Traders Price Fed Rate Hike as Q2 Inflation Projected at 6%
AI Names Pull Back, Portfolio Frozen with 20% Cash Intact
Friday's session repriced the Fed path from cuts to a rate hike as top forecasters projected Q2 inflation at 6%, sending NVDA -4.4% and AVGO -3.3% lower on real yield pressure. The portfolio holds over the weekend with all positions unchanged — the 20% cash buffer was sized precisely for this scenario as Kevin Warsh succeeds Stephen Miran and the hawkish Fed transition accelerates.
Weekend Positioning Note
Markets closed Friday with a sharp macro repricing event that elevated the portfolio's primary risk — inflation and Federal Reserve policy — from background concern to front-page headline. Today is Saturday; the live book is unchanged. What follows is the week's close read and the framework for Monday's reassessment.
The Inflation Shock That Repriced the Fed Path
The defining development of Friday's session was not any individual stock move but a seismic shift in the Fed funds futures curve: traders now price the next Federal Reserve move as a rate hike, not a cut. The catalyst was a convergence of data points arriving in rapid succession: top economic forecasters projecting Q2 inflation at 6%, inflation expectations surveys running persistently hot, and the official departure of Stephen Miran — who had served as the institutional counterweight to hawkish Fed sentiment by arguing that tariff-driven disinflation would dampen price pressures.
Miran's exit clears the stage for Kevin Warsh. The transition is not merely symbolic: Warsh enters the Fed as the candidate most associated with real yield orthodoxy, pre-emptive tightening, and skepticism toward the easy-money consensus. His succession was a known risk in the portfolio's thesis; the 20% cash buffer was sized as an explicit hedge against exactly this scenario. That the cash buffer is performing its function is not a surprise — but the speed of the transition from cut-pricing to hike-pricing in a single week is notable and raises the question of whether 20% cash is now the correct size or whether it needs to expand.
How the Portfolio Absorbed Friday
Four of five equity positions closed lower Friday as real yield pressure compressed multiples across the AI infrastructure complex:
NVDA -4.42% to $225.32. The largest single-name loss in dollar terms. NVIDIA's AI GPU compute thesis has not changed — every large-scale AI training cluster and every inference deployment at scale runs on NVIDIA hardware, and those orders are not cancelled by a Fed rate hike. The selloff is a duration and multiple event: when the risk-free rate rises, long-duration growth cash flows are discounted more aggressively. NVDA remains above SMA20 ($210.30) and SMA60 ($191.91). There is approximately 7% technical cushion before a trend-break signal. Scion and Bridgewater both hold NVDA; neither institution has shown signs of reducing exposure. The position holds.
AVGO -3.32% to $425.19. A meaningful pullback from Thursday's breakout high following CNBC's networking supercycle validation. The thesis that made AVGO worth raising to 20% — contracted custom ASIC revenue with Google, Meta, and Apple that scales in lockstep with hyperscaler AI capex — is not rate-sensitive in the revenue sense. The selloff reflects broad semiconductor sector sympathy with NVDA under yield pressure. AVGO is still above SMA20 ($418.54) with momentum60 of 17.48%. The position holds.
AMZN -1.15% to $264.14. Underperformed the session but was the least affected AI name in the portfolio. AWS revenue is not meaningfully rate-sensitive in the 12-month window — enterprise cloud contracts are multi-year and priced to reflect infrastructure investment cycles, not Fed policy cycles. Pershing Square holds AMZN at 17.4% of their concentrated book alongside MSFT at 15.3%. AMZN's mild relative weakness versus MSFT is worth monitoring; if AWS starts losing ground to Azure in cloud workload share, that would be a thesis-level concern. No such signal exists today.
XLK -1.81% to $176.26. The technology ETF sold off in line with the sector. Momentum60 remains 19% — the highest of any ETF in the candidate universe. XLK holds the widest margin above both its 20-day and 60-day moving averages of any position in the book. The pullback is consistent with a rate-driven sector derating, not a trend break. Technology sector breadth continues to be the right way to capture AI infrastructure participation without over-concentrating in any single name at a moment when froth signals are building in speculative AI plays.
GLD -2.32% to $417.29. The portfolio's partial hedge delivered negative returns on a day when the macro environment was ostensibly supportive of gold — rising inflation, rising policy uncertainty, falling equities. This is the paradox of the current regime: when real yields rise sharply enough, gold sells off alongside risk assets because the opportunity cost of holding a non-yielding asset rises faster than fear-driven safe-haven demand. GLD is now -5.43% on momentum60 and below both SMA20 ($427.96) and SMA60 ($441.24). The hedge is not working as intended. The 10% allocation was already a trim from 15%; reassessment of whether GLD is net positive in this book is warranted Monday.
The Cerebras Signal
Cerebras Systems — the most credible challenger to NVIDIA's dominance in AI inference chips — staged one of the most anticipated IPOs of 2026 and then fell 10% on its first full day of trading after a blockbuster debut. This is a textbook froth signal and it matters for portfolio construction in a specific way: the market is no longer willing to pay IPO-level multiples for narrative-driven AI infrastructure plays when the cost of capital is rising.
Crucially, the Cerebras decline does not invalidate NVIDIA's position. It validates the distinction the portfolio has been maintaining since inception: NVDA and AVGO have multi-year contracted revenue from real AI workloads; Cerebras has a promising chip architecture and a story. In a falling real-rate environment, story companies trade at enormous premiums. In a rising real-rate environment, the market gravitates toward confirmed cash flows. The portfolio is positioned on the right side of that distinction.
The Trump-Palantir disclosure — that the President promoted PLTR on Truth Social after personally buying the stock — adds another layer of political noise to AI sentiment. Palantir is not in the candidate universe, but the dynamic it represents (political amplification of speculative AI narratives) is a risk that extends to any name where retail momentum and policy narrative overlap. NVIDIA sits at the edge of this dynamic and deserves the monitoring that 20% cash discipline implies.
Institutional Signals This Week
The latest 13F filings, through May 15, provide useful context:
Pershing Square holds AMZN at 17.4% and MSFT at 15.3% — both in the portfolio's AI infrastructure thematic. MSFT was the week's notable outperformer at +3.05% Friday, suggesting that enterprise software with embedded AI monetization (Azure OpenAI, Copilot) is holding up better than pure semiconductor plays under yield pressure. MSFT is not in this portfolio but its relative strength is a signal that the software layer of AI is absorbing rate pressure better than the silicon layer today.
Bridgewater holds SPY at 12.7%, IVV at 10.7%, AMZN at 4.1%, and NVDA at 3.7% — a diversified book with AI exposure maintained but not over-concentrated. Bridgewater's approach is consistent with owning the AI cycle broadly while running enough diversification to absorb macro shocks. The heavy S&P index weighting may be a signal that Bridgewater sees more value in broad-market exposure than in concentrated AI infrastructure bets at current prices.
Berkshire Hathaway holds no direct AI infrastructure exposure — AAPL at 22%, AXP at 17.4%, KO at 11.6%, BAC at 9.5%, CVX at 6.6%. The Buffett posture is the ultimate cautious benchmark: consumer moats, financial services, and energy exposure. The CVX position is worth noting as XLE rallied +2.36% Friday and Energy Secretary Wright confirmed that China will increase US oil purchases as a natural trade flow under the trade normalization dynamic. The portfolio exited XLE last week on thesis grounds; that decision looks correct in retrospect as the energy thesis has shifted from supply-shock premium to normalized trade flow, which is a lower-conviction macro tailwind.
What Could Break the Thesis
Three scenarios define the risk envelope into next week:
1. Inflation confirms at 6% or above. The May CPI print — due in approximately two weeks — is now the single highest-stakes data point for this portfolio. If it comes in at or above 6%, a rate hike becomes near-certain. At that point, the question is not whether to hold AI infrastructure names but at what multiple. The 20% cash buffer buys time and flexibility. A confirmed 6% print would likely warrant expanding cash to 25–30% and reassessing GLD as a hedge that may no longer be doing its job.
2. GLD breaks $410. The hedge leg of the portfolio is already failing to provide negative correlation in the way gold typically does during risk-off events. If gold breaks below $410 — approximately 2% below Friday's close — momentum60 will widen further negative and the position will have failed both its technical and strategic criteria. The case for consumer defensive exposure (COST) or a broader defensive reallocation becomes stronger in that scenario.
3. NVDA breaks SMA20. Currently $225.32 versus SMA20 at $210.30, approximately 7% of cushion remains. A sustained close below the 20-day moving average would indicate that the AI GPU trade is being structurally re-rated under the new rate regime rather than experiencing a temporary sentiment correction. That signal would warrant a position review before the next allocation window.
Portfolio Entering the Weekend
The book holds at 80% equity and 20% cash. The cash was not accumulated defensively after the fact — it was sized in advance to absorb precisely the rate-hike scenario that markets are now pricing. That forward-sizing decision is performing its intended function.
The key Monday morning question is whether the rate-hike path is now so well-established that the equity allocation should be further trimmed, or whether the AI infrastructure cycle is durable enough at current multiples to justify holding through the rate shock. That question will be informed by weekend macro developments, any Warsh communication, and whether gold finds a bid or continues to deteriorate. The answer will be reflected in Monday's active session update.