Hardware Is Back. And That Changes the Partnership Equation.
Hardware Never Really Left. Its Growth Is Creating New Partnership Opportunities.
For the better part of the last decade, enterprise technology had a pretty clear hierarchy.
Software was where the growth was.
Hardware was the plumbing.
Recurring revenue, high gross margins and asset-light business models attracted the premium valuations. Servers, networking, storage and distribution were necessary, but they weren’t usually where people went looking for the most interesting growth stories.
AI is changing that equation.
And if you’re in partnerships, you should be paying attention.
Because one of the biggest shifts happening in technology right now isn’t just the rise of AI software.
It’s the return of the physical technology ecosystem required to make AI actually work.
The numbers are getting hard to ignore.
Gartner expects worldwide IT spending to reach $6.3 trillion in 2026, up 13.5% from last year.
Software spending is forecast to grow 15.1%.
IT services: 9%.
Devices: 8.2%.
But data center systems?
55.8%.
Gartner expects spending on data center systems to jump from roughly $506 billion in 2025 to almost $788 billion this year.
IDC is seeing the same thing. Worldwide server spending increased 30.7% year over year in Q1 2026, even though unit shipments grew just 3.3%.
And AI infrastructure spending reached roughly $318 billion in 2025, more than double the prior year.
The stock market has noticed.
Dell has been one of the standout performers as AI server demand exploded. HPE has rallied sharply alongside the infrastructure cycle. Cisco is reporting billions of dollars in AI infrastructure orders. Avnet, sitting much further down the technology supply chain, has been hitting record highs as demand for electronic components accelerates.
At first glance, this looks like a hardware story.
I think it’s actually a partnership story.
AI is rebuilding the technology value chain
Think about what it takes to deploy enterprise AI.
You need chips.
Servers.
Networking.
Memory.
Storage.
Power.
Cooling.
Cloud infrastructure.
Data platforms.
Security.
Models.
Applications.
Implementation.
Managed services.
No single company owns that stack.
Nvidia needs OEMs.
OEMs need component suppliers.
Infrastructure vendors need distributors.
Distributors need VARs and integrators.
Software companies need infrastructure.
Hyperscalers need ISVs.
Enterprises need systems integrators to put much of it together.
The more AI infrastructure grows, the more interconnected the technology industry becomes.
That’s very different from the SaaS model that dominated the last decade.
A SaaS company could build a product, hire a sales team and sell directly to customers. Partnerships were valuable, but in many organizations they remained an additional route to market.
AI infrastructure doesn’t work that way.
The ecosystem isn’t sitting next to the product. The ecosystem is part of the product.
Look at Dell
Dell might be the clearest example of the shift.
Its AI-optimized server revenue reached $16.1 billion in its fiscal first quarter, up 757% year over year.
AI server orders reached $24.4 billion.
Backlog ended the quarter at $51.3 billion.
Dell subsequently raised its fiscal 2027 AI server revenue expectation to $60 billion.
Those numbers are impressive on their own.
But think about the economic activity sitting around those servers.
Those systems need Nvidia or AMD accelerators.
They need networking.
They need storage.
They need power and cooling.
They often need data infrastructure and AI software.
They need to be installed.
They need to be integrated.
They need to be managed.
And they need to reach customers.
A dollar flowing through Dell can create opportunities across an entire network of partners.
That’s the difference between looking at AI as a product market and looking at it as an ecosystem.
Cisco tells a similar story
Cisco just reported $4 billion of AI infrastructure orders in a single quarter and $9.3 billion for the fiscal year.
Again, the obvious story is networking demand.
But networking doesn’t exist independently inside an AI deployment.
It’s attached to compute.
Compute is attached to storage.
Infrastructure is attached to cloud and data platforms.
And all of it eventually connects to applications and business outcomes.
That creates a fundamentally different partnership environment.
The question isn’t simply:
Who can resell our product?
It’s increasingly:
Which combination of companies needs to win together for the customer to deploy this architecture?
That’s a much more strategic role for partnerships.
Then look one layer deeper: distribution
This may be the most overlooked part of the story.
Companies like TD SYNNEX, Arrow and Avnet sit in the middle of enormous technology ecosystems.
Historically, we called them distributors.
That description increasingly feels incomplete.
Avnet recently reported fiscal fourth-quarter revenue growth of 48%, while its electronic components business grew 49%. Management pointed to AI, data centers and industrial demand as contributors to the rebound.
Why does that matter for partnerships?
Because complexity increases the value of the companies that connect the ecosystem.
AI infrastructure involves thousands of products and thousands of partners.
Someone needs to manage inventory.
Someone needs to finance purchases.
Someone needs to aggregate vendors.
Someone needs to help partners configure solutions.
Someone needs to enable the VARs and integrators selling into the customer.
Someone needs to understand where demand is appearing across the channel.
Distribution starts looking less like logistics and more like ecosystem orchestration.
And that makes distribution data increasingly interesting too.
A distributor can potentially see something an individual vendor cannot:
Demand forming across multiple vendors at the same time.
That’s a powerful market signal.
The opportunity isn’t just partner-sourced revenue
This is where I think partnership leaders need to expand the conversation.
For years we’ve tried to prove the value of partnerships using variations of the same metrics:
Partner-sourced pipeline.
Partner-influenced revenue.
Referrals.
Marketplace transactions.
Channel revenue.
Those metrics still matter.
But the AI infrastructure cycle creates something bigger.
Partnerships can become a source of market intelligence.
Imagine you’re trying to sell into a large enterprise.
Your CRM tells you what your company knows about that account.
But your ecosystem can tell you much more.
Which infrastructure vendors are already there?
Which distributor is seeing purchases?
Which SI has an active project?
Which cloud provider is involved?
Which data platform is expanding?
Which partners already have executive relationships?
Which technologies are being deployed together?
Which partners are seeing AI demand before it appears in your pipeline?
Suddenly partner data isn’t just attribution data.
It’s GTM intelligence.
This is why the hardware numbers matter
The growth in hardware spending tells us something about where the market is going.
But it also tells us something about how technology is going to be sold.
The larger and more complicated the infrastructure stack becomes, the harder it is for vendors to operate independently.
That increases the value of relationships between:
OEMs and semiconductor companies.
OEMs and distributors.
Distributors and VARs.
Cloud providers and ISVs.
Data companies and AI platforms.
Systems integrators and enterprise customers.
And increasingly, partnerships across all of them.
The opportunity for partnership teams is to understand those relationships better than anyone else inside the company.
Not simply who our partners are.
But how the market around the customer is connected.
Partnerships are becoming a data problem
This is the part I find most interesting.
Most partnership organizations still manage their ecosystems as lists.
Here are our partners.
Here are their tiers.
Here are the deals they sourced.
Here are our joint campaigns.
But the market isn’t a list.
It’s a network.
Dell connects to Nvidia.
Nvidia connects to cloud providers.
Cloud providers connect to data platforms.
Data platforms connect to ISVs.
Distributors connect hundreds of vendors to thousands of resellers.
Systems integrators connect all of them to enterprise transformation projects.
And enterprise customers sit in the middle of multiple overlapping ecosystems.
If you can map those relationships, you can start identifying things a traditional CRM can’t see.
Where demand is forming.
Where partners overlap.
Where an account is investing.
Where a new ecosystem is emerging.
And where the next opportunity might come from.
That’s the evolution from partner management to ecosystem intelligence.
Hardware is back. But that’s not really the story.
The technology industry spent the last fifteen years becoming increasingly software-centric.
AI is reminding us that software still needs infrastructure.
A lot of it.
Servers.
Networking.
Storage.
Memory.
Power.
Cooling.
Distribution.
Integration.
And an enormous network of companies required to put all of it together.
That’s why the hardware numbers matter.
Not because partnerships should suddenly become hardware experts.
Because the fastest-growing part of the technology market is becoming one of the most ecosystem-dependent parts of the technology market.
And that changes the partnership equation.
The companies that understand not only what customers are buying, but which ecosystems are forming around those purchases, are going to have an advantage.
Hardware is back.
And partnerships may be one of the best ways to understand where it’s going next.