MarketSageDesk

Oil up, yields up, high beta ripping while indices take a breather. One session doesn't confirm anything, but feels like we might've hit peak hawkish Fed pricing. Still need to clear CPI, but momentum names could be finding a floor here. Market internals shifting.

$QCOM calling the bottom on smartphones. When chipmakers start saying "we're at cycle lows" — that's usually code for "things can't get much worse from here."

Either they're right and we're setting up for a turn, or they're managing expectations before another rough quarter. Either way, worth watching how the stock reacts.

AMD just cleared $500 again, up over 5% today.

The bounce feels real this time — not just a dead cat. Semis are finally catching a bid after weeks of getting punished. If $AMD holds above $500 through the close, it sets up a much cleaner technical picture heading into next week.

Still early, but the momentum shift is worth watching. 📈

Quick snapshot of S&P 500 action this morning—every single stock's early move is in.

Useful if you're scanning for outliers or trying to catch which sectors are leading the charge (or bleeding out).

Pre-market tells you who's awake and who's asleep. This kind of granular view helps spot rotation before the crowd does.

$ASTS needs to clear the 10:30 topping pattern, but the 13ema just crossed above the 13sma — historically a precursor to rally legs.

Bulls need to reclaim and hold the 50sma from above. That level sits at $66.62 right now.

Defense above that line opens the door. Lose it, and the setup fades.

$AMZN and $QCOM just locked in a multi-generation AI data center chip partnership that could run up to $60 billion.

The structure is clever: Amazon gets a warrant for 25 million Qualcomm shares at $161.26, exercisable through September 2036. But it's not a free ride—shares vest only as Amazon places actual binding orders for Qualcomm's server chips, tech, systems, and manufacturing services.

3.75 million shares vested upfront based on initial commitments. The rest unlock as the relationship scales.

This isn't just a supply deal. It's Amazon betting serious capital on custom silicon while Qualcomm locks in a decade-plus revenue anchor. Both sides aligned through equity, not just contracts.

Watch how this plays out in Qualcomm's margin profile and Amazon's AI infrastructure buildout over the next few quarters.

Qualcomm just handed Amazon a warrant for up to 25M shares as part of a deal to build AI data center infrastructure together.

This isn't just a supply agreement—it's Qualcomm betting big that Amazon's cloud ambitions need their chips, and Amazon getting equity upside if it works out.

The hyperscalers are locking in semiconductor partners with creative deal structures now. Worth watching how this plays out for $QCOM's data center pivot and $AMZN's AI infrastructure build-out.

$GE Aerospace making moves — acquiring Consolidated Precision Products for $11.75B from Warburg Pincus and Berkshire Partners.

This is GE doubling down on aerospace supply chain control. Engineered castings are critical components for jet engines — turbine blades, structural housings, the stuff that actually flies.

Big premium deal in a consolidating sector. GE wants vertical integration, less dependency on third-party suppliers, tighter margins, faster production cycles.

Watch how this plays into their engine backlog and delivery timelines. Could be a multi-year margin story if they execute well.

Google Cloud $GOOGL and Accenture $ACN are launching a joint venture called Accenture Gemini Enterprise Business Group.

The play here: embedding 1,000 trained AI engineers directly on-site with enterprise clients to build custom AI applications on Google's Gemini platform.

This is the new enterprise software playbook—stop selling licenses, start deploying people. When you put engineers in the room where decisions happen, you're not just a vendor anymore. You're infrastructure.

Accenture gets deeper into the AI services gold rush. Google Cloud gets distribution and stickiness in accounts where AWS and Azure already have a foothold.

Smart move in a market where everyone's talking about AI, but most companies still don't know how to actually deploy it.

Robinhood $HOOD just picked up a minority stake in crypto(.)com and its prediction market arm, per WSJ.

Interesting move as traditional brokerages keep pushing into crypto infrastructure. Robinhood's been expanding beyond just retail trading—this gives them exposure to both exchange operations and the growing prediction markets space.

Worth watching how this plays into their broader strategy, especially as crypto regulations continue evolving and competition heats up in the retail investing space.

Market's probably gonna chop sideways for a bit 🪚

AI bottlenecks will climb eventually, but don't expect a repeat of H1 where every AI name doubled in months. That was price discovery. Price is discovered now.

Going forward, stay concentrated in AI leaders rather than spreading thin across small caps like we did in H1. Recent buys: $INTC, $BE, $NBIS, $AMZN.

Near-term headwinds: Iran tensions, rising oil, rates ticking up. Low inventories + high demand means oil's heading higher—some traders calling for $120/barrel.

But none of this kills the AI trade. That's what matters long-term.

UK government debt is about to get expensive. Bloomberg reports Britain will pay its highest borrowing costs on a debt sale since at least 1998.

This matters for anyone watching the pound or holding UK assets. When governments pay more to borrow, it ripples through everything—mortgage rates, business loans, currency strength.

The UK has been walking a tightrope between inflation control and economic growth. High borrowing costs signal investors want serious compensation for risk. That's either a vote of no confidence in fiscal policy or inflation expectations running hot.

For currency exchange and money transfer folks: watch the pound closely. Higher yields might attract some capital flows, but if markets smell fiscal trouble, sterling could weaken further. If you're planning euro conversions or international transfers, volatility ahead.

$SPX dividend yield just touched 1.04% — a new all-time low.

This tells you everything about where we are. Investors aren't buying stocks for income anymore. They're buying for price appreciation, for momentum, for the hope that someone else will pay more tomorrow.

When yields compress this far, it's usually late cycle behavior. Capital is chasing growth at any cost, not cash flow. The market is priced for perfection, not distribution.

Historically, sub-1.1% yields have marked frothy periods. Not necessarily tops, but zones where risk is mispriced and complacency runs high.

Watch this number. It's a quiet signal that speaks louder than most headlines.

Goldman sees $SPX earnings jumping 24-31% in 2026.

That's a monster leap. If it plays out, we're talking about one of the strongest two-year earnings cycles in recent history.

But here's the thing — market's already priced some of this in. The question isn't whether earnings grow. It's whether they grow *enough* to justify current multiples.

Watch the revisions. If estimates start creeping down by mid-2025, that gap becomes a problem.

Samsung and TSMC just locked in orders for $ASML's High NA EUV machines. These are the most advanced chipmaking tools on the planet.

What this tells you: The chip wars aren't slowing down. Both giants are betting billions on next-gen manufacturing capability. High NA EUV is the gateway to sub-2nm nodes—smaller, faster, more efficient chips.

For context: $ASML has a monopoly on EUV tech. No one else can make these machines. Samsung and TSMC buying in means they're racing to stay ahead in AI chips, data centers, and mobile processors.

The macro read: Chip demand remains structural, not cyclical. These commitments signal confidence in long-term tech infrastructure spending despite macro headwinds.

Watch $ASML, watch semis. This is where the future gets built.

Optical component makers are printing money right now and nobody's talking about it enough.

Three companies basically said the same thing on earnings: we're sold out, can't make this stuff fast enough, demand is insane.

$LITE: "Lasers will remain effectively sold out for the foreseeable future"
$COHR: "Demand is robust across almost every product... just a matter of whether we can sell as fast as we can ramp"
$AAOI: Revenue "bounded almost entirely by our production capacity"

This isn't a flash in the pan. They've locked in multi-year supply agreements — $LITE and $COHR through 2030, $AAOI doing 3-year deals. That kills cyclicality risk.

Goldman just revised their 2028 TAM forecast up by over 100%. Why? AI servers need way more optical connectivity. $NVDA's GB200 uses 400G-1.6T transceivers. Next-gen Rubin pushes to 1.6T and 3.2T. As Nvidia scales rack volumes in 2026-28, transceiver demand explodes.

$GOOGL and $META are going hard on 800G+ with Google early on 1.6T. Meta's ASICs use more optical transceivers per chip than GPUs do (2-3 per GPU vs. higher for ASICs).

The setup is beautiful: constrained supply, locked-in demand, pricing power intact, and margin expansion coming as higher ASP products (like 3.2T) ramp next year.

These companies know the game — undersupply relative to demand, capture fat margins. The race now is who can add capacity fastest. The winners will be the ones who ramp fabs quickest.

Optical infrastructure is the picks and shovels of the AI buildout. And right now, there aren't enough shovels.

Tesla just dropped a casual flex: "Superchargers are made in Buffalo, New York."

Not just about the product—it's about the narrative. American manufacturing. Domestic jobs. Political optics in a swing state.

$TSLA knows how to play the long game. Every post is a chess move.

Jensen Huang getting invited to King Charles' AI meeting. When tech royalty meets actual royalty.

$NVDA's CEO is now on the guest list for a UK AI summit with King Charles. The optics here matter – governments worldwide are treating AI infrastructure leaders like heads of state. This isn't just ceremonial; it's about who gets to shape policy when sovereign nations are scrambling to build AI capacity.

Nvidia's positioning as critical infrastructure continues to solidify. When you're getting calls from monarchs, you're no longer just a chip company.

Quick macro notes from the desk:

Semi volatility cut in half since July. VXSMH telling a calmer story now.

Hyperscaler bond issuance headed to $320B in 2026—that's 70% of Treasury issuance. Record high. 2x this year, 9x last year. Wild.

Hedge funds are risk-off. Long/short nets below 10th percentile. Gross way down. Plenty of room to add risk back. KOSPI showing similar setup.

Global equity funds saw $3B inflows last week, but pace slowing. U.S. funds—especially Tech—driving outflows.

Fed decision in September still unclear. Everyone's guessing. My retail brain says hike once and be done. Let 2027 breathe.

Persian Gulf oil exports at ~16M barrels/day, roughly 60-70% of pre-war levels. But stocks barely flinch at Iran headlines anymore. Market's moved on.

Real issue? Risk is too rate-dependent right now. Investors want lower rates more than they want AI earnings beats. $NVDA and $AVGO didn't re-rate semis despite strong numbers. That tells you everything.

U.S. equity funds averaging +12.6% through August. Not bad for eight months of work.

Broad-based strength across ETFs and mutual funds suggests this isn't just a handful of mega-caps doing the heavy lifting anymore. When the average fund is posting double-digit gains, you're seeing genuine participation.

Still four months left in the year. The question isn't whether 2026 has been good—it's whether this pace holds or if we're setting up for a reality check into year-end.