Google is preparing to make one of the largest technology investments in corporate history. Alphabet expects its 2026 capital expenditure to reach between $180 billion and $190 billion, with much of that money supporting AI computing, servers, data centers and networking infrastructure.
Yet the company’s most anticipated new AI model has missed its expected launch window.
Gemini 3.5 Pro, Google’s next flagship model for advanced coding and agent-based tasks, was originally expected in June 2026. The model remains unavailable, creating an uncomfortable question for investors: if Alphabet is investing this much in artificial intelligence, why is one of its most important products still late?
The delay does not prove that Google has lost the AI race. But it does turn Alphabet’s coming earnings results and future product announcements into a major test of execution.
What Happened to Gemini 3.5 Pro?
Google introduced the Gemini 3.5 family during its May 2026 developer conference. Gemini 3.5 Flash arrived first, while the more capable Pro version was expected the following month.
That June release did not happen.
According to Reuters’ reporting on the Gemini delay, Gemini 3.5 Pro was designed to strengthen Google’s position in AI coding and agentic tasks—two areas attracting significant enterprise spending.
A subsequent Reuters update on Google’s Gemini releases said the Pro model reportedly fell short of internal performance goals, particularly in coding.
Google has not abandoned the model. The company says Gemini 3.5 Pro is being tested with partners and is coming soon, although it has not provided a revised public release date.
Meanwhile, Google has released several less expensive models, including Gemini 3.6 Flash, Gemini 3.5 Flash-Lite and a cybersecurity-focused version called Gemini 3.5 Flash Cyber. Training for the future Gemini 4 generation has also reportedly begun.
This makes the situation more complicated than a simple product failure. Google is still releasing AI products, but the flagship model that Wall Street is watching remains missing.
Why Coding Performance Matters So Much
Coding has become one of the most commercially valuable applications of generative AI.
Businesses can use advanced coding models to write software, detect errors, modernize older systems, create automated workflows and help developers complete projects faster. These benefits are easier to measure than many consumer AI features because companies can compare development time, labor costs and software output.
Agentic AI increases the potential value further. Instead of answering a single question, an AI agent can plan a task, use tools, examine results and continue working toward a broader objective.
Google highlighted these capabilities during its official I/O 2026 announcements, positioning Gemini 3.5 as a model family that combines advanced intelligence with the ability to take action.
A flagship model that struggles with coding would therefore create more than a technical problem. It could affect developer adoption, enterprise confidence and Google Cloud’s ability to compete for valuable AI workloads.
Delaying the model may be frustrating, but releasing it before it meets expectations could be more damaging. A weak launch would give developers a reason to build their workflows around competing platforms.
Understanding Alphabet’s $190 Billion Investment
The headline figure requires context.
Alphabet is not spending $190 billion exclusively on Gemini. The range represents total expected capital expenditure for 2026, including servers, data centers, networking equipment, energy infrastructure and other long-term assets.
Still, AI is a central reason for the increase.
In an official June 2026 investor presentation, Alphabet said it expected full-year capital expenditure of $180 billion to $190 billion. The company also indicated that spending could rise significantly again in 2027.
An Alphabet filing available through the US Securities and Exchange Commission said the company was experiencing unprecedented demand for AI computing resources.
The investment includes two major components:
- Servers and specialized AI processors needed to train and operate models.
- Data centers, networking systems and related infrastructure required to deploy those models globally.
Alphabet is effectively building the physical foundation for its future AI business. The risk is that infrastructure spending occurs immediately, while the revenue and profits it is supposed to generate may take years to fully appear.
Why Investors Are Concerned
The Gemini 3.5 Pro delay arrives at a sensitive time for Alphabet.
Reuters reported that Alphabet shares had fallen approximately 9% since late April as of July 21, even though the stock remained positive for the year. The decline reflects a broader debate about whether Big Tech companies are investing too aggressively in AI infrastructure.
Investors are watching several specific risks.
1. Spending Is Rising Faster Than Product Visibility
Data centers and AI chips require enormous upfront investment. If new products are delayed, shareholders may need to wait longer before seeing a return on that spending.
One delayed model does not invalidate Alphabet’s strategy. However, repeated delays could make the company’s capital expenditure look less disciplined.
2. Competition Is Moving Quickly
Google is competing with OpenAI, Anthropic and a growing number of Chinese open-source developers.
Open-source models can be downloaded, modified and deployed at relatively low cost. They may not outperform every closed model, but they can put pressure on pricing and make it harder for major AI companies to charge premium rates.
Google must therefore compete on performance, cost, reliability and distribution—not merely on the size of its models.
3. Coding Is Becoming an Important Competitive Benchmark
Coding tools are becoming gateways into broader enterprise AI platforms. Once a company builds its development process around one provider, moving to another platform can become difficult.
A delay gives competing companies more time to attract developers, collect feedback and improve their own products.
4. Infrastructure Creates Continuing Expenses
Data centers do not stop costing money after construction. They require electricity, cooling, maintenance, networking and frequent hardware upgrades.
These expenses can place pressure on margins even when revenue continues growing. Investors will therefore examine whether Alphabet’s AI-related income is expanding fast enough to justify the increasing cost base.
Google Cloud Is the Strongest Counterargument
The bearish interpretation overlooks one major fact: Google Cloud is growing rapidly.
Alphabet’s June investor presentation said Google Cloud generated approximately $20 billion in first-quarter revenue, representing about 63% year-over-year growth. Its operating margin reached approximately 33%, while the Cloud backlog climbed to roughly $462 billion.
Alphabet also said AI solutions had become the largest contributor to Cloud growth.
For the second quarter, analysts cited by Reuters expected Cloud sales to grow by approximately 64%. That was a forecast rather than a confirmed result, but it illustrates the strength investors expected from the business.
This matters because Alphabet is not betting on a single chatbot. It is building a complete AI ecosystem involving:
- Gemini models
- Google Cloud
- Search
- YouTube
- Android
- Workspace applications
- Consumer subscriptions
- Enterprise AI tools
- Custom processors
- Global data-center infrastructure
A temporary delay in one model may not significantly damage that ecosystem if Cloud demand and AI adoption continue rising.
Google’s Custom Chips Could Be a Major Advantage
Google has spent years developing its own Tensor Processing Units, commonly known as TPUs.
Unlike general-purpose processors, TPUs are designed specifically for machine learning and AI workloads. Google can use them to train its own models, operate AI services and offer computing capacity to Cloud customers.
According to Google Cloud’s official TPU overview, these processors are designed to accelerate the mathematical operations that power neural networks, large language models and reasoning agents.
Owning custom silicon may give Google greater control over performance, cost and supply than companies that depend entirely on third-party chipmakers.
This advantage becomes especially important as global demand for AI computing continues to exceed available capacity. If Google can combine competitive Gemini models with efficient in-house hardware, its infrastructure could become a profitable business even when individual model launches experience delays.
However, hardware cannot compensate indefinitely for weak software. Alphabet still needs Gemini 3.5 Pro and its successors to deliver results that developers and businesses consider competitive.
Is Google Actually Falling Behind?
The honest answer is that it is too early to know.
The delay is a meaningful warning about execution, particularly because coding and AI agents are strategically important. Google’s spending also increases the cost of getting product decisions wrong.
But several facts argue against declaring Google defeated:
- Gemini 3.5 Pro is delayed, not cancelled.
- Google continues releasing other Gemini models.
- The company says the flagship model is being tested with partners.
- Google Cloud is experiencing strong revenue and backlog growth.
- Alphabet owns a vast distribution network across consumer and enterprise products.
- Google controls important parts of its AI hardware and infrastructure.
- The company has the financial resources to keep investing through temporary setbacks.
The real issue is not whether Google can release another model. It is whether Gemini 3.5 Pro will be competitive enough to justify the wait.
A late but exceptional model could restore confidence quickly. A late model that still trails competitors would raise much more serious questions.
What Investors Should Watch Next
Alphabet’s next earnings call and future Gemini announcements should provide clearer evidence about the company’s progress.
Five indicators deserve particular attention.
A Revised Gemini 3.5 Pro Release Date
“Coming soon” is not a precise timeline. Investors and developers will want to know when the model will become broadly available.
Coding and Agent Performance
Benchmark scores are useful, but real developer feedback will matter more. The market will examine how reliably Gemini handles complicated coding projects and multi-step tasks.
Google Cloud Growth
Strong Cloud revenue could demonstrate that Alphabet is already monetizing its AI infrastructure, even before the flagship model arrives.
Capital Expenditure and Depreciation
Alphabet may explain how its infrastructure spending is divided and how quickly increasing depreciation costs could affect margins.
Enterprise Adoption
New customers, expanding contracts and AI-related backlog would show whether businesses view Google as a serious alternative to competing AI platforms.
The Broader Lesson From the Gemini Delay
The AI race is often presented as a sequence of model launches and benchmark victories. In reality, building a sustainable AI business requires much more.
A leading company needs capable models, affordable computing, reliable infrastructure, developer support, distribution, security and a clear path to monetization.
Google possesses most of those ingredients. Its challenge is coordinating them quickly enough to remain competitive.
The Gemini 3.5 Pro delay shows that even a company with enormous financial resources, world-class researchers and custom AI infrastructure cannot guarantee that every model will be ready on schedule.
AI development is becoming more expensive at the same time that competition is making models cheaper. That tension may define the next stage of the industry.
Final Verdict
Gemini 3.5 Pro’s delay is a setback, but it is not yet evidence that Alphabet’s AI strategy has failed.
Google’s planned $180 billion to $190 billion capital expenditure raises the stakes considerably. Investors will expect the company to translate that infrastructure into better models, faster Cloud growth and sustainable profits.
For now, Alphabet still holds several powerful advantages: Google Cloud, custom chips, billions of users and distribution across some of the world’s most important digital products.
But those advantages only matter if Google executes.
The next version of Gemini does not merely need to arrive. It needs to demonstrate that Alphabet’s historic AI spending is producing technology that can compete at the highest level.
Frequently Asked Questions
Why was Gemini 3.5 Pro delayed?
Reports indicate that the model had not yet met Google’s internal performance goals, particularly for coding tasks. Google has not announced a new public release date.
Has Google cancelled Gemini 3.5 Pro?
No. Google says the model is being tested with partners and will be released soon.
Is Alphabet spending $190 billion only on Gemini?
No. The figure represents Alphabet’s expected 2026 capital expenditure across servers, data centers, networking equipment, energy infrastructure and other long-term investments. AI computing is a major driver of that spending.
Why is Gemini 3.5 Pro important?
It is expected to be Google’s flagship model for advanced coding, reasoning and agentic AI workloads—areas with significant enterprise demand.
Does the delay make Alphabet stock a bad investment?
A single product delay cannot determine whether a stock is attractive. Investors must also consider valuation, Cloud growth, profitability, competition, capital spending and their own risk tolerance. This article is informational and not investment advice.
Next RBI Polymer Notes: Are ₹10 and ₹20 Plastic Notes Coming to India?
Join the discussion