $50M for immediate capacity while Q2 deals pay back in under two years and customers fund roughly half the capex upfront. That pricing power is already showing up in the market with its first capacity auction clearing 15% above prior $NVDA Blackwell pricing, AI cloud margins reaching 50% and management saying it could sell all of 2027 capacity today.

Futurum Equities raises its cumulative data center capex forecast to $12.5T through 2030 as AI infrastructure spending expands far beyond $MSFT $GOOGL $AMZN and $META. By 2027, Tier 2 operators and third-party financing are projected to overtake hyperscalers as the largest source of spending pushing more of the buildout toward neoclouds like $NBIS and $CRWV plus private capital and lease structures.

Futurum Equites @rolfbulk raises 2030 cumulative Data Center Capex forecast Base case from $10.7 trillion to $12.5 Trillion. Bull case rises to $14.3 Trillion from $11.7. Massive demand for AI and higher input costs as key drivers. 💪🏻🚀

Nebius swung to $2.25B in operating cash flow this quarter, from a $168M outflow a year ago, with over $9B in customer prepayments expected in 2026. Prepaid multi-year contracts are how the compute buildout gets financed before the racks ship. $NBIS

Supermicro posted Q4 revenue of $11.1B, short of the $11.55B estimate, even as EPS and FY27 guidance ran well ahead. The top-line miss keeps the margin-versus-scale question open on a name where execution risk has long been the real overhang. $SMCI

The Gemini app has crossed 1B monthly users, Google's fastest-growing product ever and 14th to hit that threshold. Owning the full stack from TPU to consumer surface is what scales adoption this fast, and these numbers are the demand side of that build. $GOOGL

I came away from $ASTS earnings more bullish because the business is moving to actually building the network with revenue roughly doubling sequentially to $32M, backlog reaching ~$1.3B, 13 satellites now in orbit and production already extending through BlueBird 46. With more than $3.7B of liquidity, financing the first phase looks much less risky while the bigger execution question is how quickly ~$610M of quarterly capex can turn satellites, launches, manufacturing and gateways into a functioning commercial network. That spending is what eventually unlocks the model because AST is effectively building a cellular network before turning it on and if service launches on schedule then management’s path toward ~$1B of revenue in the first full year starts to look much more realistic..

ThE mArKeT iS iN sUcH a BuBbLe Yet these are the 2027 earnings multiples for some of the world’s biggest companies: • $AMZN ~23x • $GOOGL ~22x • $NOW ~22x • $MSFT ~20x • $AVGO ~20x • $DELL ~20x • $TSM ~19x • $NVDA ~16x • $META ~16x • $ORCL ~11x • $SNDK ~7x • $MU ~6x • $SKHY ~4x

$INTC has emerged as one of 2026’s biggest AI winners by benefiting across Xeon CPUs, purpose-built silicon and advanced packaging. That exposure is expanding because every accelerator rack still needs general-purpose compute to manage data, storage and workloads making agentic AI a new growth driver for server CPUs. The same buildout is pulling Intel deeper into custom silicon which nearly tripled YoY to a ~$2B annual run rate and advanced packaging where management already sees billions in backlog potential.

$GOOGL Ironwood TPU packs ~190GB of HBM3E so every new deployment directly increases premium memory demand across the cluster. That demand compounds as Gemini APIs process ~22B tokens per minute giving $SKHY the clearest near-term upside and $MU more upside if it gains share. Google has both the returns and the incentive to keep expanding memory capacity with Cloud margins at 36% and capex rising another $15B.

Anthropic launched Claude Opus 5, which it says approaches frontier intelligence at half the price of the prior tier. Every halving of inference cost pulls agentic workflows out of pilots and into always-on production, where compute demand compounds.

$QCOM has told customers that prices are going up by a double-digit percentage, Bloomberg reports. Qualcomm said it can no longer absorb the increasing costs it has been carrying.

Nvidia's Jensen Huang used his debut X post to argue AI will be built by every country and that open models enable sovereignty. Sovereign AI is the next demand leg: national build-outs need their own compute, and that lands directly in NVIDIA's data center order book. $NVDA

FUTURUM EQUITIES PODCAST EPISODE #50# Daniel and I will be live today at 1 PM ET exclusively on the Futurum Equities YouTube channel covering: • Why $GOOGL may be the most mispriced winner of the AI buildout and what its rising capex means for $NBIS, $CRWV, $CIFR, $DOCN, $AMZN & $MSFT • Whether $INTC breakout quarter marks a real turnaround and what it still needs to prove across 18A, 14A & external foundry demand • How $IREN $2.8B in new AI cloud contracts and customer prepayments are changing their story • Why $NOW rose while $TSLA fell after earnings & what that says about visible AI revenue versus future AI promises • What Kimi K3’s compute constraints signal for memory and networking across $MU, $SKHY, $AVGO, $CRDO, $ALAB & $MRVL We’ll leave the last 15 minutes for live Q&A so bring your questions!