Happy Monday. The Nvidia board investigation into Jensen Huang is the story the industry will be watching all week. It touches export controls, national security, and the question of whether the most consequential CEO in semiconductors stays in his seat.
The rest of today's issue is a stress test on AI returns. Hyperscalers are spending $730 billion on AI infrastructure this year, yet the evidence that it moves a balance sheet remains thin. Two stories dig into why the gap exists and what to do about it.
ENTERPRISE SOFTWARE
Nvidia Board Investigates Huang Over Alleged Chinese Spy Affair
Nvidia's board launched an independent investigation Sunday after a U.S. intelligence memo alleged that Alice Huang, a top executive at Singapore-based Megaspeed, was recruited by China's Ministry of State Security to gather information on CEO Jensen Huang. Jensen Huang admitted an extramarital affair in a staff memo and confirmed to the board the relationship was with Alice Huang. Megaspeed's Malaysian subsidiary reportedly purchased close to $2 billion in Nvidia chips restricted from direct export to China across more than 200 shipments between June 2024 and early 2026.
The board says it has no reason to believe national security or Nvidia secrets were compromised, and Huang will remain CEO during the inquiry. That may change. Analyst Dan Ives noted outright that Huang's removal is a real possibility depending on the investigation's findings. For enterprise software and hardware buyers, the immediate risk is regulatory: congressional pressure is already building, three Republican House chairs have requested a closed-door briefing on November 18, and any finding that tightens chip export enforcement would reshape procurement timelines and supply chain assumptions across the AI industry.
ARTIFICIAL INTELLIGENCE
AWS SVP: Measure AI by Business Metrics You Already Track
AWS SVP Colleen Aubrey told enterprise leaders to stop creating AI-specific metrics and start demanding that AI move numbers they already track: revenue per planner, forecast accuracy, inventory turns, time to resolution. Her rule is concrete: before any project gets funded, it must be tied to an existing metric with a clear baseline, and if that metric does not move within a defined window, the project does not scale. Amazon Connect, her proof point, hit a $1 billion annualized revenue run rate by starting narrow, with superior voice quality, rather than chasing feature parity with incumbents.
Aubrey's framework is a direct counterweight to the AI adoption playbook most vendors are selling. She is not arguing against AI investment. She is arguing that the measurement problem is the deployment problem: companies that cannot name the business metric a project will move have already lost control of it. Her point about bottleneck relocation is worth attention. Amazon's coding-assistant gains showed up not in code output but in test coverage and pipelines. Teams that budget only for the initial deployment often discover the next constraint after the fact, with no resources ready to address it.
ARTIFICIAL INTELLIGENCE
Hyperscalers Spend $730B on AI While Only 12% of CEOs Show Returns
Amazon, Google, Microsoft, and Meta are on track to spend roughly $730 billion on AI capital expenditure in 2026, up 78% from $410 billion the year before. The returns are not matching the investment. PwC's survey of 4,454 CEOs found that 56% reported neither cost savings nor revenue gains from AI over the past 12 months. MIT found 95% of AI pilots deliver zero measurable P&L impact. Goldman Sachs said AI's contribution to U.S. GDP growth in 2025 was essentially zero.
The aggregation paradox explains much of the disconnect. Task-level productivity gains of 10% to 70% do not survive the trip from individual worker to company financials to macroeconomic output. Each step dilutes the signal. For founders and executives making budget decisions, the more urgent number is this: 94% of companies plan to keep investing in AI even without immediate returns, per BCG. That consensus spending posture, combined with thin proof of returns, is exactly the environment where companies that can measure cost per outcome will separate from those that cannot.
Created by the robots at The SaaS Sentinel