On Friday 30 October, US tech stocks had a brutal session. The Nasdaq fell 2.9%, its largest one-day percentage drop since 5 June, and Intel lost 28.6% after disappointing quarterly results. Reuters’ report on the Nasdaq selloff after Intel and Amazon results links the rout to growing worries about AI-related companies’ massive spending on the technology. For anyone running or funding AI projects, the takeaway is simple: “we’ll find the return later” just got a lot harder to say out loud.
What happened
Intel and Amazon both reported late on Thursday, and both disappointed. Amazon pared its losses as Friday’s session went on. Intel did not recover, closing down 28.6%. The damage spread across the hardware stack. Nvidia fell 4.2%, AMD dropped 10.9%, and the broader chips index lost 7.2%, which left it down more than 15% for the month. Server makers were hit too, with Dell down 10.8% and HP down 9.6%.
The Nasdaq Composite closed at 25,437.00, down 747.73 points on the day. It now sits 7.8% below the all-time closing high it set on 6 October. The S&P 500 fell 1.82% and the Dow 1.09%. Over the month, the S&P 500 tech sector had its largest percentage loss since March 2025, and communication services had its biggest monthly drop since April 2022. The Dow, which leans less on tech, still finished October with a gain.
The bigger picture is not a collapse in earnings. According to LSEG data cited by Reuters, 82% of the 299 S&P 500 companies that had reported beat consensus expectations, and analysts expect third-quarter earnings growth of 31.6% year on year. The market isn’t selling because companies are failing. It is selling the ones whose AI spending hasn’t yet earned its keep.
Oliver Pursche of Wealthspire Advisors put it plainly to Reuters: the week was about “whether there will be a positive return on investment and if so, how quickly.” He added that companies falling short of expectations “are getting punished.”
What it means
For the past couple of years there has been an unwritten deal. Spend heavily on AI, call it ambition, and the market will reward you. This selloff is that deal breaking in public. Investors still reward companies that deliver. They have stopped giving credit for spending on its own.
That shift won’t stay on Wall Street. CFOs read the same headlines as everyone else, and a sharp monthly drop in the tech-heavy index is exactly what gets raised at board level when next year’s budget is being set. Enterprise procurement doesn’t happen in a vacuum. When the companies selling chips, servers and cloud capacity are being questioned on returns, the companies buying AI tools will face the same questions about their own projects.
Our view is that this is healthy. A lot of AI work over the last two years has been approved on narrative: a roadmap slide, an “adoption curve” position, a fear of being left behind. That kind of spending was always going to get a hard look eventually. October 30 is a date people will point to when that review starts.
What it means for builders
If you build or commission AI systems, expect the opening question in any budget meeting to change. It was “what could this do?” Now it is “what does this save or generate, and when does it pay back?” Projects with a clear answer will get funded. Projects without one will be paused, shrunk or folded into something that has one.
That favours narrow, measurable work over broad platform bets. An agent that clears a specific back-office queue, cuts the handling time on a known document type, or removes a manual reconciliation step is easy to defend, because the baseline already exists. A general “AI layer” for the business, with value to be found later, is now a much harder sell.
What to do now
Don’t wait for finance to ask. Put a number on every live AI initiative this month, even a rough one. Our guide to calculating process automation ROI properly goes through the method. The short version: measure the current cost of the process, estimate the realistic change, and include build, run and maintenance costs, not just licence fees. If you’re still scoping, our breakdown of what it costs to build an AI agent will help you avoid underestimating the second half of that equation.
Then sort your portfolio honestly. Put projects with a measured baseline and a payback date in the first column. Put projects that are promising but unmeasured in the second, and give each one a short deadline to produce a number. Anything left over is a candidate to stop. Run that triage yourself and you go into budget season with a defensible plan. Leave it, and the plan gets written for you.
The era of unlimited AI spending is over. Accountable AI spending starts now. The useful question for every team is whether the thing they’re building can be put into a number.
Frequently Asked Questions
How far did the Nasdaq fall on 30 October?
According to Reuters, the Nasdaq Composite fell 747.73 points, or 2.86%, to 25,437.00. That was its largest one-day percentage drop since 5 June, and it left the index 7.8% below its all-time closing high from 6 October.
Why did Intel shares drop 28.6%?
Intel released disappointing quarterly results late on Thursday 29 October. Reuters reported that those results, along with Amazon's, amplified growing worries about AI-related companies' heavy spending on the technology.
Does the selloff mean companies should stop investing in AI?
No. It means AI investment now needs a measurable return attached to it. Projects with a clear baseline, a quantified saving or revenue gain, and a payback date are much easier to defend than broad bets with value to be found later.
Want this working in your business?
D3X Solutions LABS builds AI agents and automation that pay back fast. Set up a free 30-minute call and we will map your first win together.
Get in touch
