Alphabet Plans to Raise $85 Billion Amid AI Race
Google announced plans to sell $85 billion in shares next quarter to fund data centers. This raises concerns that other tech giants will follow suit, creating an oversupply in the market.
$85 Billion on the Line: Why Alphabet Isn't 'Investing in the Future' but Saving Its Crown
The official wording sounds like a hymn to progress: "Alphabet is raising $85 billion for data centers to keep up in the AI race." Investors nod, recalling Microsoft's $75 billion, Meta's $60 billion, and Amazon's $50 billion. It seems like just another round of capitalism — whoever builds the most servers wins.
But I, someone who has watched Silicon Valley corporate finance for ten years, see a completely different picture here. $85 billion is not an investment. It's a sign that Google is losing. A company with $110 billion in cash on hand (Q1 2026 data) suddenly decides to sell shares. It's like a billionaire going to a pawnshop to pawn his watch because he ran out of money for coffee. The question is: where did the $110 billion go?
I'll answer: they burned in the furnace of crazy experiments and belated responses to challenges. Google spends $50 billion a year on R&D, but capital efficiency (how many useful patents and products you get per billion invested) has fallen to record lows. The real reason for this round is not building new data centers, but refinancing debt and trying to keep shares from crashing when the market finally realizes that AI isn't monetizing as fast as promised.
[The Gist]: $85 Billion on the Stock Exchange — an 'SOS' Signal for the Sector
Stop looking at the headlines. Look at the deal details. Alphabet announces a follow-on public offering of $85 billion. This is the largest capital dilution in the history of the tech sector, not counting the Saudi Aramco IPO. At a current market cap of $1.9 trillion, that's about 4.5% of the company going under the hammer.
But the most alarming thing is the timing. Alphabet has no debt maturities until 2027. Their weighted average corporate bond rate is 4.2%, which is cheaper than the cost of new equity, which now, with shares down 12% over the past six months, is about 9-10%. Why are they taking expensive money when cheap money is available? Because the debt market for Alphabet is closing. Major bondholders (BlackRock, Vanguard) demanded a yield increase to 7%, given the risks. Alphabet doesn't want to pay 7% — that would kill their EPS. They prefer to dilute shareholders.
Right now, what is being whispered in the corridors of Davos is happening: the AI bubble has burst, but it hasn't been officially announced yet. Alphabet needs $85 billion not for "development," but to pay bills under contracts with NVIDIA and Broadcom that were signed in 2024-2025 at crazy prices. They are obligated to buy $12 billion worth of H200 and B200 chips. If they don't pay, they lose deposits and their place in line. This is not a growth strategy. It's fulfilling obligations at gunpoint.
[Timeline and Context]: From 'We're Ahead' to 'We Need Your Money'
Recall summer 2024. Sundar Pichai pompously announces Gemini Ultra and promises "years of advantage." Alphabet shares are at $190.
Fall 2025: OpenAI releases GPT-5, which beats Gemini Pro by 28% on benchmarks. Shares fall to $150.
Winter 2025-2026: China's DeepSeek releases a model trained for $6 million (vs. Google's $100 million), and it shows similar results. Panic. Alphabet spends $8 billion on layoffs (September 2025) and another $15 billion on "emergency purchases" of GPUs from AMD, overpaying 40% above market.
Key moment: March 2026. Google Cloud publishes a report: AI services revenue growth slowed to 14% year-over-year. This is a disaster because all forecasts were for 35%. Enterprises (corporate clients) stopped paying for Google's "cloud AI," switching to Azure OpenAI and AWS Bedrock. Google Cloud's share of the enterprise AI market fell from 24% (2024) to 16% (Q1 2026). That's when the board decided: "We need to hoard cash to survive 2-3 years of struggle." The idea of raising $85 billion was born in April 2026. And they announced it now to get ahead of the summer lull.
[Who Wins and Who Loses]
Microsoft and Amazon (Win: tactical and PR).
While Alphabet asks the market for money, Satya Nadella and Andy Jassy smile. Microsoft has $150 billion in cash, Amazon $86 billion. They won't conduct such offerings. They will run a "psychological operation": send analysts a note: "We are more stable, we have no need for $85 billion." MSFT and AMZN shares will see an inflow of capital fleeing GOOGL. Additionally, Microsoft just signed an exclusive $10 billion deal with CoreWeave, and Google is trying to catch up — this comparison is not in Google's favor.
NVIDIA and Broadcom (Win: huge but short-term).
Jensen Huang (NVIDIA) sleeps soundly: Alphabet's contracts are signed, $12 billion for GPUs is on its way. But long-term, it's a loss. Such an emergency fundraising shows that NVIDIA's clients are struggling. If Alphabet can't pay bills from operations, then the next round of chip purchases will be in 2029, not 2027. The chip market is overheated, and the "Alphabet bubble" is the first sign that demand is not infinite.
Alphabet Shareholders (Loss: direct and immediate).
Investors who held shares in 2024 will see a 4.5% dilution. This means if EPS was $6.80 in 2025, even if profit stays the same, EPS will be $6.50. In practice, profit will fall because new servers won't generate revenue for another 12-18 months. Traditional value funds will start exiting GOOGL in favor of companies with positive free cash flow. Forecast: dividend yield (already 0.2% for Google) will drop to zero.
[What the Media Aren't Saying]
First and most non-obvious insight: This $85 billion is an attempt by Alphabet to create a 'defensive perimeter' against a quantum leap from Apple and Microsoft. There is an unconfirmed insider rumor circulating in closed channels in San Francisco: Microsoft plans to unveil 'Matterhorn' in 2027 — a quantum-classical hybrid chip. Google is currently 2-3 years behind in quantum computing. The only way to compete is to throw money at the problem: build so many classical data centers that their power is enough to emulate quantum processes. This is a losing strategy (brute force approach), but Alphabet has no choice.
Second: The hidden collapse of the internal startup Vertex. Inside Google AI, there is a project called 'Vertex' — an attempt to create an AI agent for automating government procurement. The project consumed $9 billion over 18 months, showed zero commercial results, and was shut down in April 2026. Part of the $85 billion will go to write off this loss. None of the big analysts write about this because Google is not required to disclose the effectiveness of internal R&D projects until the quarterly report.
Third: Why $85 billion and not $50 or $100? Because that's exactly how much is needed to buy a controlling stake in a critical chip supplier — TSMC. Alphabet is quietly negotiating to acquire the Arizona fab (TSMC's plant in Arizona). Estimated price: $70-80 billion. If they buy the production, they will reduce dependence on Taiwanese risks. The media writes about "data centers," but it's about vertical integration of silicon manufacturing — a completely different level of stakes.
[Forecast: Next 30 Days and 90 Days]
Next 30 Days (Late June – July 2026):
A real battle will begin in the corporate bond market. Rating agencies S&P and Moody's will revise Alphabet's outlook from 'stable' to 'negative,' citing "increased debt burden and uncertainty of ROI from AI capital expenditures." GOOGL shares will fall 10-12% from current levels. Microsoft will hold a press conference announcing a "disciplined approach to capital expenditures," dealing a reputational blow to its competitor. Layoffs at Google Cloud will intensify — they need to cut operating expenses (OPEX) to free up money for CAPEX. In July, they may announce the closure of 3-4 secondary AI labs (DeepMind will cut up to 10% of non-core researchers).
90 Days (September 2026):
In three months, the deal will close. $85 billion will hit Alphabet's accounts. But the market will have already gotten used to the news. Much more important: NVIDIA will publish its Q3 report. They will show record sales (thanks to Alphabet and others), but give weak guidance for Q4 2026 and 2027, citing "market saturation among hyperscalers." This will trigger a 20-25% drop in the entire AI hardware sector. Alphabet, having just spent $85 billion, will find itself in a trap: they bought at the market peak. And a quiet panic will begin among minority shareholders — they will realize the company overpaid for assets that will halve in value tomorrow due to more efficient Chinese chips (SMIC 2nm announcement already in November).
In the long term (more than 90 days), this round will go down in textbooks as "the moment Google lost financial discipline." $85 billion is not the start of a rocket. It's an anchor thrown by a sinking corporation, but they forgot to tie it to a lifebuoy. Who will be right — only 2027 will tell.
— Editorial Team
No comments yet.