AMD Earnings Fuel AI Rally|Iran War Rate Hike Risk

AI Chips vs. War

The S&P 500 and Nasdaq hit record highs on May 6 — while oil still sat above $100 a barrel and the Iran war showed no sign of ending. That contradiction is not a paradox. It is a signal about what is actually driving this market.

Advanced Micro Devices reported Q1 revenue of $10.25 billion, up 38% from a year earlier. Its data center segment alone delivered $5.78 billion — up 57% year over year — making AI chips the single biggest engine of the quarter. CEO Lisa Su said customer engagement around the MI450 accelerator series is exceeding initial expectations. The stock surged 17% on the day, extending a one-month rally of more than 60%.

Nine Wall Street firms raised their price targets in a coordinated response. KeyBanc went to $530. Bernstein upgraded outright, projecting over $14 in earnings per share by 2027. Barclays, Bank of America, and Cantor Fitzgerald all moved to $450 or above. The unifying thesis: agentic AI workloads are doubling the server CPU market that AMD competes for, with the total addressable market projected to exceed $120 billion by 2030.

The broader market followed. Meta reported 33% revenue growth in its latest quarter. Alphabet and Amazon both showed accelerating cloud spend. According to Yahoo Finance, roughly 85% of companies reporting recently beat expectations. The AI supercycle is not a sentiment story — it is showing up in actual revenue and margins.

The counter-signal worth watching: AMD now trades at 137 times earnings after a 253% one-year rally. Morgan Stanley raised its target to $410 but held an Equal Weight rating — essentially saying the business is excellent but the price already reflects it. A miss on MI450 deployment timelines, or any softening in hyperscaler capital expenditure guidance, would put the stock's stretched multiple under immediate pressure.

The Rate Hike Nobody Priced

That same AI boom feeding AMD's earnings brought an unexpected voice of caution from inside the Federal Reserve — and it came with a specific warning about interest rates.

Chicago Fed President Austan Goolsbee, speaking at the Milken Institute conference, said the Fed may need to raise rates if massive AI-driven investment spending occurs before productivity gains actually show up in the data. "It's not at all obvious that the Fed would need to lower rates in that case," Goolsbee said. "They might need to raise the rates." That sentence matters because markets have spent two years pricing in the opposite.

The mechanism is straightforward. If companies pour hundreds of billions into data centers based on an extrapolation of recent AI improvements — and those improvements hit diminishing returns — then you get inflation-driving spending without the productivity offset. Goolsbee pointed to autonomous vehicles as a precedent: progress was extrapolated indefinitely, then hit a wall.

The Iran war is compounding the pressure from a different direction. The conflict has pushed WTI crude to nearly $100 a barrel, up from $60 at the start of the year. National gas prices crossed $4.50 per gallon on May 6 — with California averaging $6.16. Gasoline outlays jumped to $503.7 billion annualized in March, up from $422.4 billion in February.

Morgan Stanley cut its US growth forecast by 0.4 percentage points, stating directly that higher gas prices will be "more than enough" to offset the boost from larger tax refunds this year. Real GDP grew just 2.0% in Q1, with personal consumption contributing only 1.6% — the weakest reading in the recent cycle. The savings rate has slipped to 4.0%.

Goolsbee was explicit: the longer oil prices stay elevated, the greater the chance people start factoring higher inflation into expectations — and that would be "extremely problematic" for the central bank. The Fed's dual mandate becomes a trap when energy-driven headline inflation rises while the job market softens. Cutting would feed inflation. Holding punishes growth. Raising — as Goolsbee hinted — would reprice every rate-sensitive asset in the market.

The verification benchmark to watch: if core PCE prints above 3.5% in May while gasoline holds above $4.50, the probability of a rate hike discussion at the June FOMC meeting rises sharply. That scenario would hit REITs, long-duration growth stocks, and mortgage-sensitive housing names hardest. The same AI stocks driving today's record highs would face their first real macro headwind.

Nuclear's Moment

The rate hike risk Goolsbee outlined — driven partly by runaway AI investment — leads directly to the third force moving markets today: the desperate scramble for power to run that investment.

Oklo surged 12% after the US Nuclear Regulatory Commission approved the Principal Design Criteria for its Aurora powerhouse — a foundational regulatory step that directly de-risks the company's path to commercial deployment. The stock has risen 170% over the past year. Its pipeline includes a 12-gigawatt non-binding agreement with Switch through 2044 and a 500-megawatt letter of intent with Equinix backed by a $25 million pre-payment.

Nano Nuclear Energy jumped 27% after announcing a memorandum of understanding with Super Micro Computer to explore powering AI data centers with small nuclear microreactors. The stated goal: create a self-powered, grid-independent AI infrastructure class that can be built anywhere without depending on strained local grids.

The structural reason these moves are not speculative is the power math. Goldman Sachs projects AI-driven data center power demand to rise 165% by 2030. Solar and wind cannot provide round-the-clock baseload supply without storage costs that destroy the economics. Natural gas conflicts with the net-zero commitments hyperscalers have already made to shareholders. Microsoft anchored the Crane Clean Energy Center restart with a 20-year nuclear power purchase agreement with Constellation. Amazon and Google have signed similar deals.

Cameco, the largest publicly traded pure-play uranium producer in the Western world, received a price target raise from Scotiabank to $175 — above the consensus average of $150 — after signing a $2.6 billion CAD uranium supply agreement with India spanning 2027 through 2035. The stock is up 151% over the past year. NuScale Power climbed 10% in sympathy with Oklo's NRC approval.

The bear case on nuclear is real and deserves equal weight. Oklo has no operating reactors. Its Aurora design targets late-2027 commercial deployment at Idaho National Laboratory — a date that has a history of slipping in this industry. Insider selling has been flagged by skeptics. The stock carries a trailing loss of $0.72 per share and runs on regulatory milestones, not revenue.

The condition that would confirm the nuclear thesis: Oklo converting its Switch and Equinix letters of intent into binding contracts before year-end, and AMD or another hyperscaler announcing a direct nuclear power agreement for a new data center campus. The condition that would break it: an NRC licensing delay pushing Aurora's timeline past 2029, or a meaningful drop in hyperscaler capital expenditure guidance that reduces the urgency of the power search.

The three forces in play today — AI chip earnings, a potential Fed rate pivot, and nuclear power demand — are not independent. They share a single root cause. The AI buildout is large enough to move oil-price-sensitive inflation, move the interest rate calculus, and move the energy grid simultaneously. The market is pricing all three in the same session. The question for investors holding any of these names tomorrow is which force the Federal Reserve decides to prioritize first.