The Fed’s $8 Trillion Balance Sheet Is Sending a Clear Signal
The Fed's $8 trillion balance sheet keeps flooding liquidity as Core PCE and M2 both hit the 90th percentile of their 12-month range. The Federal Reserve is still carrying roughly $8 trillion in assets on its balance sheet, and it is doing so while its preferred inflation gauge climbs and long-term Treasury yields sit near the top of their 12-month range. That is the number that frames every other market signal flashing on July 2, 2026. Look at what has moved alongside this number. M2 money supply has climbed to $23.05 trillion as of May 1, 2026, up $0.25 trillion in a single month, a 1.1% jump that puts the broad money measure at the 90.9th percentile of its 12-month range. At the same time, the Fed's preferred inflation gauge is not cooperating. Core PCE has risen from 126.43 in July 2025 to 130.08 in May 2026, with the most recent monthly reading up 0.3%. That index now sits at the 90.9th percentile of the past year. The bond market is voting with its feet. The 10-year Treasury yield closed at 4.48% on July 1, 2026, up 0.07% on the week and sitting at the 92.4th percentile of its 12-month range.
Trump Says ‘We Hit Them 20 to 1.’ Wall Street Says That Means a Rate Hike Is Coming in October
U.S. strikes on 90 Iranian targets for a second straight day pushed Brent crude toward $79, markets now pricing an October Fed rate hike. The June Fed minutes showed some officials had already made a case for a hike, though the committee stayed on hold. That is the setup Mark Cranfield seized on. "The market is now pretty much convinced that we will get at least one 25 basis points hike probably in October," he told the program. The rates market is corroborating that view in Treasuries: the two-year yield, the maturity most sensitive to Fed policy, climbed to a 2026 high.
Memory monthly sales hit record $74.6B as analysts see price surge ahead
Global memory monthly sales reached a record $74.6 billion, surging 31.7% month-on-month and running 2.8 percentage points above the 10-year seasonal average, according to UBS's July Memory Monthly report — a data point that has both UBS and Bernstein forecasting sharp contract price increases ahead, even as the two firms diverge sharply on how long the rally can last. The primary public beneficiaries of the memory upcycle are Micron Technology, as well as South Korean giants Samsung Electronics and SK Hynix, as the three companies together account for the vast majority of global DRAM and NAND supply. All three stand to capture the contract price gains that UBS and Bernstein now see accelerating through the second half of 2026. DRAM posted record monthly sales of approximately $48.0 billion, up 27.7% month-on-month, though that reading came in roughly 8.0 percentage points below the 10-year seasonal average, suggesting demand is not yet firing on all cylinders across every end-market. NAND, however, was a different story entirely. NAND sales rebounded to a record $25.8 billion, surging 40.7% month-on-month and landing approximately 16.9 percentage points above the 10-year seasonal average — a sign that AI-related storage demand is pulling the segment well ahead of historical norms. "Our July Memory Monthly suggests that the memory upcycle is strengthening further amid accelerating AI-driven demand and ongoing LTA negotiations," UBS wrote in its report. On pricing, UBS raised its DDR contract pricing forecast to increase 32%/18% quarter-on-quarter in the third and fourth quarters of 2026, respectively, with NAND pricing expected to rise 30%/12% QoQ over the same period. The firm projects the DRAM industry to remain structurally undersupplied through at least the second quarter of 2028, underpinned by 2027 bit demand growth of 36.2% year-on-year that is expected to significantly exceed supply growth of 19.3% YoY. High-bandwidth memory is at the core of that demand thesis: UBS projects HBM demand to grow 90% YoY to approximately 33.1 billion gigabytes in 2026, then 77% YoY to roughly 58.7 billion gigabytes in 2027, driven by continued AI accelerator deployments across hyperscalers.
Bond issuance backing AI investment tops $250B, testing limits of voracious investor demand
$31.9 billion of new AI-related bonds through July 8 this year, all but $4 billion of which backed new data centers. IG-rated bonds from traditional hyperscalers (Amazon, Alphabet, Meta, Oracle), data center developers (including Hut 8, Beacon Point, QTS), and other AI-focused concerns (SpaceX, Nvidia, NTT, among others) reached $218 billion through July 8, blasting past an $80.5 billion 2025 total, virtually all of which was placed in the second half of the year. New SpaceX 6.65% 30-year bonds traded above T+200 this week, from pricing at T+175. Meta's 6.30% 2056 bonds, inked April 30 (as part of a $25 billion package), traded to its widest level yet, at T+145, or 13 bps wide of pricing, and versus trades as tight as T+120 a month ago. CoreWeave's par-priced June 11 offering of 9.625% six-year senior notes slumped to 96.50 (10.42%) as the prospect of competition from one of its primary clients (Meta Platforms) dovetailed with heavy markets.
Can AI answer the $3 trillion question?
All told, he calculates that the AI industry will have to earn $3 trillion to justify all those chips and other data center expenditures. And that’s probably an underestimate—the rising costs of memory and the increasing use of exotic or inference-specific chips will drive that number up. On the other side of the ledger, Anthropic is thought to have hit $60 billion in ARR, while OpenAI reportedly earned $13 billion in 2025 (although in November 2025, it said it was at $20 billion ARR) and is presumably making more this year. But there’s clearly a large gap to be closed. Someone minding that gap is Torsten Slok, the chief economist at Apollo, the giant asset manager. In a recent note, he points out that the hyperscalers — Google, Meta, Microsoft and Amazon — are all predicting massive accelerations in their free-cash flow in 2028. That is, they expect to see the pay-back from all those chips they bought. What if they don’t? Slok notes a risk we’re currently seeing across AI usage: More organizations turning to cheaper open weight models, often Chinese, not those built by the frontier labs, and overall token prices falling. OpenAI’s latest model, per CEO Sam Altman, is 54% more token efficient on coding tasks. That’s good for users fretting about the cost of their AI agents, but it may be bad for companies building token factories should users not wildly increase their overall token usage with them. Slok worries that if hyperscalers don’t meet their cash flow goals, the market reaction could be severe— “with so much riding on so few names,” he writes, “a slower payoff wouldn’t just be a sector problem, it would risk tipping the economy into recession and the S&P 500 into a correction.”
The Coming Power War That Will Define the AI Era
Industry forecasts from McKinsey now put AI data center capital expenditure at roughly $5.2 trillion between now and 2030. Goldman Sachs Research projects global data center power demand will surge up to 165% by 2030 compared to 2023 levels. The world simply does not have enough clean, reliable, large-scale electricity to deliver on what the AI industry is promising. Not in the United States. Not in Europe. Not in Asia. The shortage is everywhere, and the timeline to fix it through new generation, transmission, and interconnection runs ten to fifteen years at a minimum. The companies that control electricity may likely be able to dictate terms to the rest of the AI economy for the next two decades. Total contracted revenue runs approximately $2.6 billion, with implied annual revenue of $178 million at full capacity and a net operating margin of 85%.
SpaceX Added to Nasdaq-100 and Could Trigger $4.3B in Forced Passive Buying
Index funds must absorb an estimated $4.3B in SPCX shares as $800B+ in Nasdaq-100 tracking assets rebalance after SpaceX's rapid inclusion. Only 281 million of SpaceX's 7.57 billion shares trade publicly, making the $4.3B forced bid collide with an extremely thin float. The company underneath the flows For readers new to the ticker, SpaceX is now a three-legged business. It launched more than 80% of the world's mass to orbit in recent years, operates the Starlink broadband network of approximately 9,600 satellites serving customers across 164 countries, and folded in xAI's Grok model after an early-2026 acquisition. Trailing twelve-month revenue sits at $19.3 billion per SpaceX's most recent SEC filings, with a diluted EPS of -0.68 and an EV/Revenue multiple near 111x.
Fed minutes expose deep divide over interest-rate outlook
The rate-setting Federal Open Market Committee voted unanimously last month to hold its benchmark Federal Funds Rate target in a range of 3.5% to 3.75%. Federal Reserve Chair Kevin Warsh said July 1 that inflation risks have come down in recent weeks although he didn't offer data or other numbers to support his argument. In a scenario featuring moderating inflation, "most" participants said they expected the central bank would "maintain or eventually lower the target range for the Federal Funds Rate." But "most" participants said that "some policy firming would likely be warranted" if inflation remains elevated. New York Fed President John Williams said July 7 that monetary policy was well positioned and that he expected Headline PCE, the Fed's preferred inflation gauge that's been hitting close to 4%, will dip over the next several months as energy prices stabilize.
Apollo's Sløk: The market faces big risks if hyperscalers' AI profits get delayed
Spending in 2026 by Amazon (AMZN), Meta (META), Google (GOOG, GOOGL), and Microsoft (MSFT) is now expected to cross $700 billion as those companies have gone from being the "source of cash" to the "user of cash," Bank of America's Savita Subramanian and Vivek Arya noted.
The 'Magnificent 7' stocks are trading at their cheapest valuation in more than a decade
The Magnificent Seven stocks are now trading at their cheapest valuation relative to the S&P 500 (^GSPC) in more than a decade, according to Morgan Stanley. Wall Street is growing increasingly impatient with Big Tech's astronomical capital expenditures on artificial intelligence, projected to balloon 70% and exceed $700 billion this year.
The Tech Download: Teen social media bans are missing one big piece — AI chatbots
Roughly half of U.S. teens now use chatbots like ChatGPT, Copilot, and Character.AI for schoolwork, information, or just for fun, according to Pew Research Center. It appears that although AI safety and protecting children are topics dominating headlines, the government is again missing the mark on where the real dangers lie. News Elon Musk's SpaceX joined the Nasdaq 100 index on Tuesday, less than a month after its stock market debut on June 12.
Anduril CEO says it's bad to IPO in ‘middle of a hype cycle’
Spending on the technology is hitting all-time highs, with the defense budget on track to reach $1.5 trillion. Anduril, which makes drones and AI-powered weapons, doubled its valuation in May to $61 billion, becoming one of the most richly valued private tech companies.