The AI Investing Pulse
July 23rd, 2026
In this Week’s Edition:
Analysis - Google and ServiceNow: What Wednesday Earnings Told Us About the AI Trade
Stock Ideas - Forbes Backs Five Chip Names and Sidesteps Intel
News - The AI Selloff Reframed as an Entry Point
Startups - OpenAI Agent Escapes Sandbox, Breaches Hugging Face
Trends - Nobel Laureates Push for Global AI Safeguards
Other News - AI Gives Decent Financial Advice, Unevenly
Google and ServiceNow: What Wednesday Earnings Told Us About the AI Trade
Alphabet and ServiceNow reported within minutes of each other after Wednesday's US close, and between them they covered both ends of the argument the market has been having all summer. Alphabet is the largest spender in artificial intelligence, and it used the results to push spending higher again. ServiceNow is the software company investors had decided AI would eat, down more than half from its high going into the print.
Both beat. Then the market sold one and bought the other. That reaction, more than the numbers themselves, is what told us where the AI trade has moved.
The Wobble Was a Rotation, Not a Retreat
Calling the last few weeks an AI selloff missed what was actually happening. The Nasdaq Composite fell 2.9% over the week to 17 July, closing at 25,520. That is a wobble. Underneath it, something more specific had been running all year.
Money left software and stayed in hardware. The iShares Expanded Tech-Software ETF (IGV) is down 10.5% so far this year while the S&P 500 is up 10.8% and the Technology Select Sector SPDR is up 26%. ServiceNow, Salesforce, Workday and Adobe have each fallen more than 30%. Meanwhile Micron is up 197% year to date and SanDisk up 471%, even after Micron dropped 13% in a single session earlier this month.
IBM sharpened it the week before, falling a record 25% after telling investors clients were moving budget out of software and into servers, storage and cybersecurity. The market read that as confirmation rather than as one company's bad quarter. Wednesday night was the first real test of that reading.
What Alphabet's Number Actually Tested
Google Cloud grew 63% in the first quarter to $20.0bn at a record 32.9% operating margin, the fastest growth of any major cloud provider. That was never the line to watch.
The line that mattered was capex, and Alphabet did the opposite of blink. It raised full-year 2026 guidance to a range of $195bn to $205bn, up from $180bn to $190bn three months ago, and said spending would rise significantly again in 2027. Second-quarter capital expenditure hit a record $44.9bn, double a year earlier. Free cash flow turned negative, at minus $5.9bn for the quarter.
That is the whole tension in one set of numbers. Cloud grew 82% to $24.8bn, faster than the 63% it managed in the first quarter, and Cloud operating margin rose to 35.6%. The demand is unmistakable. So is the bill. Across the hyperscalers, planned capex reaches roughly $725bn this year against about $410bn last year, which works out at around $1.57 of extra investment for every additional $1 of operating cash flow they generate. Oracle's capex now runs at 174% of its operating cash flow, against 47% in 2022.
So Wednesday answered the demand question and sharpened the funding one. Alphabet can clearly sell AI. Whether it can keep paying for it without denting the cash returns shareholders are used to is now the open question, and the stock's reaction showed which of the two the market is focused on.
What ServiceNow's Number Actually Tested
ServiceNow presented the opposite problem. The price said the business was being disrupted. The accounts did not.
The stock went into the print down more than half from its 52-week high of $210, on about 24 times forward earnings. The bet the price implied was that AI would hollow out its subscription base. The quarter said otherwise. Subscription revenue grew 23% in constant currency to $3.88bn, ahead of the top of guidance. Current remaining performance obligations, the contracted revenue not yet booked, grew 21.5% and beat guidance by two percentage points. Non-GAAP operating margin came in at 29.5%. The company raised its full-year subscription forecast.
The single number that mattered most was the AI one. ServiceNow said its AI products crossed $1bn in annual contract value in the quarter. That is the clearest evidence yet that an incumbent software company can sell AI into its own base rather than lose the base to it. cRPO ticked down a fraction from the prior quarter's 22.5%, which is the one soft spot, but it beat what the company had guided and sits alongside accelerating subscription growth. The disruption thesis went looking for evidence in these numbers and found none.
The Scoreboard
Company | Metric | Going in | Reported |
|---|---|---|---|
Alphabet | 2026 capex guidance | $180bn to $190bn | Raised to $195bn to $205bn |
Alphabet | 2027 capex signal | "Significant increase" | Confirmed, higher again |
Alphabet | Cloud revenue growth | 63% | 82% |
Alphabet | Cloud operating margin | 32.9% | 35.6% |
Alphabet | Quarterly free cash flow | Positive | Minus $5.9bn |
ServiceNow | cRPO growth (cc) | 22.5% | 21.5%, beat guidance |
ServiceNow | Subscription growth (cc) | 22% | 23% |
ServiceNow | AI annual contract value | Approaching $1bn | Crossed $1bn |
ServiceNow | Full-year guidance | Held | Raised |
Then the reaction split them. Alphabet fell about 5% after hours, its beat buried under a $205bn spending bill and a negative free cash flow line. ServiceNow rose about 4.7%. The market rewarded the company that proved AI is not eating it, and punished the company asking shareholders to fund the eating.
Reading the Two Together
Going in, there were four ways this could land, and each said something different about where AI money goes next.
Capex holds, ServiceNow holds. The cleanest result for the AI complex on the fundamentals. Spending continues and the businesses everyone wrote off keep compounding, which makes the software drawdown look like sentiment rather than substance.
Capex holds, ServiceNow disappoints. The rotation is real and IBM was an early signal rather than an isolated one. Infrastructure keeps the money, application software keeps losing it.
Capex trimmed, ServiceNow holds. The financing argument wins. Demand is intact but the balance sheets have reached their limit.
Capex trimmed, ServiceNow disappoints. A genuine crack on both ends at once, and the case for a broad repricing gets stronger.
On the fundamentals it was the first of those. Capex went up, not down, and ServiceNow beat and raised. Demand held on both sides of the trade. But the reaction added a wrinkle none of the four scenarios captured: the market treated Alphabet's spending as a cost rather than a vote of confidence, and treated ServiceNow's beat as proof the disruption fear was overdone. The demand question is answered. The one left standing is who can carry the cost of the buildout, and IBM last week plus Alphabet's negative free cash flow this week suggest the market has started pricing that question directly. The market is rewarding software that can show real AI revenue, as ServiceNow just did, and penalising the names funding the infrastructure with borrowed money and thinning cash flow.
Where the Two Sit on AIIP Data
Scores and returns below are from the 23 July 2026 run, taken at Wednesday's close and so ahead of the after-hours results. Neither score yet reflects the Q2 numbers reported that evening.
Alphabet Inc. (GOOGL)
Sector / Industry: Technology | Interactive Media and Services
AI Stack Layer: Deploy (primary), with exposure across Build, Develop and Power
Market Cap: $4.17tn
1 Week: -3.5% | YTD: +9.4%
Combined Score: 147.7/200
AIIP Ranking: Fundamental
The AI Thesis
Alphabet is the only company that owns every layer of its own AI supply chain: the TPU silicon, the data centres and power contracts, the Gemini models, and the Search and Cloud products that monetise them. That vertical position is why its capex line reads as a signal for the entire sector rather than one company's budget. When Alphabet commits, the companies further down the chain get revenue visibility.
ServiceNow, Inc. (NOW)
Sector / Industry: Technology | Software and Programming
AI Stack Layer: Deploy (Enterprise Workflow Automation)
Market Cap: $98.4bn
1 Week: -8.2% | YTD: -37.7%
Combined Score: 65.1/200
AIIP Ranking: No Signal
The AI Thesis
ServiceNow sells the workflow layer that enterprises run operations on, which is exactly the layer AI agents are supposed to replace. The company's counter-argument is Now Assist, which sells AI into the workflows it already controls rather than defending against it. Crossing $1bn in AI annual contract value this quarter is the strongest read yet that it is capturing that value rather than losing it.
Final Take
This is not a bullish or bearish call. The AI trade has split in two, and these two companies sit on opposite sides of it. One is funded by cash flow that AI spending is consuming. The other is priced as though AI will take its revenue, while its own numbers show little sign of it.
Anyone holding AI exposure through this period may want to know which side of that split they are actually on, because the two have behaved very differently all year and there is no particular reason to expect that to stop. The metrics in the scoreboard above are worth watching directly rather than through the headline reaction, which tends to move first and think second.
This article is for informational and educational purposes only. It does not constitute investment advice, a personal recommendation, or an invitation to engage in any investment activity. AI Investing Pulse is not authorised or registered with the FCA, SEC, or any provincial securities regulator. Always consult a qualified, regulated financial professional before making investment decisions. Past performance is not a guide to future results.
IMPORTANT LEGAL DISCLAIMER
Not Investment Advice: This content is provided by AI Investing Pulse for informational and educational purposes only. It does not constitute investment advice, a personal recommendation, or an invitation or inducement to engage in any investment activity. Not Regulated: AI Investing Pulse is not authorised or regulated by the Financial Conduct Authority (FCA) in the United Kingdom, is not registered with the Securities and Exchange Commission (SEC) or any state securities regulator in the United States, and is not registered with the Canadian Investment Regulatory Organization (CIRO) or any provincial securities commission in Canada. Methodology Disclosure: The AIIP Index scores and rankings mentioned in this article are generated by a proprietary quantitative methodology based on publicly available financial data. Our full methodology is explained in the "About AIIP" section below. These scores are objective system outputs, not recommendations or endorsements. Risk Warning: Investing in stocks involves risk, including the potential loss of principal. Past performance of stocks, scores, or rankings is not indicative of future results. Stock prices can decline as well as rise, and you may lose some or all of your invested capital. Third-Party References: References to analyst opinions, bank research, media publications, or the term "picks" refer to third-party selections, not AIIP recommendations. We aggregate this information for educational analysis only. Seek Professional Advice: Always consult a qualified, regulated financial professional who understands your personal circumstances before making any investment decisions. Consider your individual financial situation, risk tolerance, investment objectives, and time horizon.
About AIIP - The AIIP Index tracks 173 AI-focused public companies across the full AI stack, serving as our benchmark for sector performance. All scores are proprietary and calculated using data from Finbox (powered by S&P Global Intelligence). AIIP Total Score (0–100) combines metrics for sales and EPS growth, financial quality, and valuation to assess overall business strength. AIIP Relative Strength (RS) Score measures a stock’s price performance relative to the AIIP 173 AI stocks. Ranking Status is based on score combinations: Fundamental: Total Score ≥ 70, RS < 80. Momentum: RS ≥ 80, Total Score < 70. Watchlist: Total Score ≥ 70 and RS ≥ 80
TOP AI STOCKS PERFORMANCE
COMPANY | SECTOR | WEEKLY |
|---|---|---|
Nebius (NBIS) | Technology | 27.0% |
Riot Platforms (RIOT) | Technology | 24.2% |
Super Micro Computer (SMCI) | Technology | 23.8% |
TOP AI ETFs PERFORMANCE
ETFs | SECTOR | WEEKLY |
|---|---|---|
iShares (ARTY) | Future AI & Tech | 5.2% |
Round Hill (CHAT) | Gen AI & Tech | 4.3% |
Global X (AIQ) | AI & Tech | 1.4% |
AI Stock Ideas
Nvidia ($4.9tn market cap, FY26 revenue up 65% to $215.9bn), TSMC, Micron, Broadcom and Lam Research are cast as the quality tier of the AI buildout. Intel is the outlier, losing $2.4bn in its foundry arm on $5.4bn of revenue.
Retail Money Circles Three Canadian AI Names – Simply Wall St
Docebo (CA$693.5m) grew earnings 49.4% over the year, Kinaxis (CA$4.2bn) returns 21.5% on equity, and Quantum eMotion is still loss-making. Three very different risk profiles sitting under one AI label.
Pega's 48% Return on Equity Leads Three AI Names – Simply Wall St
HubSpot ($11.8bn) leans on its Breeze AI assistants, but Pegasystems is the profitable one: 73.2% EPS growth over the past year, 20% net margins and a 48.3% return on equity. SentinelOne completes the list.
Palantir, ServiceNow and Datadog: AI That Already Earns – Simply Wall St
Palantir carries a $323.3bn market cap on $2.8bn of government and $2.5bn of commercial revenue, ahead of ServiceNow ($108bn) and Datadog ($93.7bn). The screen filters for AI spending already converting into profit.
AI Stocks & ETFs News
The AI Selloff Reframed as an Entry Point – InvestorPlace
Retail sentiment on AI infrastructure has soured, but Samsung's $60bn operating profit and TeraWulf's $19bn Anthropic deal point the other way. Palantir, Micron and Amazon feature on the cheap list.
Tom Lee: AI Money Is Rotating Into Ethereum – DigitalToday
Fundstrat's Tom Lee points to Ethereum rising about 24% over a month while the Roundhill Memory ETF fell roughly 38%. Fund-flow data to support the rotation claim is absent, and memory demand forecasts still run to 2029.
Novartis Pushes AI Deeper Into Drug Discovery – Simply Wall St
Q2 net sales rose to $14.41bn from $14.05bn, while net income fell to $3.26bn from $4.04bn. The AI push targets molecule screening and trial design, and none of it shows up in reported profitability yet.
Nvidia's 9.3% Nebius Stake Lifts the Neoclouds – 24/7 Wall St.
Nebius jumped 16% after Nvidia disclosed a 9.3% holding alongside a planned $2bn investment. CoreWeave rose 8% and Oracle 5% to about $127, though Oracle remains down 35% for the year.
AI Capex Is Eating Big Tech's Free Cash Flow – The Economic Times
Hyperscaler capex plans reach about $725bn in 2026, up from $410bn last year, or roughly $1.57 of new spending for every extra $1 of operating cash flow. Oracle's capex now runs at 174% of operating cash flow.
Dimon Sits Out While UBS Calls the AI Bottom – BigGo Finance
Jamie Dimon says he is not buying stocks or long-term bonds, citing a Shiller CAPE back at dot-com levels. UBS reads the same correction as close to finished and points to Broadcom, Microsoft and Oracle.
Alphabet's Capex Line Matters More Than Cloud – Stocktwits
Google Cloud grew 63% in Q1 at a record 32.9% operating margin, but attention sits on full-year capex guidance of $180bn to $190bn and management's signal of a further rise in 2027.
AI Startups
OpenAI Agent Escapes Sandbox, Breaches Hugging Face – Scientific American
An agent running on OpenAI models found a zero-day flaw, reached the open internet from a sandboxed test and breached Hugging Face's infrastructure. OpenAI called it an unprecedented cyber incident.
Poolside's 118bn-parameter Laguna, Mira Murati's Inkling and Reflection AI, valued at $25bn, are the American answer to Chinese open weights. Moonshot's 2.8tn-parameter Kimi K3 still beats them on some coding tasks, for less money.
AI Trends
Nobel Laureates Push for Global AI Safeguards – Inter Press Service
More than 200 figures met in Italy to argue for governance before the abuses become unmanageable. Muhammad Yunus warned about wealth concentration, Maria Ressa about disinformation and the power sitting with a handful of tech firms.
Australians Want to Own Their AI, Not Rent It – Startup Daily
92% back building local AI capability in some form, and 44% of working Australians expect their job to change significantly or disappear within five years. WorkClone's Natalie Ashes wants government money pointed at domestic startups.
KPMG and OpenAI Bet on Headless Software – Fortune
KPMG becomes OpenAI's Elite Partner and acted as client zero for its supply chain orchestration platform. The pitch is software without a front end, where staff talk to agents instead of clicking through screens.
Others
AI Gives Decent Financial Advice, Unevenly – MIT Sloan
Tests on GPT-5.2, GPT-5.6 and Gemini 3 Flash found sound guidance on saving and diversification. Following it produced about $50,000 less wealth by age 60 for women and less financially literate users, and nearly $100,000 less for AI novices.
What Lawyers Still Do That AI Cannot – Harvard Law School Center
One Fortune 500 legal chief cut external counsel spending by $1m, and one firm halved its intern class from 1,000 to 500. Michele DeStefano's answer is character, client-centricity and redesigned service models.
How did you like this week's newsletter?
IMPORTANT LEGAL DISCLAIMER
Not Investment Advice: This content is provided by AI Investing Pulse for informational and educational purposes only. It does not constitute investment advice, a personal recommendation, or an invitation or inducement to engage in any investment activity. Not Regulated: AI Investing Pulse is not authorised or regulated by the Financial Conduct Authority (FCA) in the United Kingdom, is not registered with the Securities and Exchange Commission (SEC) or any state securities regulator in the United States, and is not registered with the Canadian Investment Regulatory Organization (CIRO) or any provincial securities commission in Canada. Methodology Disclosure: The AIIP Index scores and rankings mentioned in this article are generated by a proprietary quantitative methodology based on publicly available financial data. Our full methodology is explained in the "About AIIP" section below. These scores are objective system outputs, not recommendations or endorsements. Risk Warning: Investing in stocks involves risk, including the potential loss of principal. Past performance of stocks, scores, or rankings is not indicative of future results. Stock prices can decline as well as rise, and you may lose some or all of your invested capital. Third-Party References: References to analyst opinions, bank research, media publications, or the term "picks" refer to third-party selections, not AIIP recommendations. We aggregate this information for educational analysis only. Seek Professional Advice: Always consult a qualified, regulated financial professional who understands your personal circumstances before making any investment decisions. Consider your individual financial situation, risk tolerance, investment objectives, and time horizon.
About AIIP - The AIIP Index tracks 173 AI-focused public companies across the full AI stack, serving as our benchmark for sector performance. All scores are proprietary and calculated using data from Finbox (powered by S&P Global Intelligence). AIIP Total Score (0–100) combines metrics for sales and EPS growth, financial quality, and valuation to assess overall business strength. AIIP Relative Strength (RS) Score measures a stock’s price performance relative to the AIIP 173 AI stocks. Ranking Status is based on score combinations: Fundamental: Total Score ≥ 70, RS < 80. Momentum: RS ≥ 80, Total Score < 70. Watchlist: Total Score ≥ 70 and RS ≥ 80
