Alphabet stock sheds $700B as AI bills climb because investors have stopped rewarding Google's AI spending the way they did a year ago. The stock is down roughly 15% from its all-time high, and the drop lines up with rising capex guidance, a negative free cash flow quarter, and an unusually public shake-up inside Google DeepMind.
Short answer: Alphabet stock has fallen about 15% from its May 2026 peak, erasing roughly $700 billion in market value. The slide follows a raised 2026 AI capex forecast ($195–205 billion), Alphabet's first negative free cash flow since its 2004 IPO, a Google DeepMind leadership shake-up, and senior researchers leaving for OpenAI and Anthropic.
I track vendor pricing and earnings pages for a living on this site, because a tool I recommend is only as good as the company still funding it next year. When I pulled the actual filings and earnings coverage behind this headline instead of just reading the round number, three separate stories turned out to be stacked on top of each other: a spending story, a cash-flow story, and a people story. Here's what's verified in each one, a comparison table of the timeline, and how to check a story like this yourself the next time an AI-spending headline shows up in your feed.
What's actually driving the sell-off
Alphabet hit an all-time high on May 13, 2026, after a roughly 150% run over the prior year. Since then it's given back about 15% and more than $690 billion in market value — call it $700 billion, which is the number most headlines use, though Bloomberg's own tracker put the exact figure at $692 billion the same day. That gap matters less than the direction: Alphabet is now the single biggest point-drag on the S&P 500 of any stock, according to AP’s August 28, 2026 report. None of this happened on one bad day. It built up over about ten weeks from three separate triggers, which is the part most coverage skips past.
Step-by-step: how to make sense of a story like this yourself
1. Check the capex guidance change, not just the stock move
At its Q2 2026 earnings call on July 22, Alphabet raised its 2026 capital expenditure forecast to $195–205 billion, up from $180–190 billion guided a quarter earlier. Q2 capex alone was $44.9 billion, roughly double the year-ago quarter, according to Alphabet’s own 8-K filing with the SEC. The stock fell about 6–7% that day despite Cloud revenue climbing to $24.8 billion and the Gemini app crossing 950 million monthly users — a beat on the business, a miss on the spending math investors wanted.
2. Read the cash-flow line, not just revenue growth
The same filing shows free cash flow of negative $5.9 billion for Q2 2026 — the first negative quarter since Alphabet's 2004 IPO, driven by that $44.9 billion in capex outrunning operating cash flow. A company this profitable running cash-negative for one quarter isn't a crisis by itself, but it's the exact number a skeptical investor points to first. My AI bubble breakdown goes deeper on how this compares to Amazon, Microsoft, and Meta's own 2026 capex guidance.
3. Separate a spending headline from a people headline
Two weeks later, on August 5, Google DeepMind CEO Demis Hassabis stepped aside to become the lab's chairman and Alphabet's chief scientist; DeepMind CTO Koray Kavukcuoglu took over day-to-day Gemini development as a senior VP reporting to Sundar Pichai, not as CEO, per Fortune’s coverage. Alphabet shares fell again on that news alone — a separate move from the capex-driven drop two weeks earlier.
4. Track the talent exits by name and destination
In my testing, this is the detail most casual coverage buries: Transformer co-author and Gemini co-lead Noam Shazeer left for OpenAI, and AlphaFold lead John Jumper — the 2024 Nobel Chemistry laureate — left for Anthropic, both reported by TechCrunch on June 24, 2026. A few weeks later, 27-year Google veteran and chief scientist Jeff Dean departed to co-found an AI-for-science startup, Discovery Loop, alongside fellow longtime researchers Sanjay Ghemawat, Oriol Vinyals, and Quoc Le — with Alphabet itself listed as a founding investor.
5. Check whether the product is actually behind, not just the headlines
Gemini 3.7 Flash shipped August 13, which I covered in my Gemini 3.7 Flash guide — a real release, not vaporware. But Gemini 3.5 Pro, the flagship model investors are actually waiting on, is still running behind its original schedule. That gap between "we're shipping" and "we're shipping the model that matters" is worth checking directly on Google’s own AI product page rather than taking either side's spin.
6. Decide what it changes for you, not for the stock
If you don't hold Alphabet stock, the practical question isn't whether Google "wins" the AI race this quarter — it's whether Gemini stays competitive enough to keep using. My Gemini vs. ChatGPT comparison is the more useful read if that's your actual concern.
Timeline: what happened and when
| Date | Event | Stock/market impact |
|---|---|---|
| May 13, 2026 | Alphabet stock hits all-time high | Peak market cap |
| Jun 24, 2026 | Noam Shazeer leaves for OpenAI; John Jumper leaves for Anthropic | Talent-exodus concerns build |
| Jul 22, 2026 | Q2 earnings: capex raised to $195–205B; FCF −$5.9B (first negative quarter since 2004) | Stock fell ~6–7% same day |
| Aug 5, 2026 | Hassabis steps down as DeepMind CEO, becomes chairman; Kavukcuoglu takes over Gemini as SVP | Stock fell further; single-day drop reported near $186B |
| Aug 6, 2026 | Alphabet sells $25B in bonds (drew ~$115B in demand) | Financing move, not a loss |
| Aug 13, 2026 | Gemini 3.7 Flash ships; Gemini 3.5 Pro still delayed | Mixed signal on product pace |
| Aug 27–28, 2026 | Cumulative losses reach ~15% off peak | ~$700B erased; biggest S&P 500 point-drag |
Example prompts you can copy
Use these to check a headline like this yourself instead of taking one outlet's number on faith:
- Fact-check a specific figure: "Using [source URL], confirm the exact stock-price percentage drop and dollar market-value figure reported for [company] on [date] — quote the sentence it comes from."
- Read the primary filing: "Summarize the capex guidance and free cash flow figures from this earnings release in three sentences, with no commentary: [paste filing text or URL]."
- Track a talent story: "List every senior departure from [company]'s AI division reported since [date], with each person's prior role, new destination, and the outlet that reported it."
- Sanity-check a chatbot's memory: "Do you have live, current data on [company]'s stock price, or is this from your training data? Tell me the cutoff before answering."
Common mistakes to avoid
The mistake I see most is treating a single day's stock move as the whole story instead of checking which of three different triggers caused it — a capex-guidance day, a leadership day, and a cumulative-toll day each moved this stock separately, not once. Second, don't ask a general-purpose chatbot for today's stock price or market cap and trust the number without checking the date; most models have a training cutoff and will guess rather than say so unless you ask directly, which is why I only trust figures pulled through live search grounding. Third, don't confuse raised capex guidance with money already lost — a spending plan for the rest of the year is not the same as a realized loss, and that mix-up was the single most common misreading in this week's coverage. Fourth, don't repeat a rounded headline number as exact — "$700 billion" and Bloomberg's own $692 billion figure are close enough not to matter for the story, but different enough that you should say "roughly" if you're citing it yourself.
Tools that make this easier
You don't need a Bloomberg terminal to check a story like this — you need somewhere to verify a number and somewhere to compare the AI tools actually affected by it. My AI bubble guide walks through the capex-versus-revenue math across all four major hyperscalers, not just Alphabet, and Apple’s low exposure to the same bubble risk is a useful contrast if you want to see how a company with almost no AI capex is positioned differently. If your actual question is which chatbot to use given Google's AI turbulence, my best AI models roundup and Gemini vs. ChatGPT head-to-head cover that directly, and how to use Gemini is worth a look if you're deciding whether to keep depending on Google's assistant day to day.
My take
The spending and cash-flow numbers are real and verifiable straight from Alphabet's own SEC filing — that part of the story isn't exaggerated. What I'd push back on is treating this as one clean narrative. A capex increase, a negative free-cash-flow quarter, a CEO stepping into a chairman role, and four researchers leaving for two different rivals are four separate, independently confirmed events that happened to land in the same ten weeks. Stacked together they read as "Google is losing," which is a stronger claim than any single one of them supports on its own. If you're not holding the stock, the number worth watching isn't the market cap — it's whether Gemini 3.5 Pro ships competitively once it finally lands.
Frequently Asked Questions
Did Alphabet really lose $700 billion in market value?
Close to it. Alphabet's stock is down about 15% from its May 13, 2026 all-time high, and most coverage rounds the loss to $700 billion. Bloomberg's own tracker put the more precise figure at $692 billion on the same day — both describe the same drop, just rounded differently.
Why is Alphabet's stock falling if AI spending is supposed to help Google compete?
Investors reacted to the cost side, not the strategy: raised 2026 capex guidance ($195–205 billion), a first-ever negative free cash flow quarter, and a leadership shake-up inside Google DeepMind all landed within about ten weeks. The market is pricing in near-term spending pressure, not necessarily doubting the long-term AI bet.
Is Google losing the AI race to OpenAI and Anthropic?
It's lost specific people to both — Noam Shazeer to OpenAI and John Jumper to Anthropic, among others — and Gemini 3.5 Pro is behind schedule. But Gemini 3.7 Flash shipped on time in August, and the Gemini app has grown to 950 million monthly users, so "losing" oversimplifies a picture that's mixed rather than one-directional.
Should I stop using Gemini because of this news?
Not based on this story alone. None of the reported issues are about Gemini's current models becoming unavailable or unreliable — they're about spending, cash flow, and personnel. If you're deciding between AI tools generally, judge Gemini on how it performs for you today, covered in my Gemini vs. ChatGPT comparison.
How can I track a fast-moving AI-industry story like this myself?
Go to the primary source first — an SEC filing or the company's own newsroom — before trusting a headline's rounded number, and ask any AI tool you use for research whether it has live data or is guessing from a training cutoff. The example prompts above are close to what I used to build the timeline in this piece.