Subscribe
Share
Share
Embed
Hosts: Alex Torres & Sarah Chen
In this episode:
• Welcome to Pivot Invest for Saturday, May 9th, 2026. I'm Alex Torres.
• And I'm Sarah Chen. Markets closed the week with the S&P up 0.4%, the Nasdaq flat, and the ten-year yielding 4.18%. A quiet tape, b
Daily AI news for investors and financial professionals. Two expert hosts break down how artificial intelligence is reshaping markets, portfolios, and the future of finance.
Alex Torres: Welcome to Pivot Invest for Saturday, May 9th, 2026. I'm Alex Torres.
Sarah Chen: And I'm Sarah Chen. Markets closed the week with the S&P up 0.4%, the Nasdaq flat, and the ten-year yielding 4.18%. A quiet tape, but plenty happening underneath.
Alex Torres: Here's the question I keep circling back to: are we watching AI capex finally translate into earnings, or are we watching the most expensive infrastructure buildout in history search for a return?
Sarah Chen: The numbers from this week argue both sides. Hyperscaler capex guidance for 2026 now totals roughly $560 billion across Microsoft, Google, Amazon, and Meta. That's up 22% year over year.
Alex Torres: And yet enterprise AI revenue, the actual top-line return, is tracking closer to $90 billion industry-wide. The ratio still doesn't pencil out cleanly.
Sarah Chen: Which brings us to Nvidia. The stock pulled back 3% this week on reports that Anthropic and OpenAI are both accelerating custom silicon partnerships. Anthropic deepened its Trainium commitment with AWS, and OpenAI's Broadcom chip is reportedly entering production this fall.
Alex Torres: The narrative shift here is subtle but important. For two years, the story was Nvidia's monopoly. Now it's Nvidia's margin. Even if they keep the volume, the pricing power erodes when your three biggest customers are also your three biggest competitors.
Sarah Chen: Wall Street consensus still has Nvidia growing data center revenue 18% next year, but margin assumptions are quietly compressing. Gross margin estimates dropped from 74% to 71% over the past quarter.
Alex Torres: Three points of margin on $200 billion in revenue is real money. That's six billion dollars walking out the door, and it shows up in valuation multiples fast. Let's move to banking. JPMorgan rolled out its agentic AI platform to commercial clients this week.
Sarah Chen: The pitch is autonomous treasury operations, cash management, FX hedging, working capital optimization, all running with human oversight rather than human input. JPMorgan claims early pilots cut treasury operating costs by 35%.
Alex Torres: If that number holds at scale, every regional bank in America has a problem. You can't compete on a service your competitor automated 35% cheaper.
Sarah Chen: Citi and Bank of America have similar products in pilot. The interesting question for business leaders is whether your treasury team should be evaluating these tools now or waiting six months for the comparison shopping to mature.
Alex Torres: I'd argue now. The switching costs compound. Whoever onboards first locks in workflow and data integration, and that's a moat that's very hard to dislodge later.
Sarah Chen: On the regulatory front, the SEC finalized its AI disclosure rules Thursday. Public companies will need to disclose material AI dependencies, model risk exposures, and any AI-driven decision systems affecting financial reporting, starting with fiscal year 2027 filings.
Alex Torres: This is the moment AI stops being a marketing line in the 10-K and starts being a liability section. Auditors are going to have a busy 18 months.
Sarah Chen: Deloitte and PwC both announced expanded AI assurance practices this week, clearly anticipating the demand. Expect audit fees to rise 8 to 12% for AI-heavy companies.
Alex Torres: Quick hits. Tesla's robotaxi network expanded to Phoenix and Miami this week, bringing the total to seven cities. Reported utilization is around 60% during peak hours.
Sarah Chen: Revenue contribution is still rounding error, under 2% of Tesla's quarterly top line, but the unit economics are improving. Per-mile cost is down to about 31 cents.
Alex Torres: Stripe filed confidentially for an IPO. Reported target valuation is $110 billion, down from the $95 billion private mark last year, but up from the 2023 low of $50 billion.
Sarah Chen: Their AI-powered fraud product, Radar, is now generating over $1.2 billion in annual revenue on its own. That's the line item bankers are pitching hardest.
Alex Torres: In crypto, Bitcoin held above $98,000 this week despite the broader risk-off tone. The spot ETF complex saw $2.1 billion in net inflows.
Sarah Chen: And in energy, Constellation signed a 20-year nuclear PPA with Meta for 1.8 gigawatts. The AI power deals keep getting bigger. Constellation stock is up 14% year to date on this theme alone.
Alex Torres: That's the trade nobody talks about enough. The picks-and-shovels play for AI isn't chips anymore. It's electrons, and increasingly it's baseload electrons that can run 24/7.
Sarah Chen: Utility valuations reflect that. The XLU is trading at 22 times forward earnings, well above its ten-year average of 17. That's a sector that historically traded like a bond proxy now trading like a growth story.
Alex Torres: Looking ahead to next week: CPI on Tuesday, retail sales Thursday, and earnings from Walmart, Cisco, and Applied Materials.
Sarah Chen: Applied Materials will be the tell on semiconductor capex. If equipment orders are softening, that's an early signal the AI buildout is pacing itself.
Alex Torres: The takeaway for business leaders this week: the AI economy is bifurcating. Infrastructure spend keeps climbing, but the returns are starting to concentrate in companies with proprietary data and direct customer relationships.
Sarah Chen: Translation: if your AI strategy is buying tools, you're a customer. If it's building workflows on your own data, you're a competitor. The market is starting to price that distinction, and the gap will only widen from here.
Alex Torres: That's our show. Onwards.
Sarah Chen: Stay sharp.