Oil’s renewed surge reset the day’s market tone before most other stories could settle. Brent crude climbed toward $105 a barrel as strikes on shipping in and around the Strait of Hormuz intensified, including a rare attack on a tanker off the coast of Qatar reported by a unit of the UK Navy. Shipments through the waterway have recovered somewhat in recent weeks, but the tempo of attacks has risen with them. Equities fell for a second session, short-dated Treasury yields jumped as traders raised the odds of another Federal Reserve increase this year, and longer-dated bonds swung as investors weighed inflation risk against an auction calendar that still found buyers at elevated yields.
The transmission was straightforward. Higher energy costs revived the fear that inflation would prove stickier, and that the central bank would have less room to wait. Two-year yields rose nearly five basis points to 4.81 percent. The ten-year yield edged up to 5.29 percent. Thirty-year yields, after an early loss, finished a basis point lower at 5.66 percent, just off the highest levels since 2002, ahead of a $22 billion sale of the long bond. Traders priced roughly a one-in-five chance of a rate increase later this month and a full hike by year-end. The S&P 500 declined, with chipmakers adding to the drag. Fed Governor Christopher Waller, speaking in Istanbul at an event organized by Turkey’s central bank, said further hikes would likely be needed if the data continued to arrive as expected, though officials had flexibility on timing and did not need to tighten at consecutive meetings. “The hikes do not need to come at consecutive meetings, but they should be in place in an acceptable period of time,” he said. Weekly jobless claims eased to the lowest level since July, reinforcing the picture of limited layoffs even as energy prices complicated the inflation path.
Fawad Razaqzada at Forex.com framed the oil move as intensifying inflation concerns and adding upward pressure on bond yields, with depleted inventories supporting replenishment demand and the longer-term direction of prices still tilted higher absent progress toward a US-Iran arrangement. Bret Kenwell at eToro argued that stocks can absorb higher yields but that the margin for error is narrowing after the market’s recent advance: elevated borrowing costs pressure valuations and raise the bar for earnings, shifting the burden of proof back to companies that have so far supported the rally with profit growth. Angelo Manolatos at Wells Fargo Securities read the turnaround in the long bond as a sign of strengthening end-user demand, especially after Wednesday’s $39 billion ten-year auction cleared at a yield below the prevailing market and drew a record 97.5 percent take-down from non-dealer investors. Molly Brooks at TD Securities said yields appeared to be meeting resistance that was drawing buyers back, and that a rally that held for a few days could pull in accounts that had been waiting for a clearer signal. The Treasury’s long-bond sale was to be followed by a buyback of as much as $6 billion of bonds maturing in 20 to 30 years, the fourth such operation since the department expanded the program as long-term yields surged.
The oil market itself, however, has become less reactive to presidential rhetoric even as the conflict continues to shape the macro outlook. A Bloomberg analysis of price moves found that comments from President Donald Trump, which early in the war could move crude by several percent in a session, now often pass with little lasting effect. On April 1, a televised warning that the United States would strike Iran and send it “back to the stone ages” unless a peace deal was signed helped push prices more than 7 percent higher, above $109, amid threats against electric generation plants that were not carried out. On October 1, a warning that Iran should sign a ceasefire or cease to exist barely moved the market. Traders interviewed by Bloomberg said they had shifted attention from headlines to physical supply and cargo flows, and had cut position sizes because US strategy remained unpredictable. John Kilduff of Again Capital described months of “headline-o-rama” and said participants he speaks with are trading smaller. Rob Thummel at Tortoise Capital Advisors said traders were paralyzed by White House reversals. Michael O’Rourke at JonesTrading noted that after repeated rounds of comments the market becomes desensitized, and that physical-flow data is treated as more actionable.
Intraday ranges have narrowed relative to the opening phase of the war, and volatility has eased alongside fewer strikes on energy infrastructure. Open interest in Brent futures has fallen to the lowest since March 2025. JPMorgan analysts wrote that they no longer know how to model the endgame. Algorithmic commodity trading advisers, which tend to follow momentum, can still amplify swings that later reverse, creating brief disconnects between paper and physical markets. The underlying supply shock remains severe. Combined with Ukrainian strikes on Russian refiners, the conflict has pushed retail fuel prices to records in many places, prompted European and other central bankers to flag stubborn energy-driven inflation, and created political pressure for Trump’s party ahead of November midterms. The desensitization matters beyond crude: with oil now embedded in the inflation debate, a murkier and less liquid market clouds the path for rates and growth even when presidential statements no longer clear the tape.
Singapore Prime Minister Lawrence Wong, speaking at the Forbes Global CEO Conference, treated the energy risk as less acute than the technology cycle. He said the situation was one in which Singapore was prepared less for a sudden and severe dislocation around the Strait of Hormuz, and that that risk had come down as stockpiles and alternative supplies contained costs. On the tech rally that has powered regional trade, he was more cautionary: a correction will come, though neither the timing nor the casualties are knowable, and no boom is indefinite. Singapore’s approach, he said, is to use the present window. Non-oil domestic exports have surged on semiconductor and hardware demand; the government recently lifted its GDP growth forecast to a range of 4.5 to 5.5 percent. The city-state accounts for roughly 10 percent of global chip production and 20 percent of semiconductor manufacturing equipment.
The AI capital cycle supplied the day’s clearest evidence that public markets are beginning to discriminate inside that boom. Australia’s Firmus Grid, an Nvidia-backed data-center developer marketing a $5.5 billion listing at A$11 a share and a path toward a $30 billion valuation, closed order books without a clear indication of price or structure after demand failed to support the marketed terms. Days earlier the company had cited indications of interest well above the offer size. Investors and advisers cited the absence of a proven operating record—Firmus runs two data centers against a pipeline of 912 megawatts, of which 46 megawatts have been built—customer concentration on names such as Meta and OpenAI, a free float that would leave about 58 percent of shares eligible to trade from day one, regulatory scrutiny of data centers, and the need for repeated equity and debt raises. UniSuper’s chief investment officer, John Pearce, said the story was compelling and the valuation was not: “So much has to go right.” Phil Wool of Rayliant Global Advisors called the episode a historical fail for what was set to be one of Australia’s largest IPOs. Maas Group, which holds a stake and at least A$855 million in related electrical contracts, fell as much as 30 percent in Sydney before paring the loss to 22 percent, and said there had been significant speculation over whether the IPO would proceed. Firmus was founded in 2019 as a bitcoin miner by Oliver Curtis, who served a prison sentence for insider trading, with Tim Rosenfield and Jonathan Levee. A fundraising round in early August that included Jane Street and Blackstone valued the company at $10.5 billion; revenue in the 2026 financial year was $51 million. JPMorgan was a joint lead manager alongside Bank of America, Morgan Stanley, and Morgans, even as the bank pitched an approximately 11 percent yield on a $5 billion leveraged loan for Volta Infrastructure’s Norway data-center complex.
The Firmus stumble sat inside a wider set of warnings. Accelevation priced a US debut below range last month. South Korea’s Kospi is down 27 percent from its June peak as the memory-chip surge around Samsung and SK Hynix cooled. Ray Dalio described AI as a classic bubble near bursting because of the debt taken on to fund it and rising rates. Michael Burry wrote on X that the market was in the first stage of grief—denial—and that in 2000 and 2008 that stage lasted six to nine months. Bain projects the industry needs $6 trillion in annual revenue by 2031 to justify data-center capital. KKR estimates $8 trillion is required to complete the global buildout. Oracle cited force majeure in late September on a New Mexico project. Apollo’s infrastructure head, Olivia Wassenaar, said the firm is working to keep financed deals from being derailed by local opposition. Maxence Visseau of Arkevium Capital said investors still believe in AI but will not pay any price for companies that spend heavily, depend on a few customers, and promise profits years out. Anthropic is targeting a large IPO as early as next month, so the test of public appetite is not finished.
Banks are financing that buildout even as the cracks appear. Barclays analysts Peter Troisi and Ishika Goyal estimated that the six largest US banks will issue about $41 billion of debt in the fourth quarter, 30 percent above the fourth-quarter average since 2015, after $50 billion of senior issuance in the third quarter, more than double the prior year. Year to date the group has borrowed $192 billion globally, up about 40 percent, with a full-year estimate of $233 billion. For US banks overall, Barclays now forecasts $294 billion of senior debt, 25 percent above its initial estimate and up from $240 billion in 2025. The funds support hyperscaler and tech spending—from SpaceX to Oracle—and the trading activity that has accompanied it. Yield premiums have stayed stable enough that the supply has been relatively well digested, giving banks an incentive to stockpile. Potential delays in tech IPOs could cap issuance. The seasonal post-earnings window opens next week: Wells Fargo, JPMorgan, Goldman Sachs, and Citigroup report on October 13, Morgan Stanley and Bank of America the following day. Bloomberg Intelligence’s Arnold Kakuda said Goldman and Morgan Stanley, leaders of 2026 issuance, may be the most active afterward. JPMorgan’s own strategists, including Kabir Caprihan, forecast $24 billion of US bank supply in October and noted that Wells Fargo and Citigroup already raised $18 billion off-cycle last month, so the bulk of post-earnings deals may come from Morgan Stanley and Goldman, with Bank of America possibly staying out. Higher benchmark yields and growing credit unease around hyperscalers were flagged as restraints.
Private-market winners from the same cycle continued to mark paper gains. Tiger Global, which agreed to invest $150 million in OpenAI in 2021 at a $15.7 billion valuation after a call from Sam Altman a year before ChatGPT’s public release, and which has added to the stake since, is positioned for a paper profit of about $5 billion if talks to raise $30 billion at a $1.4 trillion valuation conclude on those terms. The firm cannot fully realize the gain until an IPO that Altman has pushed at least into next year. The position has helped the hedge fund and the public-private crossover fund, up 9.4 percent and 19.3 percent respectively through the first nine months. The latest venture vintage, Private Investment Partners 16, a $2.2 billion fund that is the firm’s smallest in about a decade and also holds Anthropic, Waymo, and Temporal, showed a gross internal rate of return of 65 percent and a net IRR of 44 percent as of June 30. Tiger, now managing about $60 billion after assets fell from $100 billion to $50 billion in the 2022 drawdown, was the first hedge fund into OpenAI; earlier venture and individual backers included Khosla Ventures in 2019 and, around 2015, Y Combinator, Reid Hoffman, and Peter Thiel. Chase Coleman has taken a more direct role in illiquid bets after the firm returned to smaller, more concentrated wagers. Prior large wins included more than $6.5 billion on JD.com and $3.5 billion on Flipkart.
Josh Kushner’s Thrive Capital has used a similar run of AI and SpaceX outcomes to widen its mandate. Bloomberg’s Billionaires Index places Kushner’s fortune at about $17.5 billion. Thrive, founded 16 years ago, now manages more than $65 billion, up from about $15 billion three years ago, when investors including Mukesh Ambani, Jorge Lemann, and Henry Kravis came in at a $5.3 billion valuation. More than half of assets now come from investment gains. Bloomberg calculates more than $6 billion in carried interest since founding, most of it in the past three years, and more than $1 billion of liquidity in the past 12 months. A similarly sized minority stake is being sought; on asset growth since the last round, Bloomberg calculates a firm value of about $23 billion. Thrive first invested in SpaceX at a $38 billion valuation ahead of the company’s $1.77 trillion IPO in June, and in Cursor at $400 million and later $9.9 billion before SpaceX acquired it for $60 billion. OpenAI, Stripe, and Anduril remain large private holdings. Thrive Holdings has raised more than $2 billion, including from SoftBank, Altimeter, and D1, at a $12 billion valuation, and has acquired more than 70 accounting, IT, and similar businesses it intends to reshape with AI, with a dedicated team of 35. Thrive Eternal targets assets Kushner describes as rooted in identity, culture, community, history, and physical experience. It took a minority stake in the San Francisco Giants in April and, with Bob Iger, moved within a day to help finance the Los Angeles Lakers at $12.5 billion after Mark Walter became embroiled in a federal investigation—25 percent above what Walter paid less than a year earlier, and well above the $6.1 billion Celtics sale. The NBA amended its private-equity rules to allow firms to invest alongside managers, a change accelerated by the transaction. Projections circulated to investors envisage team earnings rising from $129 million last season to about $600 million over a decade through ticketing, sponsorship, and a possible new arena. A FIFA commercial-arm financing of as much as $4.2 billion that Thrive Eternal was expected to lead collapsed within days amid opposition from regional bodies; UEFA has asked a federal judge to authorize subpoenas for documents and testimony as it explores potential action. Kushner said the firm failed to appreciate the political dynamics of global football. UEFA has not accused Thrive of wrongdoing.
Google moved to convert infrastructure spending into corporate revenue with a Gemini workplace agent that operates inside Gmail, Drive, Docs, and Calendar, choosing models, applying cost controls, and carrying administration and governance features. Thomas Kurian, chief executive of Google Cloud, said the agent is built to work like a colleague, with its own corporate identity and the ability to pursue a goal across several days, rather than as a chatbot confined to a single app or a preset task. The launch is aimed at the business contracts Google needs to justify hundreds of billions of dollars in AI infrastructure, in competition with Microsoft, OpenAI, Anthropic, and Amazon. Some Google employees have worried that cutting-edge capabilities, particularly in coding, have fallen behind rivals. The company has begun sharing a next flagship model, Gemini 4 Argon, with trusted cybersecurity partners.
The same Washington that is courting technology investment moved against a core labor channel for the industry. The administration said it would indefinitely suspend multiple firms, including Microsoft, from the Permanent Labor Certification Program that lets employers sponsor foreign workers for permanent residency, alleging illegitimate representations in applications. Vice President JD Vance said no company had abused the system more than Microsoft, and urged Microsoft and Adobe to employ more Americans rather than “importing an indentured servant.” Labor Secretary Keith Sonderling named Cognizant, Infosys, Tata, Wipro, HCL Technologies, and Capgemini, and said multiple federal investigations were active. The companies did not immediately comment. The suspensions sit alongside tighter H-1B rules and a broader restriction of legal and undocumented migration pathways. Business groups have warned that the labor supply for growth is at risk. The timing was conspicuous: Vance’s remarks came hours before President Trump was set to award Satya Nadella a National Medal of Science at a Washington summit on US technology, and a week after Nadella joined other executives at a White House luncheon on AI safeguards. Vance said the relationship with Microsoft was strong and that the aim was not to harm the company but to have it thrive by employing Americans.
Wall Street’s own compensation cycle produced a separate, internal windfall. About 20 Goldman Sachs executives are in line for equity awards, to be finalized later this month, exceeding $500 million at the current share price under the Shareholder Value Creation Award introduced in October 2021. Chief Executive David Solomon is set for more than $100 million of the pool; President John Waldron and business heads Ashok Varadhan, Dan Dees, and Marc Nachmann are also among the recipients. Half the value depended on the stock reaching targets that Goldman’s price has exceeded by enough to trigger the maximum; the other half is relative performance against six peers, where Goldman has trailed only Bank of New York Mellon, whose shares returned 174 percent over the period under Robin Vince, Solomon’s former chief risk officer. Goldman shares are little changed this year but have returned about 150 percent over five years including dividends. The program was initially limited to the top two executives, then widened after shareholder questions; their combined grant was then worth about $50 million. Spokesperson Jennifer Zuccarelli said the incentives aligned pay with performance, continuity, and retention, and that the firm has performed exceptionally well since. The board’s “war for talent” rationale had irritated employees after a $10 million deduction from the CEO’s pay earlier in 2021 over the 1MDB scandal. Most of the current 45-person management committee will not share in the 2021 award. JPMorgan’s Jamie Dimon holds a long-term award from the same era now worth about $270 million. After 2008, Goldman options granted in lieu of bonuses eventually yielded billions for partners.
Consumer staples offered a more ordinary earnings revision. PepsiCo cut its outlook for a measure of earnings growth to 1 to 2 percent for the fiscal year, from 4 to 6 percent, as North American beverages weakened and the snack recovery lagged. Third-quarter organic revenue declined slightly in North America for both food and beverages, while international sales continued to grow. Chief Executive Ramon Laguarta said the company does not feel good about the beverage business and is concentrating urgency there. North American beverage volumes are down 3 percent this year, with zero-sugar and flavored options outperforming full-sugar; investment will increase behind Poppi, acquired last year, as well as Pepsi and Mountain Dew. Earlier price cuts on salty snacks reversed a sales decline into slightly positive territory but delivered less volume than hoped, and margins were hit after bags had reached more than $7 and cost shelf space. Some prices will rise in coming months, though Laguarta said they should still sit below 2025 levels and that the prior cuts give flexibility to absorb a new wave of inflation. The company continues to add protein, fiber, simpler ingredients, and alternative oils including avocado. A pilot integrating snacks and beverages is working in some regions; elsewhere PepsiCo will look at refranchising bottling. Shares rose 1.2 percent by late morning; they were down 14 percent on the year through Wednesday against a 14 percent gain in the S&P 500. RBC’s Nik Modi called the results marginally better than feared, while noting that the rate of improvement has stalled under inflation pressure.
A separate consumer-sector rumor moved restaurant stocks without confirmation. Chipotle jumped as much as 8.6 percent after the Financial Times reported that Starbucks has worked with advisers in recent months on a takeover proposal. The status could not be learned, and a deal of that scale might never proceed. Brian Niccol has led Starbucks since 2024 after six years at Chipotle. Starbucks declined to comment; Chipotle could not be reached. Starbucks shares fell as much as 6.7 percent, the most intraday in more than a year, before paring the loss. Chipotle’s market value is about $41 billion, which would place a transaction among the largest in the restaurant sector; the stock is down about 20 percent over the past year. Melius analyst Jacob Aiken-Phillips said synergies were not obvious given the operational differences, though Niccol’s knowledge of Chipotle, access to its fresh-ingredient supply chain, and Starbucks’ effort to improve food could be part of any rationale. Niccol has criticized Starbucks food counters as insufficiently bountiful and objected to cardboard menu photography as the company reworks store design, beverage speed, and the food lineup.
Brazil’s presidential runoff supplied an emerging-market counterpart to the market’s search for credible fiscal stewards. Flávio Bolsonaro’s first-round finish ahead of incumbent Luiz Inácio Lula da Silva on October 4 has drawn a deeper bench of potential finance ministers than the campaign possessed when its main economic face was Daniella Marques, former president of Caixa Econômica Federal and now on leave from Legend. Names circulated among campaign and financial-circle contacts include Marcelo Kayath, 62, former Credit Suisse investment-banking head in Brazil and now a partner at QMS, who has argued for curbing mandatory spending and stabilizing debt, though he backed Lula in 2022 and was considered for a central-bank board seat in 2023; Mansueto Almeida, 59, former Treasury secretary under Jair Bolsonaro, architect of spending and pension efforts before leaving in 2020, and now chief economist at BTG Pactual; Roberto Campos Neto, 57, central-bank president from 2019 through 2024, now vice chairman and global head of public policy at Nubank, whose tenure included banking competition and the Pix payments system but whose experience is monetary rather than fiscal; and Alexandre Bettamio, Bank of America’s co-chair of global corporate and investment banking, recently considered for the top job at B3, with no prior government role. No decision has been made. Campaign manager Senator Rogério Marinho has indicated major personnel announcements are unlikely now. Marques remains a possible finance minister or a candidate for another post.
Adriana Dupita of Bloomberg Economics said a market-fluent name could extend the asset rally if Bolsonaro wins, but that what is most lacking is a feasible plan for the fiscal adjustment markets expect, and that knowledge of the public sector and Congress matters as much as market ties. The budget deficit is running near 10 percent of GDP, and fiscal concerns have kept borrowing costs high even as the central bank lowers its benchmark. Bolsonaro has pledged to curb spending, reduce bureaucracy, and cut taxes without a detailed debt-stabilization plan. He is treated as a heavy favorite; no second-place finisher in the first round has won a runoff since Brazil returned to democratic elections in 1989. Lula is considering naming Vice President Geraldo Alckmin finance minister if reelected, Reuters reported. Nelson Marconi of Fundação Getulio Vargas said any president will weigh political implications heavily, and that market pressure for a severe adjustment will take time to produce agreement. The elite’s renewed interest echoes the US corporate re-engagement with Trump between the aftermath of January 2021 and the 2024 election.
In Europe, the personnel calendar at the European Central Bank advanced without resolving the presidency. Nominations to succeed Isabel Schnabel, who is moving to the International Monetary Fund and whose board term ends in January, must be submitted by October 28, with agreement on a candidate possible as soon as November 9, according to people familiar with a timetable set by euro-zone finance ministers in Luxembourg. The schedule does not address a successor to President Christine Lagarde. Lagarde has hinted at leaving before her term ends next October, most recently telling a French newspaper she might quit a few months early. Greek leadership has spoken of a possible grand bargain covering her post, Schnabel’s seat, and Chief Economist Philip Lane’s term ending in May. Italian Finance Minister Giancarlo Giorgetti has asked her to clarify; she has not. German Finance Minister Lars Klingbeil said Germany will retain an important role and will put forward a proposal from within the federal government. Potential German names include Princeton’s Markus Brunnermeier, Berkeley’s Ulrike Malmendier, Tobias Adrian, who recently left the IMF post Schnabel will take, and former Finance Minister Joerg Kukies. EU leaders hold the final say; a December summit decision is plausible given the January vacancy.
Prediction markets supplied a microstructure story with macro pretensions. A Bloomberg analysis of more than a billion Kalshi transactions found that since the start of 2025 about half of trading volume has involved one side chasing potential returns of at least 1,000 percent against more sophisticated counterparties and market-makers; in recent weeks that share has exceeded 65 percent as sports parlays proliferated. On Polymarket International’s public ledger, trades on unlikely outcomes accounted for 52 percent of volume over 21 months. Polymarket said the cash put up by longshot bettors is about 1 percent of volume—the bulk of the stake is on the offering side—but that 1 percent still exceeded $600 million in September across the two platforms. After wins and losses since early 2025, money on “yes” longshots lost 15 percent by settlement on Kalshi and 27 percent on Polymarket, before fees that can reach 7 percent. Contracts priced to pay out 2.2 percent of the time paid out 1.9 percent on Kalshi. Parlays were 58 percent of Kalshi contracts in September; roughly two-thirds of recent parlay contracts carried less than a 1 percent chance of resolving yes, and ultra-long-odds parlays paid out about 40 percent less often than prices implied. World Cup winner contracts drew more than $1.5 billion in notional volume, more than two-thirds of it on teams with less than a 10 percent chance, including $165 million on Mexico and the United States. Kalshi’s monthly volume has grown roughly twenty-fold in a year, surpassing $60 billion in September, with only a small share outside sports, parlays, and crypto. Philip Newall of the University of Bristol called the puzzle why participants cannot see they are the sucker; academic work since Richard Griffith’s 1949 horse-racing study has documented longshot bias. The platforms, overseen by the CFTC rather than state gambling regulators, have begun voluntary safeguards—Kalshi as a platinum member of the National Council on Problem Gambling, Polymarket with deposit limits and self-exclusion—while maintaining that exchange structure differs from a sportsbook taking the other side.
Two further institutional stories sat at the edge of the capital-markets day. Michigan’s attorney general sued Blue Cross Blue Shield of Michigan, alleging the plan conspired with other Blue Cross plans to divide territory and reduce competition, raising costs for state workers’ health plans and for consumers and businesses. The case is State of Michigan v. Blue Cross Blue Shield of Michigan Mutual Insurance Company in the Eastern District of Michigan. Blue Cross plans nationally have agreed to pay billions to resolve other allegations of anticompetitive arrangements; the Michigan plan had no immediate comment. Separately, the NFL’s $192 million public pledge since the end of 2025, alongside Ariel Investments, Sixth Street, and Bessemer, toward flag football—and a professional league intended to launch next year—has coincided with a 48 percent annual jump in high-school girls’ participation, to more than 100,000, and a rise of more than 550 percent in five years, with the league estimating about 800,000 girls ages 6 to 17 competing in some form. Forty states have sanctioned or are sanctioning the sport; the NCAA is expected to recognize a women’s championship in January; the Olympic debut is set for Los Angeles in 2028. Brian Flinn, the league’s senior vice president for global flag football, described the effort as both fan development and a long-term business decision, citing research that childhood sports participation predicts adult fandom and internal findings that half of girls joining flag teams had not been playing anything. The girls’ championship broadcast peaked at 1.34 million viewers. The capital logic is the same one visible in Thrive’s sports book: scarce cultural assets and new participation channels are being financed on the expectation that audiences compound.
The immediate calendar is dense. Bank earnings begin October 13, and with them the next wave of senior debt tied to AI financing. The thirty-year auction and the accompanying buyback will test whether the bid that appeared in the ten-year persists as oil holds near $105 and two-year yields price a hike by year-end. Schnabel nominations close October 28. Firmus has yet to state a price or confirm the deal. Brazil’s runoff later this month will determine whether the new names around Bolsonaro become officeholders or footnotes. OpenAI’s raise, if completed at the discussed valuation, will mark another paper gain for Tiger and Thrive without creating exit liquidity. Waller’s flexibility on timing leaves the Fed’s next move dependent on data that energy prices are now complicating. Across these threads the common constraint is the cost of capital: it is high enough to rattle an AI infrastructure IPO, rich enough to pull buyers into long Treasuries, and still low enough, in credit-spread terms, for the largest banks to keep borrowing against a buildout whose public-market test has only begun.
