TSMC’s July sales have surged nearly 45% from a year earlier, providing fresh evidence that demand for artificial intelligence hardware remains exceptionally strong. Taiwan Semiconductor Manufacturing Company reported NT$467.58 billion ($14.5 billion) in revenue for July, representing a 44.7% year-on-year increase and a 5.6% rise from June.
The latest figures reinforce the strength of the AI infrastructure cycle that has powered TSMC’s growth throughout 2026. The company manufactures advanced chips for some of the world’s largest technology companies, including Nvidia and Apple, and has become one of the most important companies in the global AI hardware supply chain.
TSMC July revenue jumps 44.7%
TSMC’s July performance was significantly stronger than the growth rates seen in many traditional semiconductor markets.
| Metric | July 2026 |
|---|---|
| Revenue | NT$467.58 billion |
| US-dollar equivalent | ~$14.5 billion |
| Year-on-year growth | 44.7% |
| Month-on-month growth | 5.6% |
| First seven months revenue | NT$2.87 trillion |
| January-July growth | ~37% |
TSMC’s revenue for the first seven months of 2026 reached approximately NT$2.87 trillion, up around 37% from the same period a year earlier.
TSMC JULY SALES
July 2025
████████████████████
July 2026
████████████████████████████
+44.7%
Revenue:
NT$467.58 billion
≈ US$14.5 billion
The numbers show that TSMC’s growth is not being driven by a single strong quarter. Revenue momentum has continued into the second half of the year.
AI remains the biggest growth engine
The central driver behind the strong performance is continued demand for advanced chips used in AI infrastructure.
AI companies are deploying increasingly powerful models, while cloud providers are expanding data-centre capacity to support training and inference.
That creates demand for increasingly sophisticated processors.
AI MODELS
↓
More computing demand
↓
More AI accelerators
↓
More advanced chips
↓
TSMC manufacturing
↓
Higher revenue
TSMC sits at a critical point in this chain because many of the most advanced AI processors are manufactured through its foundry network.
Reuters has previously reported that increasing adoption of AI models across consumer, enterprise and sovereign applications is driving demand for greater computing power and advanced semiconductor chips.
Nvidia is a major part of the story
One of the biggest beneficiaries of TSMC’s advanced manufacturing capabilities is Nvidia.
Nvidia designs the AI accelerators used in many large-scale AI data centres, while TSMC manufactures those advanced processors.
NVIDIA
↓
AI chip design
↓
TSMC
↓
Advanced manufacturing
↓
AI server
↓
Data centre
↓
AI applications
This makes TSMC a key indirect beneficiary of the enormous investment flowing into Nvidia-based AI infrastructure.
The relationship also explains why strong Nvidia demand can translate into higher manufacturing demand at TSMC.
The AI hardware cycle is still accelerating
The July sales figure is particularly notable because TSMC was already coming off a strong first half of the year.
The company’s June revenue rose 67.9% year over year to NT$442.68 billion, according to Reuters, while second-quarter revenue increased 36% from a year earlier and exceeded market expectations.
TSMC REVENUE MOMENTUM
June 2026
+67.9% YoY
↓
July 2026
+44.7% YoY
↓
First 7 months
+37% YoY
The growth rates fluctuate from month to month, but the overall trend remains exceptionally strong.
Why advanced chips matter
Not all semiconductor manufacturing is equally valuable.
AI accelerators and other high-performance processors require some of the industry’s most advanced manufacturing technologies.
These chips involve:
- Extremely small transistor structures
- Advanced packaging
- High-performance computing
- Large memory bandwidth
- Complex chip designs
- Sophisticated power management
AI CHIP
Advanced process
+
High transistor density
+
Advanced packaging
+
High-bandwidth memory
↓
High-performance AI system
TSMC’s ability to manufacture these chips at scale is one of the company’s biggest competitive advantages.
TSMC is at the centre of the AI supply chain
The company’s importance extends beyond manufacturing individual processors.
TSMC connects chip designers with the physical infrastructure needed to produce advanced semiconductors.
CHIP DESIGNERS
│
├── Nvidia
├── AMD
├── Apple
└── Other customers
↓
TSMC
↓
Advanced manufacturing
↓
Chip packaging
↓
AI / computing
This position gives TSMC exposure to multiple parts of the technology industry without requiring it to develop its own consumer devices or AI models.
First seven months revenue reaches NT$2.87 trillion
The cumulative revenue number is another important indicator.
TSMC generated approximately NT$2.87 trillion during the first seven months of 2026, representing growth of roughly 37% year over year.
JAN–JUL 2026
NT$2.87 trillion
████████████████████████████████
Growth:
~37% YoY
This puts TSMC on a strong trajectory for the full year.
The cumulative figure also reduces the risk of interpreting one unusually strong month as an isolated event.
Demand is extending beyond AI training
The AI boom initially centred heavily on training large models.
But inference is becoming increasingly important.
Once an AI model is trained, enormous computing resources may be required to serve millions or billions of user requests.
MODEL TRAINING
+
MODEL INFERENCE
+
AI AGENTS
+
ENTERPRISE AI
↓
CONTINUOUS COMPUTE DEMAND
↓
MORE AI CHIPS
The growth of AI agents could further increase inference demand because an agent may perform multiple model calls while completing a task.
That could create a longer-lasting demand cycle for AI processors.
AI data centres need more than GPUs
The AI infrastructure boom is also increasing demand for a broader range of semiconductor products.
A modern AI data centre requires:
| Component | Role |
|---|---|
| AI accelerators | Model training and inference |
| CPUs | General-purpose computing |
| Networking chips | Connect servers and clusters |
| Memory | Store and move model data |
| Power-management chips | Manage electricity |
| Storage controllers | Data storage |
| Connectivity chips | Network communication |
AI DATA CENTRE
│
┌─────┼─────┬─────┐
↓ ↓ ↓ ↓
GPU CPU MEMORY NETWORK
│ │ │ │
└─────┴─────┴──────┘
↓
AI SYSTEM
This means the AI boom can benefit multiple areas of the semiconductor supply chain.
TSMC’s advanced packaging business is increasingly important
AI processors are becoming more complex, with multiple components needing to work together at extremely high speeds.
That makes advanced packaging increasingly important.
CHIPLET
+
MEMORY
+
COMPUTE
+
INTERCONNECT
↓
ADVANCED PACKAGING
↓
AI PROCESSOR
TSMC has invested heavily in advanced packaging technologies to support the growing needs of AI customers.
This is important because simply producing smaller transistors is no longer enough to deliver the performance required by cutting-edge AI systems.
TSMC’s growth supports the broader AI investment story
The company’s sales figures provide an important real-world indicator for the AI boom.
Technology companies can announce enormous AI infrastructure spending plans, but TSMC’s revenue shows whether that spending is translating into actual semiconductor demand.
AI INVESTMENT
↓
Chip orders
↓
TSMC production
↓
TSMC revenue
The nearly 45% July increase therefore provides another data point suggesting that AI infrastructure spending is translating into tangible hardware demand.
Big Tech’s AI spending is creating a massive chip market
Microsoft, Google, Amazon and Meta are all expanding AI infrastructure.
Nvidia is also working with major financial institutions to help mobilise hundreds of billions of dollars for additional AI infrastructure.
That creates a powerful demand environment for advanced semiconductor manufacturers.
MICROSOFT
+
GOOGLE
+
AMAZON
+
META
+
AI STARTUPS
+
GOVERNMENTS
↓
AI INFRASTRUCTURE SPENDING
↓
AI CHIP DEMAND
↓
TSMC
The latest TSMC sales figures suggest that this spending cycle remains active.
Microsoft is also developing its own AI chips
Microsoft’s reported plans to unveil its next-generation Maia 300 AI accelerator highlight another trend in the market: major cloud providers increasingly want custom silicon.
Microsoft is reportedly discussing manufacturing capacity with TSMC for more than 300,000 Maia 300 chips, with a longer-term ambition of more than one million units.
This could actually create additional demand for TSMC.
CUSTOM AI CHIPS
Microsoft
↓
Maia design
↓
TSMC manufacturing
↓
Azure AI infrastructure
Even when technology companies develop their own AI processors, they may still depend on TSMC to manufacture them.
Custom chips do not necessarily threaten TSMC
At first glance, Microsoft’s Maia, Google’s TPUs and Amazon’s Trainium could appear to challenge Nvidia.
But from TSMC’s perspective, the situation can be different.
TSMC can manufacture chips designed by many competing companies.
NVIDIA ──────┐
AMD ─────────┤
Apple ───────┤
Microsoft ───┤
Other ASICs ─┤
↓
TSMC
↓
Advanced chips
This gives TSMC an unusual position in the AI ecosystem.
It does not need every customer to use the same architecture.
It can benefit as long as demand for advanced chip manufacturing continues.
TSMC’s growth is broader than one customer
Another advantage is diversification.
Nvidia may be one of the most important AI customers, but TSMC also manufactures chips for other major technology companies.
That means the company can benefit from growth across:
- AI
- Smartphones
- High-performance computing
- Automotive
- Consumer electronics
- Networking
- Data centres
AI is currently one of the most powerful growth engines, but it is not TSMC’s only business.
Smartphone demand is a different story
The semiconductor industry is not uniformly strong.
Traditional consumer electronics and smartphone demand can be considerably more cyclical than AI infrastructure.
This creates an important distinction:
TRADITIONAL CHIP DEMAND
↓
Consumer cycles
↓
More volatile
AI CHIP DEMAND
↓
Data-centre investment
↓
Massive infrastructure buildout
↓
Currently much stronger
TSMC’s latest sales figures suggest that AI-related demand is currently providing a powerful offset to weaker areas of the semiconductor market.
The AI boom is increasing TSMC’s strategic importance
TSMC’s position has become increasingly important because advanced AI processors are difficult to manufacture.
A limited number of companies have the technology and scale required to produce the most sophisticated chips.
That makes TSMC strategically important to:
- Nvidia
- Microsoft
- AMD
- Apple
- Other AI companies
AI INDUSTRY
↓
Needs advanced chips
↓
Limited advanced manufacturing capacity
↓
TSMC
↓
Global AI supply chain
This also explains why governments and companies are increasingly concerned about semiconductor supply-chain resilience.
Taiwan’s role remains critical
TSMC’s dominance means Taiwan remains a crucial part of the global technology supply chain.
The company has expanded manufacturing internationally, including investments in the United States, but Taiwan remains its core manufacturing base.
GLOBAL AI SUPPLY CHAIN
Chip design
↓
Advanced manufacturing
↓
Packaging
↓
AI servers
↓
Data centres
↓
Global AI services
Any disruption to advanced semiconductor production could therefore have consequences far beyond Taiwan.
TSMC is investing heavily to meet demand
Strong AI demand also means TSMC needs to expand capacity.
The company has been investing in new fabs and advanced packaging facilities to meet customer demand.
The challenge is that semiconductor manufacturing capacity takes years and billions of dollars to build.
HIGH AI DEMAND
↓
Capacity shortage
↓
New fab investment
↓
Construction
↓
Equipment installation
↓
Qualification
↓
Mass production
This long investment cycle makes supply planning particularly important.
The risk of overbuilding
The biggest long-term question is whether AI chip demand will remain strong enough to justify the enormous capacity being built.
If AI adoption continues accelerating, today’s capacity investments could prove insufficient.
If demand slows dramatically, the industry could eventually face excess capacity.
STRONG AI DEMAND
↓
More fabs
↓
Higher utilisation
↓
Strong returns
WEAKER AI DEMAND
↓
More fabs already under construction
↓
Lower utilisation
↓
Pressure on returns
For now, TSMC’s sales figures suggest demand remains robust.
AI efficiency could change the equation
Another factor is the rapid improvement in AI model efficiency.
If developers can achieve the same performance with fewer chips, demand growth could eventually slow.
However, efficiency improvements can also make AI cheaper to deploy, potentially increasing usage and offsetting the reduction in compute required per task.
MORE EFFICIENT AI
↓
Lower cost per task
↓
More AI adoption
↓
More total AI workloads
↓
Potentially more overall compute
This is one of the reasons predicting long-term AI chip demand remains difficult.
TSMC’s July numbers in one infographic
┌──────────────────────────────────────┐
│ TSMC JULY 2026 │
├──────────────────────────────────────┤
│ Revenue NT$467.58B │
│ US$ value ~$14.5B │
│ YoY growth +44.7% │
│ MoM growth +5.6% │
│ Jan–Jul revenue NT$2.87T │
│ Jan–Jul growth ~37% │
│ Main growth driver AI hardware │
└──────────────────────────────────────┘
What the numbers mean for Nvidia
For Nvidia, strong TSMC sales are another positive indicator that demand for its AI hardware remains high.
Nvidia needs advanced manufacturing capacity to produce its latest generations of AI accelerators.
STRONG NVIDIA DEMAND
↓
More chip orders
↓
More TSMC production
↓
Higher TSMC revenue
The relationship works in the other direction as well: TSMC’s manufacturing capacity can influence how quickly Nvidia can supply the market.
What the numbers mean for AMD
AMD is also competing aggressively in AI accelerators.
Like Nvidia, it relies on advanced foundry manufacturing for its high-performance processors.
A strong TSMC environment therefore supports the broader AI semiconductor ecosystem rather than just Nvidia.
What the numbers mean for Microsoft
Microsoft’s Maia strategy provides another example of how cloud companies are becoming semiconductor designers.
Even if Microsoft reduces some dependence on Nvidia, it still needs a manufacturer capable of producing advanced chips at scale.
That could mean continued business for TSMC.
What the numbers mean for investors
TSMC’s sales provide one of the clearest real-time indicators of semiconductor demand.
Investors can watch monthly revenue to understand whether AI infrastructure spending is translating into actual chip production.
| Indicator | What it tells investors |
|---|---|
| TSMC monthly sales | Current chip demand |
| Advanced-node demand | AI / high-performance computing strength |
| Packaging capacity | AI system demand |
| Capex | Future capacity expansion |
| Nvidia orders | AI accelerator demand |
| Cloud capex | Future AI infrastructure demand |
The AI semiconductor flywheel
The current cycle can be represented as:
AI MODELS
↓
MORE USERS
↓
MORE INFERENCE
↓
MORE COMPUTE
↓
MORE AI CHIPS
↓
MORE TSMC ORDERS
↓
MORE FACTORY INVESTMENT
↓
MORE AI COMPUTE CAPACITY
↓
MORE AI APPLICATIONS
↓
MORE USERS
TSMC’s latest revenue number suggests this flywheel is still turning rapidly.
Why 45% growth matters
A 45% annual increase would be impressive for almost any mature technology company.
It is particularly notable for TSMC because of the scale at which it already operates.
The company generated NT$467.58 billion in a single month, yet still achieved nearly 45% year-on-year growth.
That indicates how large the incremental demand generated by AI infrastructure has become.
The bigger semiconductor story
TSMC’s performance is increasingly becoming a proxy for the health of the advanced semiconductor industry.
When TSMC reports strong growth, it can indicate strength across several technology markets simultaneously.
TSMC
│
├── AI
├── Data centres
├── Smartphones
├── HPC
├── Networking
└── Advanced computing
The latest results strongly reinforce the view that AI remains the dominant growth engine for advanced chips.
What to watch next
TSMC’s quarterly results
Investors will look for confirmation that the monthly sales momentum is translating into sustained quarterly growth.
AI chip orders
Nvidia, AMD and custom-chip customers will provide clues about future production demand.
Advanced packaging capacity
Packaging has become a potential bottleneck as AI systems become more complex.
TSMC capital expenditure
Higher spending would indicate confidence that demand will remain strong.
AI infrastructure spending
Cloud providers’ capital expenditure plans will remain one of the most important leading indicators.
AI model efficiency
If models become significantly more efficient, the industry will need to determine whether higher usage offsets lower compute requirements.
Geopolitical risks
TSMC’s central role in the global semiconductor supply chain means geopolitical developments remain an important risk factor.
Conclusion
TSMC’s 44.7% year-on-year jump in July revenue to NT$467.58 billion, or roughly $14.5 billion, provides one of the clearest signs yet that the AI hardware boom remains powerful. Revenue also increased 5.6% from June, while cumulative sales for the first seven months reached approximately NT$2.87 trillion, up around 37% from the same period a year earlier.
The numbers matter because TSMC sits at the centre of the advanced semiconductor supply chain. The company manufactures chips for Nvidia and other major technology companies, making its sales a useful indicator of whether the enormous investments being announced across AI infrastructure are translating into actual hardware demand.
The evidence so far suggests they are.
The AI industry is moving from model development toward large-scale deployment. Training increasingly capable models requires enormous computing clusters, but inference is becoming an equally important source of demand as millions of users interact with AI systems and businesses deploy AI agents.
That means demand for AI chips is becoming more continuous.
The shift toward custom AI accelerators is also unlikely to eliminate TSMC’s role. Microsoft is reportedly preparing its Maia 300 chip and discussing manufacturing capacity for hundreds of thousands of units, potentially scaling beyond one million. Google’s TPUs and Amazon’s Trainium chips follow a similar model.
For TSMC, this creates an attractive situation: it can potentially manufacture chips for companies that compete with one another.
Nvidia may dominate AI accelerators today, but Microsoft, Google, Amazon, AMD and other companies are developing alternatives. If those chips are manufactured through TSMC, the foundry can benefit from the broader expansion of AI computing regardless of which chip architecture ultimately wins.
The company also benefits from the increasing importance of advanced packaging. Modern AI processors are becoming highly complex systems involving compute dies, memory and other components that must work together at extremely high speeds.
This means TSMC’s competitive advantage is no longer simply its ability to produce smaller transistors. Its manufacturing technology, packaging capabilities and ability to produce advanced chips at enormous scale are all becoming strategically important.
There are risks, of course.
The current AI infrastructure boom assumes that AI demand will continue to grow rapidly. If companies eventually discover that they have overbuilt data-centre capacity, or if improvements in model efficiency dramatically reduce computing requirements, semiconductor demand could slow.
At the same time, efficiency improvements can lower the cost of AI and potentially encourage much greater adoption, creating even more total workloads. The balance between efficiency gains and increased AI usage will be one of the most important factors determining the industry’s long-term trajectory.
For now, however, TSMC’s numbers point in one direction: AI hardware demand remains exceptionally strong.
The company’s July sales surge, combined with strong first-half performance, suggests the AI infrastructure cycle has not yet lost momentum. As Nvidia, Microsoft, Google, Amazon, AMD and other companies continue expanding AI computing capacity, TSMC is positioned to remain one of the biggest beneficiaries of the global semiconductor investment wave.
The bigger message for the technology industry is clear:
The AI boom is no longer just a software story. It is becoming one of the largest semiconductor and infrastructure investment cycles in history—and TSMC is sitting at its centre.
Get the day’s top stories in your inbox
One concise email. No spam, unsubscribe anytime.
