The AI spending boom is accelerating — Goldman Sachs projects hyperscaler capex will hit $800 billion in 2026, up from $527 billion consensus estimates. As of June 2026, Morgan Stanley sees $805 billion from the Big Five alone. This capital surge is forcing the Federal Reserve to weigh rate hikes even as Bitcoin miners pivot to AI infrastructure.
What You’ll Learn
- Why Goldman Sachs and Morgan Stanley project $800B+ in AI capex for 2026 and which hyperscalers drive it
- How the AI spending surge feeds inflation pressure and shifts Federal Reserve rate-hike probability
- Where Bitcoin miners fit in the AI infrastructure pivot and what it means for BTC hash rate
- Whether the Cisco-era comparison for Nvidia holds water or misses key structural differences
The AI spending boom reshaping global capital markets just crossed a psychological threshold. Goldman Sachs now sees $800 billion in hyperscaler capital expenditure for 2026, Morgan Stanley lifted its forecast to $805 billion for the Big Five alone, and TS Lombard warns the global figure could breach $800 billion — all revisions clustered in the last 30 days. This isn’t incremental; it’s a 36% year-over-year leap that now rivals U.S. defense and education budgets combined. The spillover has reached the Federal Reserve, where Governor Lisa Cook explicitly tied AI investment to inflation persistence, and Bitcoin, where miners are liquidating BTC holdings to fund GPU clusters. Goldman Sachs' S&P 500 target of 8,000 underscores how AI earnings now drive the broader market narrative.
Goldman Sachs $800B Call: The Numbers Behind the Headline
Goldman Sachs strategist Amanda Lynam delivered the latest upward revision in early June 2026, telling clients the four largest hyperscalers — Meta (META), Microsoft (MSFT), Amazon (AMZN), and Alphabet (GOOGL) — will deploy a combined $5.3 trillion from fiscal 2025 through 2030. That’s an $800 billion increase from the $4.5 trillion consensus just one quarter earlier. The FY2026 consensus now sits at $527 billion, up from $465 billion at the start of Q3 earnings season, per Goldman Sachs Research.
Evercore ISI goes further, modeling ~$800 billion in 2026 spend concentrated in Alphabet, Microsoft, Meta, Amazon, and Oracle (ORCL). CreditSights pegs the top five hyperscalers at $602 billion (36% YoY), with ~75% ($450 billion) earmarked specifically for AI infrastructure — GPUs, servers, networking gear, and data center build-out. J.P. Morgan notes the leaders’ collective capex has tripled from $150 billion in 2023 to over $500 billion in 2026, and AI-related investment contributed more to U.S. GDP growth than consumer spending in 2025.
Named entities: Amanda Lynam (Goldman Sachs strategist), Meta (META), Microsoft (MSFT), Amazon (AMZN), Alphabet (GOOGL), Oracle (ORCL), CreditSights, J.P. Morgan, Evercore ISI.
Fed’s Inflation Calculus: Lisa Cook Connects AI Capex to Rate-Hike Risk
The capital intensity of the AI build-out has migrated from equity research notes to Federal Reserve speeches. Governor Lisa Cook, speaking on March 19, 2026, stated that “artificial intelligence could also be contributing to higher prices, given the large investment in the” data center ecosystem. On May 27, 2026, she told Reuters she is “prepared to raise rates if disinflation does not appear in a timely manner.” Minneapolis Fed President Neel Kashkari echoed the concern, calling the Middle East conflict an “inflationary shockwave” and noting inflation risk now exceeds labor market deterioration risk.
The FOMC’s median projection sees PCE inflation ending 2026 at 2.4% — above the 2% target — with the federal funds rate at 3.50-3.75% after three quarter-point cuts in late 2025. Polymarket’s “No change” odds for the June 2026 meeting hit 98% in May, but the December 2026 hike probability climbed to 67% post-May jobs data — and per our 48% rate hike odds analysis, J.P. Morgan now sees the next move as a 25bp hike, not a cut. The AI spending channel is now explicitly in the Fed’s reaction function: massive capex drives equipment demand, power consumption, and labor competition — all inflationary at the margin.
Named entities: Lisa Cook (Fed Governor), Neel Kashkari (Minneapolis Fed President), FOMC, PCE inflation, Polymarket, Reuters (May 27, 2026).
Bitcoin Miners Pivot to AI: Hash Rate Capital Reallocation
The crypto side of the story is equally concrete. Bitcoin miners — historically pure-play BTC revenue vehicles — are redirecting capital to AI infrastructure. Oracle’s five-year credit default swaps surged from $40 to over $160 by April 2026 as the company issued debt to fund AI capex, per Charles Schwab research. Meta Platforms laid off 10% of staff while maintaining AI spend, interpreted as operating capital repurposing. The Bitcoin hash rate faces structural pressure: Sequoia’s David Cahn estimates a $600 billion annual AI revenue gap that must be filled to justify the build-out, and miner pivots reduce BTC-denominated capital formation.
Simultaneously, Bitcoin ETF flows turned positive in June 2026 after an 18-day outflow streak — BlackRock’s IBIT saw $48M inflow, Fidelity’s FBTC added $12M — suggesting institutional appetite remains despite macro headwinds. The CME FedWatch tool shows rate-hike odds rising, which historically correlates with risk-asset volatility. CoinShares identifies the “high-beta risk” regime: Bitcoin sells off sharply when recessionary impulses hit, but AI-driven productivity narratives could cushion downside.
Named entities: Oracle, Meta Platforms, David Cahn (Sequoia), BlackRock IBIT, Fidelity FBTC, CME FedWatch, CoinShares, Charles Schwab.
Nvidia’s Cisco Moment: TS Lombard’s $800B Warning
TS Lombard’s Dario Perkins drew a direct parallel: Nvidia (NVDA) today occupies the same strategic perch Cisco Systems (CSCO) held in 1999-2000 — central to a transformative infrastructure cycle, highly profitable, yet vulnerable when earnings lag expectations. Cisco’s revenue peaked in 2000 but its stock didn’t reclaim those highs for 18 years. The $800 billion global AI capex figure from TS Lombard matches the scale of the dot-com build-out in inflation-adjusted terms.
The counterintuitive insight: [COMMON BELIEF] Nvidia’s moat is unassailable because it controls 75%+ of AI compute. [DATA] TS Lombard notes Cisco similarly dominated internet routing (80%+ share) yet saw P/E compression from 200x to 15x as revenue growth decelerated. [WHY GAP EXISTS] The difference is monetization visibility — Cisco sold boxes; Nvidia sells compute with recurring software (CUDA) and networking (InfiniBand/Spectrum) attach rates. But the revenue gap (David Cahn’s $600B) remains the critical test.
Named entities: Dario Perkins (TS Lombard), Nvidia (NVDA), Cisco Systems (CSCO), David Cahn (Sequoia), CUDA, InfiniBand, Spectrum. Related: Nvidia’s $300B semiconductor wipeout in the June 5 crash.
Morgan Stanley’s $805B Forecast: David Sacks’ 75% GDP Claim
Former White House AI and crypto czar David Sacks amplified Morgan Stanley’s June 2026 upgrade: $805 billion in 2026 hyperscaler capex (Amazon, Alphabet, Meta, Microsoft, Oracle), rising to $1.1 trillion in 2027. Sacks claimed AI could drive 75% of U.S. GDP growth — a 2.5% tailwind in 2026 and over 3% in 2027. Morgan Stanley’s Mislav Matejka noted “one can’t lose” in the U.S.-China AI race, implying capex projections only rise regardless of stock performance.
The debt financing angle is critical: Morgan Stanley Research expects structured credit markets to absorb the capex wave, citing the $27 billion Meta JV for a U.S. AI data-center campus in 2025 as template. J.P. Morgan estimates AI-related investment now accounts for ~25% of total U.S. market capex. The capital intensity (capex/revenue) for hyperscalers has reached 45-57% — historically unthinkable levels.
Named entities: David Sacks, Morgan Stanley, Mislav Matejka, Meta, J.P. Morgan, Holger Zschaepitz (German market analyst).
What This Means for Your Portfolio: Positioning the AI-Fed-Crypto Triangle
Three actionable takeaways emerge. First, hyperscaler capex visibility extends through 2027 — Morgan Stanley’s $1.1 trillion forecast implies semiconductor equipment (ASML, LRCX, KLAC), data center REITs (EQIX, DLR), and power infrastructure (ETN, VRT) remain structural beneficiaries. Second, Fed rate-hike probability now incorporates an AI capex variable — watch Lisa Cook’s speeches and PCE prints for disinflation confirmation; a surprise 2026 hike would reprice long-duration AI equities. Third, Bitcoin’s correlation regime is shifting: miner capitulation to AI reduces BTC supply pressure, but risk-off from rate hikes dominates short-term price action. The Ethereum ETF inflow reversal (BlackRock ETHA $19.3M, June 2026) suggests institutional rotation within digital assets, not exit.
This is not investment advice. Consult a SEBI-registered advisor.