Alphabet Loses About $500 Billion in Value as AI Mode Squeezes Search Advertising

Alphabet, the parent company of Google, lost roughly $500 billion in market value on Thursday after investors concluded that the artificial intelligence tools meant to protect its search business are now squeezing the advertising revenue that pays for them. According to The Wall Street Journal, Google shares fell by about 13 per cent during the day, the stock’s worst day on record, erasing roughly $500 billion from Alphabet’s value and dragging down the wider United States stock market. According to The Times, the shares closed down $39.54, or 12.3 per cent, at $282.35 in New York, the worst single-day fall in data going back to August 2004.

The sell-off followed results published on Wednesday 28 October, after the American market had closed, for the three months to the end of September 2026. According to Reuters, Alphabet missed Wall Street expectations for both profit and revenue, and its shares had already fallen by more than 10 per cent in extended trading on Wednesday evening. The disappointment was not that Google’s business had stopped growing. Revenue rose by a quarter. It was that the growth was coming in the wrong places, at the wrong cost, and with a warning attached: the new way people search earns Google less money per question than the old way did.

Alphabet reported total revenue of $127.6 billion for the third quarter, up 25 per cent from $102.3 billion a year earlier, according to Reuters, citing data compiled by LSEG. That was short of the $128.2 billion analysts had expected. Revenue excluding payments to distribution partners was $111.2 billion, also below forecasts. Net income available to common shareholders fell 25 per cent to $26.17 billion, and diluted earnings per share fell to $2.12 from $2.87 a year earlier. On an adjusted basis, earnings per share of $2.67 missed the $3.02 target Wall Street had set, according to Reuters.

The headline profit figure needs careful reading, because the operating business and the investment portfolio pulled in opposite directions. Operating income rose 30 per cent to $40.6 billion, according to the earnings release as summarised in coverage of the filing. The fall in net income was driven largely by what Alphabet reports as other income and expense, which swung from a gain a year earlier to a net expense of $8.2 billion. Alphabet reported an $8.7 billion net loss on equity securities, which it said reduced net income by $6.7 billion after tax, or 55 cents a share. Reporting on the earnings call said the loss came primarily from an unrealised loss on Alphabet’s investment in SpaceX after a fall in the value of that stake. An unrealised loss of that kind is a paper revaluation, not cash leaving the company, but it still counts in reported earnings, and it made a quarter of strong operating growth look, on the bottom line, like a quarter in retreat.

The part of the results that unsettled investors most was Google Search. Search and other advertising revenue rose 12 per cent to $63.4 billion, according to The Times, below Wall Street expectations and the slowest rate of growth in five quarters. The Wall Street Journal reported that the figure was down from 17 per cent growth in the second quarter and 19 per cent in the first quarter. Search accounts for about half of Alphabet’s revenue, so even a modest deceleration there carries outsized weight. Google Services, the division that includes search advertising, YouTube and subscriptions, generated $96.3 billion, missing Wall Street expectations of about $98.5 billion, according to The Wall Street Journal. Google’s total advertising revenue was $82.01 billion, about $1.41 billion short of expectations, according to Reuters. YouTube advertising revenue rose 9 per cent to $11.2 billion.

On the earnings call, Google’s chief business officer, Philipp Schindler, gave investors the explanation they had feared. According to The Wall Street Journal, Schindler said the strong growth in queries made through AI Mode, Google’s conversational search format, was having an impact because those queries were monetised at a lower rate than traditional search. The Times reported the same point in his words: “the strong growth in AI mode queries had an impact, because of the lower monetisation we were seeing in AI mode”. At the same time, Schindler said overall query growth was running at its fastest rate in five years. People are asking Google more questions than ever. The problem, as investors read it, is that a growing share of those questions now arrive in a form that produces fewer paid clicks.

Schindler did not attribute the whole slowdown to AI Mode. According to The Wall Street Journal, he also pointed to changing foreign exchange rates, reduced revenue from Asian retailers affected by European Union regulatory changes, and a shift of advertising spending into the first half of the year driven by the World Cup. Reporting that set out his remarks in order put foreign exchange first, followed by the Asian retailer effect and the World Cup timing, with lower monetisation in AI Mode queries last among the four factors. That ordering matters, because it means the search slowdown was not a single-cause event. But markets fasten on the cause that threatens the future, and the AI explanation was the one that could not be dismissed as a one-off. Exchange rates move both ways. A structural change in how search is monetised does not.

Google’s answer to the monetisation gap is to put more advertising into AI Mode itself. According to The Wall Street Journal, Schindler said the company is bringing more advertisements to AI Mode responses and beginning to roll them out globally, and that early results suggest the formats are working. Click-through rates for sponsored product listings in AI Mode are already comparable to those for organic listings, the company said, and its user research indicates that people find the advertisements relevant. The Times reported that Schindler described the tests as returning “promising results”, and that the company would scale advertisements in AI Mode to a number of countries, including the United Kingdom, in the fourth quarter. “To be clear, we feel very good about our trajectory here,” he said, according to The Times. Investors, for one day at least, did not share the confidence.

The other side of the quarter was the cost of building the infrastructure on which all of this depends. Alphabet spent $51.4 billion on capital expenditure in the three months, according to The Times, pushing free cash flow into negative territory for only the second quarter since the company went public in 2004. Free cash flow was negative $9.07 billion, according to Reuters, compared with analyst expectations that it would come close to breaking even, and worse than the negative $5.9 billion recorded in the previous quarter. Reuters reported that the previous quarter was the first time in Alphabet’s 22-year history as a public company that free cash flow had turned negative. The cash burn was driven by spending to expand cloud capacity and fund artificial intelligence research, including a $7.66 billion operating loss in Alphabet-level activities that the company said primarily reflected expenses related to artificial intelligence research and development, according to Reuters.

Chief Financial Officer Anat Ashkenazi told analysts that capital expenditure for the full year 2026 was still expected to reach $195 billion to $205 billion, according to coverage of the call, which implies spending of as much as about $73 billion in the final three months of the year alone. She also said capital expenditure would increase significantly in 2027, according to The Times. The commitment does not stop at the quarterly spending figure. On the earnings call, Ashkenazi disclosed more than $1 trillion in purchase commitments, future leases and other obligations, according to Bloomberg’s reporting of the call and the subsequent quarterly filing. The filing itself put purchase commitments not recognised on the balance sheet at $924.4 billion, with most contracts to be fulfilled through 2031, and reported a further $130.5 billion in leases that had not yet commenced, largely covering data centres and energy infrastructure. It also listed $98.5 billion in contingent commitments, up about $23 billion from the previous quarter, covering items such as guarantees of data centre lease payments, support for partners buying power equipment, and a possible requirement to repurchase artificial intelligence chip systems from customers if certain conditions apply, according to Bloomberg. Alphabet stressed that the headline total reflected maximum contractual obligations, with variable pricing and volume flexibility that could bring the eventual spending lower. The scale of the commitment, even with that caveat, is the point: Google is contracting for power, chips and buildings years ahead, on terms that are difficult to unwind if demand disappoints.

Against that spending, Google Cloud was the clear bright spot in the results. According to Reuters, cloud revenue exceeded estimates and more than doubled, to $30.71 billion in the quarter ended 30 September. The Times reported growth of 103 per cent year on year to $30.7 billion, with cloud operating income rising sharply. Cloud chief Thomas Kurian made his debut on an earnings call, according to Reuters, a sign of how central the enterprise business has become inside a company long dominated by consumer search. Kurian said sales of Google’s tensor processing units, the custom chips that compete with Nvidia’s graphics processors, had become a meaningful contributor to the cloud business. Google only began selling access to its tensor processing units in the second quarter of this year, according to The Times, and they are used by customers including Anthropic and OpenAI. Google Cloud’s backlog of contracts awaiting fulfilment rose to $560.8 billion at the end of September, according to Bloomberg’s reporting of the filing, of which $16.8 billion was linked to a multiyear agreement with Crux AI, a venture Alphabet formed with Blackstone. Bloomberg reported that more than a third of the quarter’s growth in the backlog came from that joint venture, a detail that tempered some of the enthusiasm for the headline backlog figure, since part of the demand was coming from a venture Alphabet itself had helped to create.

The results arrived on a day when Google had tried to set the terms of debate itself. Earlier on Wednesday, the company announced Gemini 4 Carbon, a new frontier artificial intelligence model it described as its strongest yet, to be made available first to a subset of cyber defenders. According to Reuters, chief executive Sundar Pichai told analysts on the earnings call that the model meant Google was again “leading at the frontier”, and that the announcement was intended in part to counter a perception that the company had fallen behind in the artificial intelligence race. The model announcement and the earnings release therefore told two halves of the same story: Google believes it can build frontier models, and it is spending at an unprecedented rate to prove it, while the business that has historically paid for everything is showing the strain of the transition.

The market reaction spread beyond Alphabet. According to Reuters, the Nasdaq Composite fell by more than 1 per cent on Thursday as Alphabet’s results rekindled worries about ballooning artificial intelligence costs, outweighing strength in financial and healthcare shares. Alphabet was on track at one stage to lose more than $464 billion in market value, accounting for most of the more than $500 billion the Nasdaq-100 index was set to erase if intraday losses held, according to Reuters’ intraday reporting. Microsoft fell 1.5 per cent following reports that it was planning to raise at least $25 billion through a bond sale, even though the company had delivered a strong cash-flow result and forecast robust cloud growth the previous day, according to Reuters. Amazon fell 2.9 per cent before its own quarterly results after the market closed, while Meta and Apple rose. The Philadelphia semiconductor index fell 0.9 per cent, ending a four-day winning run, according to Reuters. The pattern was not a rout in technology as a whole. It was a repricing of one specific bargain: that the companies spending most heavily on artificial intelligence would keep generating the cash to pay for it without interruption.

Analyst reaction split between those who saw a warning and those who saw a buying opportunity delayed. Mike O’Rourke, chief market strategist at JonesTrading, told The Wall Street Journal: “One of the companies most aggressively investing in AI to protect its core businesses showed vulnerability to AI. That’s why people are disappointed.” Tracy Woo, principal analyst at Forrester, told The Times: “The softer search performance is significant because it highlights the challenge of funding massive AI investments while maintaining the strength of established businesses.” Analysts at Barclays said the results “gave us a bit of PTSD and harks back to the air pocket experienced pre-pandemic”, when revenue growth slowed amid concern that the advertising market was reaching saturation, according to The Times. They argued, however, that Alphabet remained in a fairly strong position if it could find a way to monetise its artificial intelligence queries, calling the episode an investor “code yellow”, but “not some kind of AI roadkill situation”. Analysts at Wedbush Securities said Google had “a lot to prove now”, but added: “we remain confident that its complete ecosystem will prove a winner as AI advertising evolves”, according to The Times. NewStreet Research said Alphabet would face ongoing headwinds until it could improve “faith in search fundamentals and AI monetization”, which the firm estimated could take until the first half of 2027, according to The Wall Street Journal. JPMorgan analysts said the reaction showed “the margin for error” for a company spending hundreds of billions of dollars on artificial intelligence infrastructure, according to Bloomberg’s reporting.

The fall was severe enough to leave a technical mark as well. According to The Wall Street Journal, the stock completed a so-called death cross on Thursday, when its 50-day moving average dropped below its 200-day moving average, a pattern some analysts read as a signal that the price has further to fall. Technical signals of that kind do not determine a company’s prospects, and they are as often a description of what has already happened as a prediction of what happens next. But in a week when investors were already asking whether the artificial intelligence trade had run ahead of its returns, the chart gave the anxiety a visible form.

The competitive context sharpens the question Google now faces. The Wall Street Journal reported that rival artificial intelligence companies are presenting their own chatbots as better ways to find answers, and taking aim directly at Google’s advertising business. Andrew Boone, an equity research analyst at Citizens, said increased use of enterprise artificial intelligence services from OpenAI and Anthropic could be leading some of Google’s highest-spending consumers to switch to artificial intelligence chatbots for their personal searches. OpenAI, which began testing advertisements with some United States-based ChatGPT users in February, said this summer that it was on track to reach $1 billion in annualised advertising revenue, according to The Wall Street Journal, and has said that more than 100,000 advertisers have run campaigns on ChatGPT, with advertisements now available in 89 countries. Advertising buyers told the newspaper that OpenAI’s advertising business was still in its early stages, with most advertisers funding their spending from experimental budgets rather than shifting money out of search. The threat to Google is therefore not yet a measurable loss of advertising revenue to a single rival. It is the possibility that the habit of searching itself is changing faster than Google can reprice it.

That is the deeper background to Thursday’s fall. Google Search has generated hundreds of billions of dollars a year by showing a list of links and charging advertisers each time a user clicks on a sponsored one, as The Wall Street Journal set out in its account of the results. Artificial intelligence search can bypass that process and give a single direct answer, reducing the need for the lucrative links in between. Google has been moving carefully to take advantage of demand for chatbot-style answers without damaging its biggest business. The third-quarter results suggest that care has not been enough to prevent the trade-off from showing up in the numbers. More questions, answered more directly, are worth less to Google per question than the old list of links was. The company’s task is to invent, inside the answer itself, an advertising format as effective as the one it is replacing, and to do so before users, advertisers and investors lose patience with the experiment.

Analysis: Why It Matters

1. The cannibalisation question is no longer theoretical

For two years, the central question about Google and artificial intelligence has been whether the company would disrupt itself before a rival did, and whether it could do so without destroying the economics of search. Thursday’s results move that question from the realm of strategy presentations into the profit-and-loss account. Schindler’s admission was carefully worded and surrounded by other explanations, and it would be wrong to read it as a confession that AI Mode alone caused the search slowdown. Foreign exchange, regulatory change in Europe and the timing of World Cup advertising all played a part, on Google’s own account. But none of those factors explains why investors reacted with Alphabet’s largest one-day loss of value rather than a shrug. What changed was the confirmation, from Google itself, that conversational queries monetise at a lower rate, at precisely the moment when those queries are growing fastest. A company can manage a cyclical advertising slowdown. It is harder to manage a change in the unit economics of its core product, because every additional user makes the problem larger rather than smaller. The original analysis here is straightforward: Google’s defence has always been that it would rather cannibalise its own business than let someone else do it. That remains the right instinct. But self-cannibalisation still has to be paid for, and Thursday showed the market demanding evidence, not just intent, that the replacement business can carry the same weight.

2. Growth is no longer enough when cash flow turns negative

The most striking feature of the quarter is the combination it produced: revenue up 25 per cent, operating income up 30 per cent, and free cash flow of negative $9.07 billion. In an earlier era, a company growing at that pace would have been forgiven almost any level of investment. The market’s patience has shortened because the investment is no longer a choice made from surplus cash. Alphabet is now spending faster than the business generates cash, for the second quarter in a row, and has contracted to spend more than $1 trillion in future commitments. That transforms the debate from “is Google growing?” to “can Google fund its own ambitions?” The answer is plainly yes in the narrow sense: operating income alone was $40.6 billion in the quarter. The harder question is about discipline and sequencing. Capital expenditure of $51.4 billion in a single quarter, rising significantly again in 2027, on top of $98.5 billion in contingent commitments, means that Google’s balance sheet is being committed years in advance to a bet whose returns are still being invented. Cloud growth of 103 per cent suggests the bet is not fanciful. But cloud revenue of $30.7 billion, however fast it grows, cannot yet carry a $200 billion annual capital programme on its own. Search still has to pay. That is why a 12 per cent growth rate in search, respectable in isolation, felt like a crisis: it is the funding source for everything else, and it is decelerating while the bill accelerates.

3. The $1 trillion commitment changes what “optionality” means

Large technology companies have long argued that scale gives them optionality: the ability to try many things, abandon the failures, and double down on the successes. A commitment of more than $1 trillion in purchase obligations, leases and other obligations points in the opposite direction. It is a decision to narrow the future in advance, locking in power contracts, data centre space and chip supply, mostly through 2031, with some leases running far longer, on terms that include guarantees to partners and possible repurchase obligations set out separately as contingent commitments. Alphabet’s caveat matters and should not be airbrushed out: the headline figure is a maximum, with variable pricing and volume flexibility, and the eventual spend could be lower. Even so, the direction of travel is unmistakable. Google is trading flexibility for certainty of supply in a market where power and advanced chips are scarce. That may prove to be the winning move if demand for artificial intelligence computing keeps doubling. It may prove to be a burden if demand grows more slowly, or if a cheaper way of delivering the same capability emerges. The original point to grasp is that the risk has moved. It is no longer mainly the risk of falling behind a rival model. It is the risk of having bought, in advance and at scale, the infrastructure for a future that arrives later, or differently, than expected. Investors who once worried that Google was spending too little on artificial intelligence now have a new worry to weigh alongside it: that Google has spent so much, so early, that it has reduced its own room to change course.

4. Cloud is now the counter-argument, and it has to keep proving itself

Every bearish reading of Thursday’s results has a bullish reply, and it is spelled “Cloud”. A business that more than doubles in a year, with operating income rising sharply and a backlog above $560 billion, is not a side project. It is evidence that Google’s artificial intelligence spending is producing something customers will pay for, in the form of computing capacity, custom chips and enterprise services. Kurian’s debut on the earnings call underlined the point: the cloud business now merits a voice of its own. But the cloud counter-argument has two weaknesses that deserve equal attention. The first is the composition of the backlog. If, as Bloomberg reported, more than a third of the quarter’s backlog growth came from a joint venture Alphabet helped to form, then part of the demand signal is, in effect, Google selling computing power to an entity it co-created. That does not make the contracts worthless, but it does mean the backlog should be read with more care than the headline suggests. The second weakness is timing. Backlog is a promise of future revenue, while capital expenditure is cash leaving the company now. Alphabet’s own disclosure that it expects to recognise just over half of the backlog as revenue over the next 24 months, as reported in coverage of the call, still leaves a long bridge between spending and return. Cloud must therefore do more than grow quickly. It must grow quickly enough, and convert backlog into cash quickly enough, to close a gap that is widening every quarter. Thursday’s market verdict was, in effect, a judgement that the bridge is longer than investors had assumed.

5. This is a test of the whole AI spending boom, not just of Google

Alphabet’s fall dragged down the Nasdaq and unsettled shares across the technology sector because Google is the purest test of a proposition on which the entire artificial intelligence investment boom rests: that enormous spending on data centres, chips and power will be repaid by equally enormous revenues, soon enough to justify the outlay. Microsoft’s fall on the same day, despite a strong cash-flow result, and the semiconductor index’s retreat, suggest that investors are applying the same test to the whole ecosystem, not just to one company. The “manic depressive sentiment swings” described by Trevor Slaven of Barings, quoted by Reuters, capture the mood: enthusiasm for the potential profits of the artificial intelligence build-out, alternating with alarm at the leverage and debt financing that increasingly underpin it. Google matters disproportionately in that debate because it is the company with the most to lose from artificial intelligence and the most to gain. It owns the search business being disrupted, the cloud business selling the picks and shovels, and the frontier models competing with OpenAI and Anthropic. If Google cannot make the economics work, the case for the rest of the sector becomes harder to make. If it can, Thursday will be remembered as a buying opportunity that looked, for a day, like a verdict. The honest assessment is that neither conclusion is yet proven. The results proved that the transition has a cost, and that the cost is larger and arriving sooner than the market had priced. They did not prove that the transition will fail. The distinction between a costly transition and a failed one is the whole investment question now, and it will be settled not by one quarter’s share price, but by whether AI Mode advertising, cloud conversion and chip sales together produce, over the next four to six quarters, the cash flow that this quarter’s spending assumed.

What to Watch Next

1. The fourth-quarter capital bill. Alphabet’s full-year capital expenditure guidance implies spending of as much as about $73 billion in the final quarter of 2026. Watch whether that figure is delivered in full, trimmed, or pushed into 2027, and whether free cash flow stays negative for a third consecutive quarter. A third negative quarter would move the cash-flow story from an investment phase to a pattern.
2. AI Mode advertising in the fourth quarter. Google said it would scale advertisements in AI Mode to more countries, including the United Kingdom. Watch for any disclosure of advertising revenue, click-through performance or advertiser numbers from AI Mode specifically, and for whether search revenue growth stabilises, recovers, or slows further in the next quarter.
3. Cloud backlog conversion. The $560.8 billion backlog is only useful if it becomes revenue and then cash. Watch the next cloud revenue figure, the operating margin on that revenue, and any further disclosure about how much of the backlog comes from related ventures such as Crux AI versus independent enterprise customers.
4. The next hyperscaler results. Microsoft, Amazon and Meta report in the same earnings season and are making comparable artificial intelligence investments. Watch whether their results show the same pattern of strong cloud growth alongside cash-flow pressure, or whether Alphabet’s search-specific problem proves to be company-specific rather than sector-wide.
5. Regulatory and competitive pressure on search. Watch for further developments in the European Union’s treatment of online advertising and retailer exemptions, for OpenAI’s expansion of ChatGPT advertising into more countries, and for any sign that advertisers are moving budgets from search to chatbot advertising rather than merely testing the new format with experimental funds.

Sources

• Reuters, via The Hindu BusinessLine: “Alphabet misses earnings estimates as AI spending drives cash burn”
• The Wall Street Journal: “Alphabet Sheds Half a Trillion in Market Value in Stock’s Worst Day Ever”
• The Times: “Alphabet suffers $500bn wipe-out in AI search sell-off”
• PYMNTS, reporting Bloomberg: “Google’s AI Exposure Exceeds the $1 Trillion It Reported”
• Unite.AI, summarising Alphabet’s SEC-filed earnings release: “Alphabet Posts 25% Revenue Growth as Google Cloud Jumps 103%”