Intel's $15B Equity Offering Puts Financing, Not Demand, at the Center of the AI Trade. Jefferies Cuts Apple to Underperform.

Monday, August 10, 2026, written at 9:20am Eastern, ten minutes before the opening bell, so the pre-market marks below are live and several of them will move within the hour. Futures are flat rather than directional, which is the correct posture for a session sitting two days ahead of the only data point that matters this week. The interesting news this morning is not in the index level. It is that the largest single corporate announcement of the day is a company asking shareholders for fifteen billion dollars to fund the AI buildout, on the same morning a Bloomberg piece asks whether credit markets have started to doubt that same buildout, and one trading day after Berkshire Hathaway disclosed that it has finally stopped hoarding cash. Those three items are one story told from three balance sheets.

Top Call: the AI Trade Just Became a Financing Trade

Intel announced a $15 billion underwritten public offering of common stock this morning and the shares fell more than 3 percent in pre-market to roughly $98.36, breaking the $100 handle. The company is raising equity, not debt, explicitly to protect its investment-grade rating and improve its debt-to-equity ratio while funding physical AI, purpose-built silicon, advanced packaging and external foundry wafers. Read that carefully. Intel grew revenue 25 percent last quarter with data centre revenue up 59 percent, and its answer to that demand is to dilute rather than to borrow or self-fund. For eighteen months this market has priced AI on the demand line. This morning it has to price it on the funding line, and the two do not always point the same way.

Where Things Stand at the Bell

The shape matters more than the levels. Small caps lag again, the VIX is up 3.6 percent off a multi-week low, and gold gives back a fraction of a 7 percent week while crude adds 1.4 percent. That is a market taking a small amount of risk off ahead of an inflation print, not one repricing anything.

Intel: Dilution as a Strategy

The $15 billion raise is the most consequential corporate action of the day and it deserves more than the pre-market reaction it is getting. Three points.

First, the timing is deliberate. Intel is issuing into a share price that has roughly doubled off its lows and into a narrative where its 18A node is in high-volume production with named customers. Management is monetising its own re-rating. That is rational treasury behaviour, and it is also an implicit statement about where management thinks the stock sits relative to fair value.

Second, the choice of instrument is the tell. A company confident about the timing of free cash flow borrows. A company that needs multi-year capital before free cash flow turns positive, and that cannot afford a ratings downgrade while it waits, issues equity. Intel said as much when it cited its debt-to-equity ratio and its investment-grade rating in the same breath as the use of proceeds. The market read the dilution and sold it. The more durable read is that foundry capital intensity is still running ahead of foundry cash generation, and the gap is fifteen billion dollars wide.

Third, this is not an Intel-specific problem. CoreWeave will spend $31 billion to $35 billion this year against a $99.4 billion backlog and reports tomorrow with a consensus loss of $1.22 a share, wider than the $0.27 loss a year ago. Super Micro has pre-announced revenue at the low end of its range with better margins and more than $60 billion of new orders. Every one of these firms converts balance sheet into capacity faster than capacity converts into cash. Bloomberg published a piece this morning arguing that credit markets have begun to price doubt into exactly this dynamic. When the equity market and the credit market start asking the same question about the same buildout in the same week, the question tends to get answered on the tape rather than in a research note.

Apple Loses Its Pricing Story

Jefferies cut Apple to Underperform from Hold this morning, taking the target to $263.66 from $285.56, roughly 16 percent below the recent price. The stock is down about 1 percent. Analyst Edison Lee’s supply-chain checks indicate Apple has cancelled the twentieth-anniversary all-glass iPhone slated for September 2027 because production yields were unacceptable. That leaves the foldable, due next month, as the only near-term driver of higher average selling prices and margin, and Lee flags that soaring memory costs could push its starting price above $2,000.

This connects directly to the storage and DRAM story that has run through these notes for a fortnight. Samsung, SK Hynix and Micron have sold out all of 2027 DRAM and HBM with customers filled at only 60 to 70 percent of requests. Memory scarcity is not an abstract supplier win. It is now visibly a cost line inside the largest hardware franchise in the world, and it has arrived at the exact moment Apple lost the product that was supposed to carry the price increase. That is a fundamental downgrade rather than a valuation call, and it is the second sell rating to land on the name in recent months.

Asia Overnight: Japan Rips, Semis Lag the Index

Japan led the region, adding roughly 1,360 points off Friday’s payroll miss. The yen was the release valve, falling 0.6 percent to 158.89 per dollar, its steepest daily decline in five months, which is the mechanical exporter trade rather than anything domestic. Note the internal divergence: the index ripped while SoftBank fell 1.22 percent, SK Hynix slipped 0.14 percent and Samsung lost 0.43 percent. Asia bought the discount rate and left the semis behind, which is the same split the US tape has run for two weeks.

Korea is the more interesting story underneath. SK Hynix is preparing a shareholder return programme worth roughly 100 trillion won, about $72 billion, with a buyback component potentially reaching 40 trillion won, to be finalised and announced in the third quarter and funded by a commitment of half of free cash flow. That is an extraordinary number for a company that fell 10 percent on a sidecar halt eight sessions ago after a record quarter was judged a miss. Korean chipmakers are responding to sustained investor pressure by converting the memory cycle into cash returns rather than into further capacity, and if that holds it changes how the sector should be valued. The Kosdaq triggered a sidecar again during morning volatility, so the underlying tape remains jumpy.

China was mixed and unconvincing. Hong Kong finished about 0.9 percent higher and the Shanghai Composite added 0.7 percent to 3,966.59, but the blue-chip CSI 300 was lower as tech weakness offset consumer gains. Breadth in China is coming from the wrong places.

Europe: Records Held, Software Leads

The Stoxx 600 traded around 661, up roughly 0.4 percent and back in record territory after Friday’s 660.25 close, and the Euro Stoxx 50 added 0.4 percent to about 6,530. The DAX is the exception, hovering near 26,330 and down about 0.14 percent, just under Friday’s record close, though it remains up more than 12 percent year to date and one of the strongest major benchmarks in the world this year.

Leadership was software and AI infrastructure for a second consecutive session: SAP gained more than 4 percent, with Infineon and Siemens each up roughly 3 percent. That is worth flagging, because European tech leadership has historically been a lagging echo of the US move. Two sessions in a row of independent strength, with the Stoxx 600 at records while the DAX consolidates, suggests the continent is trading its own AI capex chain rather than importing sentiment. The overhang is the one bothering everyone: Hormuz, and what it does to the ECB’s inflation path.

Oil and Geopolitics: the Strait Is Not Reopening

Brent added more than 1 percent to about $84.70 and WTI rose 1.4 percent to $79.31, and this is the most important non-US development of the morning. Iranian Foreign Minister Abbas Araghchi said Tehran is not in direct talks with Washington to end the war and reopen the Strait, contradicting the US position last week that a deal was near. Iran’s stated conditions are sanctions relief and war reparations, and Tehran insists on controlling shipping through the waterway, which is contrary to international maritime law and is therefore not a position anyone can accept as written.

The physical numbers are the ones to hold onto. Before the conflict roughly 130 vessels transited the Strait daily, carrying about a fifth of global oil supply. In early August the count is 8 to 15 vessels a day. The International Maritime Organization counts at least 64 violent incidents and 17 deaths since the war began, most attributed to Iran, and ADNOC reported attacks on three vessels transiting Hormuz over the weekend, which the UAE condemned. Brent is up roughly 16 percent since the war started in late February.

The separate Iran and Oman track is reportedly close to agreement, but the mechanism involves maritime service fees paid through a newly established corporation, which is a toll booth on a waterway Iran does not own. Markets have now twice rallied on proximity to a deal and twice given it back. The reasonable base case is that the risk premium is structural until vessel counts recover, and vessel counts, not communiqués, are the indicator to watch. On the sanctions side, the Senate passed the Graham Russia and Iran Sanctions Act 86 to 11 on Friday and sent it to a House on recess, so escalation is legislated but not yet operative. Separately, Ukraine says it has zero Patriot interceptors remaining and is seeking resupply before winter, which is a live risk to European energy infrastructure in the fourth quarter.

The Weekend Print: Berkshire Finally Spends

Berkshire Hathaway reported Saturday and the stock is up about 0.4 percent pre-market. This matters beyond the ticker, because Berkshire’s cash pile has been the most-cited bearish valuation signal on Wall Street for three years.

Our take: this is Greg Abel’s signature quarter, not Warren Buffett’s. Berkshire bought roughly $20 billion of equities after fourteen or fifteen consecutive quarters of net selling, spent $4.5 billion on its own shares against $235 million the prior quarter, and put $10 billion into Alphabet, which now sits in the top five holdings alongside American Express, Apple, Bank of America and Coca-Cola. The Alphabet stake is explicitly an AI-supporting position, a notable philosophical departure for this balance sheet. The bearish reading is that $365.5 billion of cash still represents enormous caution about valuations. The more useful reading is directional: the most valuation-sensitive large allocator in the market started buying at record index levels, and it bought AI infrastructure exposure through the cheapest large-cap route available.

Earnings Today: Rocket Lab Is the One That Matters

Rocket Lab reports after the close and the stock is up about 2 percent pre-market into a roughly 41 percent rally. This is the launch-complex print of the week.

Our take: the financials are close to irrelevant tonight. Rocket Lab cleared $200 million for the first time in Q1 with 63.5 percent growth, and a print near $232 million simply confirms that the Electron and space-systems engine works. The stock trades on one sentence: does management still guide the first Neutron launch to late 2026? The Space Force awarded $663 million of contracts that give an unflown rocket a confirmed 2028 defence mission, which means schedule slippage is now a contractual problem rather than a public-relations one. A reaffirmed late-2026 window supports the rally. Any softening of that language, even phrased as prudence, is what takes the 41 percent back out. AST SpaceMobile also reports today and is up about 1 percent, with satellite deployment cadence and cash burn the focus.

Elsewhere pre-market: Hewlett Packard Enterprise is up more than 5 percent on a Morgan Stanley upgrade to overweight. Verisk Analytics is down more than 6.5 percent after Delaware Chancery Judge Bonnie David ruled it cannot walk away from its $2.35 billion acquisition of AccuLynx, finding that Verisk’s own willful conduct caused the closing condition to fail, though completion still depends on the FTC’s ongoing review. Doximity is up 32.6 percent to $27.40 after management detailed the economics of its AI search product, with the chief executive saying it earns ten times what it costs to run. IonQ is up 11.9 percent to $44.43 on Q2 revenue of $80.1 million, 20.4 percent above forecast. SpaceX is up 3 percent to $137.12, holding above its $135 IPO price and working on its first up week in five. Super Micro is up about 3 percent into tomorrow’s report. Barrick is down roughly 5 percent on Q2 EPS of $0.82 against $0.88 expected despite a $1.95 billion Nevada settlement, a reminder that gold at $4,388 does not automatically produce gold earnings. Tenax Therapeutics is down 85 percent to $2.08 after a Phase 3 trial missed its primary endpoints, taking market capitalisation from $503 million to $77.8 million.

Sector Spotlights

Semis and AI infrastructure. The sector is splitting into fundable and unfundable. Intel is diluting, SK Hynix is planning the largest shareholder return in Korean corporate history, Super Micro is carrying more than $60 billion of orders it has to finance through working capital, and CoreWeave is spending up to $35 billion this year. Global semiconductor sales hit a record $120.6 billion in May, the fifteenth consecutive monthly record. Demand is not the variable. Cost of capital is. Nvidia ended last week up more than 10 percent as AI chip fears eased, and it is worth being honest that the fear eased on price action rather than on new information.

Energy. Crude is bid on the failure of diplomacy rather than on demand. Energy equities rallied and travel names, cruise lines in particular, sold off on fuel costs. The structural point is that a persistent Hormuz premium is a tax on the consumer that arrives in CPI with a lag, which is precisely why Wednesday’s print is being watched more closely than a normal July release.

Financials and credit. The Bloomberg piece flagging doubt in credit markets is the one to read today. The argument is straightforward: transformational technology has to be financed with borrowed money long before it generates a return, and 2026 private credit is already the most stressed environment since 2008, with late-cycle behaviour in corporate credit and looser standards in private debt. Intel’s decision to issue equity rather than debt is a data point in favour of that thesis, not against it.

Healthcare and biotech. A constructive tape. The FDA granted accelerated approval to Replimune’s Tudriqev in advanced melanoma and approved Takeda’s Orzeyful for narcolepsy type 1. Tarsus Pharmaceuticals agreed to acquire Alkeus Pharmaceuticals in retinal disease. Attovia Therapeutics raised $289 million in an upsized IPO and rose about 29 percent on debut, and Blossomhill priced a $150 million upsized deal. The biotech issuance window is open, which is usually a mid-cycle risk-appetite signal rather than a late-cycle one.

Labour and restructuring. Equinix is cutting 3 percent of staff, roughly 400 people, alongside Adaire Fox-Martin replacing Charles Meyers as chief executive and the departure of the CIO and CISO. That follows AMD’s roughly 1,000 cuts and Cisco’s 6,000. There have been 322 layoff events in 2026 affecting 205,832 workers, about 927 a day. Friday’s negative payroll print was not an accident of seasonal adjustment.

Key Events Today and This Week

• Today, after the close: Rocket Lab Q2. AST SpaceMobile Q2. Roughly 226 companies report in total, with nearly 90 percent of the S&P 500 season complete and 86 percent beating on EPS.
• Tuesday: CoreWeave Q2 after the close, consensus revenue about $2.55 billion against a $99.4 billion backlog. Super Micro at 5pm Eastern. Lumentum, the optical read on AI networking.
• Wednesday, 8:30am: July CPI. Consensus core at plus 0.2 percent month over month, headline around 3.4 to 3.6 percent year over year against June’s 3.5 percent. Nebius reports before the open.
• Thursday, 8:30am: July PPI. Applied Materials after the close, the cleanest read on wafer-fab equipment spending and HBM tooling.
• September 15 and 16: FOMC. Hike odds sit near 36 to 44 percent depending on the venue, down from roughly 55 percent before Friday’s payroll report, with the committee last split 9 to 3.

What We Are Actually Watching

Three threads run through everything above and they are not independent.

The first is that the AI trade has moved from an income statement question to a balance sheet question. Intel’s raise, CoreWeave’s capex, Super Micro’s order book and the credit market’s growing scepticism are one phenomenon. As long as capital is available the buildout continues and the multiple holds. The moment financing costs rise, the same capex that reads as growth reads as leverage. That transition, if it comes, will not be signalled by a demand miss. It will be signalled by a funding announcement that gets sold harder than this one.

The second is that policy is genuinely undecided in a way it has not been for months. Friday’s payroll print killed the certainty of a September hike but did not replace it with anything. Roughly 40 percent odds is the market saying it does not know. Wednesday’s CPI is therefore not a confirmation event, it is a decision event, and the asymmetry is unpleasant: a soft core print mostly confirms what equities already rallied on last week, while a firm one reopens a hike into a labour market that just shed 23,000 jobs.

The third is that oil is no longer a headline risk, it is a supply fact. Eight to fifteen vessels a day against a pre-war 130 is not a premium that unwinds on a press conference. Every day the Strait stays constrained, the energy component of the next CPI gets a little harder and the Fed’s problem gets a little less soluble. JPMorgan raised its year-end S&P target to 8,000 from 7,800 on the strength of the earnings season, and on the earnings evidence alone that is defensible. It assumes funding stays cheap and the Strait eventually opens. Neither assumption improved this morning.

Marks above are pre-market and will move by the open. We will mark Friday’s positioning against this week’s prints and own whatever we get wrong.