Markets head into the June 2 US open with an AI capital story large enough to reset trader positioning across the entire mega-cap complex. Alphabet (GOOGL) said Monday it will raise $80 billion in equity to fund its AI compute buildout, including a $10 billion direct investment from Berkshire Hathaway. A few hours later, Nvidia CEO Jensen Huang called Marvell Technology (MRVL) the next trillion-dollar company at Computex in Taipei, and Marvell shares opened up 22% in premarket trading. With Anthropic also filing for a US IPO at a valuation near $1 trillion, the Alphabet $80 billion AI capex announcement now anchors a broader repricing of the AI infrastructure trade.

TL;DR

  • Alphabet plans an $80B equity raise to fund AI compute, including a $10B direct stake from Berkshire Hathaway.
  • Nvidia's Jensen Huang publicly tagged Marvell as the next trillion-dollar AI infrastructure name. MRVL +22% premarket, +158% YTD.
  • Anthropic confirmed it filed paperwork for a US IPO at a roughly $965B private valuation.
  • Combined hyperscaler capex is on track to exceed $700B in 2026 and could top $1T in 2027.
  • Equity market reaction skews positive for AI infrastructure suppliers and negative for hyperscaler near-term free cash flow.

What happened

Alphabet disclosed an $80 billion capital raise after the US close on Monday, June 1. According to its press release, the structure is split into three buckets: $10 billion from Berkshire Hathaway through a private placement, $30 billion in underwritten public offerings (including $15 billion of depositary shares representing mandatory convertible preferred stock), and $40 billion through an at-the-market program for Class A and Class C shares that is expected to begin in the third quarter. Goldman Sachs, JPMorgan, and Morgan Stanley are bookrunning the underwritten offerings.

The company framed the raise as a response to AI demand it says is exceeding available supply. Alphabet had already revised its 2026 capex outlook in April to between $180 billion and $190 billion, and management has flagged compute capacity, power, land, and supply chain access as the principal constraints on its build pace.

Hours later, Nvidia CEO Jensen Huang shared a stage at Computex Week in Taipei with Marvell CEO Matt Murphy and described Marvell as essential to the AI data center stack. He singled out Marvell's networking and connectivity chips as the layer that lets disaggregated compute clusters communicate at scale, then called Marvell "the next trillion-dollar company." Marvell shares were trading up 22% as of 7:28 a.m. ET in Tuesday's premarket, extending a year-to-date gain that exceeds 158%.

Separately, Anthropic confirmed Monday it has filed paperwork for a US initial public offering. Its most recent private round priced the company at more than $965 billion, ahead of OpenAI's last reported $852 billion mark. The price and number of shares have not been set. Anthropic told investors it expects to be profitable in the first half of 2026.

Why markets moved on the AI capex story

Three signals landed within roughly 12 hours of each other, and each pushed a different lever in the AI trade.

First, Alphabet's $80 billion raise quantifies how aggressive hyperscaler capex is becoming. Equity issuance of this size is unusual for a profitable mega-cap. It tells the market that internally generated cash and prior debt issuance (over $30 billion in February, $11 billion in European currencies, and a $25 billion bond sale last November) are no longer pacing demand. Berkshire's $10 billion follow-on, layered onto a roughly $20 billion existing stake, signals continued long-duration conviction from one of the most patient holders on the tape.

Second, Huang's Marvell endorsement reinforced the picks-and-shovels trade. When the dominant AI chip supplier publicly names a connectivity vendor as critical infrastructure, fast money rotates quickly. Marvell already beat estimates in late May with $2.4 billion in quarterly revenue and raised its data center outlook, so the move has fundamentals underneath it.

Third, Anthropic's IPO filing puts a public market clock on the AI pure-play category. Combined with SpaceX's expected listing and a potential OpenAI float, the IPO calendar concentrates a historic amount of pre-IPO capital into a narrow window.

Implications for traders

US equities (stocks). The dispersion between AI infrastructure beneficiaries and the rest of the tape is likely to widen at the open. Marvell, Nvidia, Broadcom, Arista, and other connectivity and networking names should see immediate flows. Alphabet itself slipped in extended trading, a typical equity-issuance discount, which can pressure mega-cap weighted ETFs at the open before a recovery if the AI thesis holds. Software stocks, which CNBC flagged as crossing a notable milestone earlier in the morning, can benefit from positive cross-currents in tech sentiment.

Index futures. Nasdaq 100 futures (NQ) will react more sharply than S&P 500 futures (ES) given the AI weight concentration. Watch for early sector rotation: industrials and utilities tied to power and data center infrastructure can attract incremental bids on the capex narrative.

Bonds and credit. Mega-cap equity issuance at this scale modestly relieves pressure on investment-grade credit supply expectations from hyperscalers. Spreads in IG tech may tighten at the margin.

Crypto. The crypto tape has been heavy on its own narrative (record Bitcoin ETF outflows and the recent Strategy headlines), so AI capex enthusiasm is not the primary driver here. Still, a strong risk-on session in US equities can spill back into BTC and high-beta altcoins if equity gains hold into the cash open.

Cross-asset context. Even though US equities printed an S&P 500 record high into Monday, a parallel oil shock complicates the AI-led rally. Traders should size the AI-infrastructure trade with the energy and geopolitical risk backdrop in mind.

Potential positives vs negatives

Positives. The capex commitment validates AI demand as durable rather than seasonal. Suppliers across the stack (Marvell, Broadcom, Arista, ASML, TSMC, Micron) have a multi-year revenue runway. Berkshire's vote of confidence reframes Alphabet's raise as long-duration capital formation, not dilution panic. Anthropic's profitability claim, if confirmed in the IPO prospectus, would be the first hard datapoint that an AI pure-play can hit operating breakeven, which would strengthen valuations across the category.

Negatives. An $80 billion equity raise is dilutive. Even with strong underlying demand, near-term GAAP earnings per share at Alphabet will face pressure from the new share count. If AI revenue growth slows before capex normalizes, free cash flow margins will compress. Analyst skepticism is already visible on the Bloomberg tape, with Wall Street strategists turning more cautious after a two-month rally. A concentrated IPO calendar (Anthropic, potentially OpenAI, SpaceX) could pull liquidity out of secondary AI positions in the second half of the year.

What could happen from here

Base case. AI infrastructure names trade firmly higher at the open, led by Marvell, with broader semis participating. Alphabet opens modestly lower on issuance overhang, then stabilizes as buyers digest the Berkshire signal. Nasdaq futures hold gains into the cash session. Confirmation: MRVL holds above the premarket level on first-hour volume, NQ stays green at the bell, IG credit spreads contract slightly. Invalidation: MRVL gives back more than half of the premarket pop in the first 30 minutes, or NQ rolls red on broader risk-off catalysts.

Upside case. The Berkshire stake reframes the Alphabet raise as a long-duration vote rather than a capital call, GOOGL recovers in cash hours, and a rotation into AI infrastructure broadens to power, cooling, and physical buildout names. Confirmation: GOOGL closes green, breadth in semis (SOX) outperforms QQQ, and utilities tied to data center load (VST, CEG, NRG) lead the S&P 500. Invalidation: GOOGL closes more than 2% lower despite a green broad tape.

Downside case. The combination of mega-cap dilution, IPO supply concerns, and existing analyst skepticism on the rally tips equities into a profit-taking session. Confirmation: NQ rolls red in the first hour, MRVL fades back to flat, and credit spreads widen. Invalidation: NQ remains positive with leadership broadening beyond the AI infrastructure cohort.

What traders should watch next

  • The Alphabet press release follow-up: confirmed timing of the $40B at-the-market program. Faster execution suggests urgency in compute deployment and can pressure GOOGL near term.
  • Marvell sell-side responses through the morning. New price targets above current levels would extend the rally; muted reactions would invite profit-taking.
  • Anthropic S-1 disclosures when they hit. Revenue run rate, gross margin, and inference cost trajectory will set the valuation framework for the broader AI category.
  • Power and grid names (VST, CEG, NRG, GEV) for capex follow-through.
  • Credit market reaction in IG tech spreads, given the volume of debt and equity issuance already in 2026 from hyperscalers.

FAQ

Why is Alphabet raising $80 billion in equity instead of issuing more debt? The company has already issued over $66 billion in bonds across multiple currencies in the last six months. Equity reduces incremental leverage and signals that the buildout is multi-year, not a one-time spend. In our internal logs at chartdojo.io we have seen active-trader users rotate aggressively into AI infrastructure leaders on prior capex announcements, so positioning for follow-through matters.

Is Marvell's 22% premarket move sustainable? The endorsement from Huang is meaningful, but moves of this size often see retracements once the cash session opens. Watch first-hour volume relative to the 20-day average and whether the stock holds the premarket VWAP.

How ChartDojo traders should approach this

When we tested this kind of catalyst-driven AI dispersion on chartdojo.io, the highest-conviction setups came from waiting for the first-hour range to form before adding exposure. Chasing premarket pops without a defined invalidation level is one of the most common pattern errors our practice-feedback engine flags. The discipline is to size into confirmation, not headline excitement.

Build real skills, not bad habits. ChartDojo gives active traders AI-powered feedback on every paper trade so you can pressure-test these setups before risking real capital. Start your free trial and put your reaction to the next AI capex print on the clock.

Sources

Feature image source: Sundar Pichai, photo by Lukasz Kobus / European Commission (CC BY 4.0).

This article is for educational purposes only and is not financial advice. Trade at your own risk.

Header image source: pexels by Pixabay. License: PEXELS.