Second-Life Compute: A Secretary Suite Note On Smartphone Standardization, Factored Networks, And End-Of-Life Civic Infrastructure

Second-Life Compute: A Secretary Suite Note On Smartphone Standardization, Factored Networks, And End-Of-Life Civic Infrastructure

DOI: To be assigned.

John Swygert

July 10, 2026

Abstract

A recent Google-supported effort with researchers at the University of California San Diego demonstrates that retired smartphones can be stripped to their motherboards, loaded with general-purpose Linux systems, and clustered into low-carbon cloud-computing infrastructure. This is an important proof of concept, but it should not remain merely a recycling story or a university demonstration. It should become a design mandate.

This paper argues that smartphones should be manufactured from the beginning as dual-life devices: first as consumer communication devices, then as standardized compute-board components in factored networks. The central claim is simple: obsolete hardware is often obsolete only in its original role. When placed into a routed, distributed, workload-aware network, the same hardware may retain substantial civic, educational, research, and infrastructure value.

Secretary Suite has previously argued for the reclassification of old hardware as recoverable functional infrastructure rather than waste. The Google / UC San Diego project now demonstrates the same direction of thought at institutional scale. The next step is not only to reuse old phones. The next step is to redesign new phones so that reuse is easy, safe, standardized, and expected.

01 Purpose Of This Paper

The purpose of this paper is to identify the missing design implication in current discussions of phone-cluster computing.

Google and UC San Diego are not merely showing that old phones can be reused. They are showing that the smartphone should be understood differently from the beginning. A smartphone is not only a finished consumer object. It is a layered compute system. When its screen, battery, camera, casing, and consumer desirability expire, its motherboard may remain a valuable processing unit.

The proper design question is therefore not:

Can old phones be reused?

The better question is:

Why were phones not designed for this second life from the beginning?

A device that contains useful compute capacity should not become hard-to-process waste merely because its first consumer role has expired. It should be built with a known second-life pathway.

This paper proposes that future smartphones should be designed for end-of-life conversion into standardized compute-board components for factored networks.

02 The Demonstration Now In Public View

Google Research has described a project in which researchers at the University of California San Diego are building a second-life computing platform from retired smartphones. The process extracts smartphone motherboards, collects them into clusters, and redeploys them as a general-purpose computing platform. Google states that UC San Diego plans to deploy a data center built from 2,000 Pixel smartphones to provide low-cost, low-carbon cloud computing for researchers and students.

This matters because the project does not treat the phone as a failed consumer object. It treats the phone as a recoverable compute layer.

The display is removed.

The battery is removed.

The chassis is removed.

The cameras and nonessential consumer hardware are removed.

The motherboard remains.

That motherboard contains the processor, accelerators, memory, storage, and core compute functionality. Google also notes that the motherboard represents approximately half of a phone’s embodied carbon footprint, making it the most carbon-significant part to recover.

This is the key reclassification.

The phone is no longer a phone.

It is a compute-board component.

Once the consumer shell is gone, the remaining board can be routed into a different system.

That is Secretary Suite logic.

03 The Secretary Suite Alignment

Secretary Suite has repeatedly treated “obsolete” systems as misclassified capacity.

An old machine is not automatically dead.

An old building is not automatically useless.

An old industrial site is not automatically blight.

An old phone is not automatically waste.

The question is not whether the object has lost its prior role. The question is whether some functional layer inside it can be routed into a new system.

This is why the Google / UC San Diego project is Secretary Suite aligned. It does not merely recycle material. It recovers function.

Recycling asks:

What material can be recovered after the object fails?

Second-life compute asks:

What function still remains before the object is destroyed?

That is a higher form of reuse.

Material recycling is important, but it is often the last stage of value recovery. It takes the object apart to salvage matter. Second-life computing takes the object apart to salvage computation.

That distinction matters.

The future should not be built only around recycling waste. It should be built around preventing functional systems from being prematurely reduced to waste.

04 Obsolete As A Consumer Object Does Not Mean Obsolete As Infrastructure

A retired smartphone may feel obsolete to its owner.

The battery may be weak.

The camera may be outdated.

The screen may be cracked.

The operating system may no longer receive ordinary consumer support.

The device may feel slow compared with the newest model.

But none of those facts prove that its motherboard has no remaining public value.

This is the category mistake.

The consumer judges the phone as a phone.

The infrastructure planner should judge the board as a compute component.

Those are different evaluations.

A retired smartphone should not be measured only against the newest consumer phone. It should be measured against the workload it can still perform inside a properly factored network.

A board that cannot compete with a modern flagship phone may still run classroom tasks.

It may still run grading jobs.

It may still host small cloud services.

It may still support local research workloads.

It may still serve as an edge-compute node.

It may still provide backup civic computing.

It may still contribute to a distributed educational cloud.

Its first life may be over.

Its second life may be just beginning.

05 The NASA Comparison And The Meaning Of Hidden Capacity

The power of this idea becomes clearer when modern “obsolete” hardware is compared with earlier mission-scale computing.

The Apollo Guidance Computer was a marvel of its time. A technical analysis describes it as having 72 KB of ROM, 4 KB of RAM, and approximately 14,245 FLOPS. It helped support one of the most significant technological achievements in human history under extraordinary constraints.

The point is not to make a careless one-to-one comparison between a modern phone and all Apollo-era computation. The exact comparison depends on which devices, which mission systems, which performance metrics, and which year are being compared. The responsible claim is still powerful enough:

Modern discarded or retired phones often contain computational capacity that would have been extraordinary by historical standards.

That should change how society thinks about the junk drawer.

When a device that would have been astonishing in the Apollo era is sitting unused in a drawer, the issue is not merely waste. It is classification failure.

The phone is called obsolete because the consumer market has moved on.

But the compute substrate remains.

The question is whether society has a system capable of receiving it.

06 Factored Networks

The phrase “component of a factored network” is essential.

A single retired phone motherboard should not be romanticized as a replacement for a modern data center server. That would be the wrong comparison. Data center servers have more cores, more memory, enterprise management tools, cooling design, reliability expectations, and maintenance pathways.

The value of a retired smartphone board appears when it becomes one component inside a factored network.

A factored network is a coordinated system in which work is divided into properly sized tasks and routed to compatible compute components. The system does not ask every component to do everything. It assigns the right work to the right capacity.

This is the difference between failure and reuse.

A retired phone board may be weak as a standalone server.

It may be strong as one node in a factored network.

The network restores value by matching workload to remaining capacity.

That means the design philosophy is not:

Old phones can replace modern servers.

The better design philosophy is:

Old phones can become useful compute factors when workloads are decomposed, routed, managed, and scheduled correctly.

Google Research describes the use of containerized applications managed by Kubernetes, with phones organized into self-managing clusters of 25 to 50 devices. It also states that UC San Diego’s early experiments showed that a 20-phone cluster could support peak submission rates for a class of more than 75 students, and that a 2,000-phone deployment could support about 100 such classes at once.

That is factored-network logic in practice.

The task is not forced onto one device.

The task is distributed.

The workload is routed.

The network does the restoration.

07 What The Article Misses

The common article-level summary is:

Old phones may become cloud servers.

That is interesting, but incomplete.

The deeper conclusion is:

New phones should be designed so that they can become cloud-server components later.

This is the missing design mandate.

If the industry now knows that retired phones can be stripped to their motherboards and clustered into useful compute systems, then future smartphones should not be built as if this second-life pathway is an afterthought.

They should be built for it.

The teardown should be easier.

The battery separation should be safer.

The board should be physically standardized.

The power input should be standardized.

The cooling interface should be standardized.

The firmware unlock pathway should be standardized.

The Linux recovery path should be standardized.

The mounting pattern should be standardized.

The board should be designed for rack conversion.

The manufacturer should publish second-life documentation.

The device should be born with a known second use.

That is the policy and design leap.

The current project proves that second-life computing is possible even when the original phones were not optimized for that purpose. The next generation should remove unnecessary friction.

A future smartphone should be designed as a dual-life object.

First life: smartphone.

Second life: compute-board node.

08 The Second-Life Compute Standard

This paper proposes the creation of a Second-Life Compute Standard for smartphones and comparable mobile devices.

The purpose of the standard would be to make end-of-life conversion safe, economical, interoperable, and scalable.

A serious standard should include at least the following requirements:

1. Safe and rapid battery removal or isolation.

2. Standardized compute-board form-factor zones.

3. Standard mounting points for second-life racks.

4. Standard board-level power input.

5. Standard cooling contact points.

6. Standard firmware unlock pathway after consumer retirement.

7. Secure wipe and chain-of-custody procedures.

8. Manufacturer-supported general-purpose Linux recovery mode.

9. Standardized cluster-management identity and attestation.

10. Published end-of-life board documentation.

11. Cross-manufacturer compatibility targets.

12. Clear labeling of second-life readiness at the device level.

13. Repair and teardown documentation available to certified recovery programs.

14. Modular board extraction designed into the original device.

15. Environmental reporting that distinguishes material recycling from functional reuse.

This would not require every phone to be identical. It would require every phone to expose enough common second-life interfaces that boards from different brands, factories, and generations could be processed into compatible compute systems.

That is the major opportunity.

A world with billions of smartphones should not depend on custom hacking and one-off teardown workflows to recover compute capacity.

Second-life conversion should become ordinary.

09 From E-Waste To Civic Compute Reserve

Global e-waste is already a major problem. The Global E-waste Monitor 2024 reported that 62 million tonnes of e-waste were generated in 2022, projected annual generation could reach 82 million tonnes by 2030, and documented collection and recycling is projected to fall to 20% by 2030 if current trends continue.

This means society is facing two related failures.

First, devices are being manufactured, consumed, retired, and discarded at enormous scale.

Second, many devices leave consumer use before every functional layer inside them has been exhausted.

Phone-cluster computing addresses both problems.

It extends the service life of the compute layer.

It reduces demand for newly manufactured hardware in appropriate workload classes.

It preserves some of the value of embodied carbon.

It creates lower-cost infrastructure.

It gives schools, universities, libraries, municipalities, and public institutions access to compute capacity that might otherwise sit unused.

That is why this should not be limited to corporate sustainability demonstrations.

It should become civic infrastructure policy.

A county could maintain a certified second-life compute program.

A university system could collect retired phones from students and alumni.

A public library network could host local services on standardized second-life clusters.

A workforce-development program could train technicians in secure teardown, testing, Linux conversion, cluster operation, and repair.

A state could create procurement incentives for second-life-ready phones.

A manufacturer could receive environmental credit not only for recycling materials, but for proving functional reuse.

The old phone in the drawer becomes part of a public compute reserve.

That is the Secretary Suite inversion.

10 Design For The Second Life, Not Only The First Sale

Most consumer devices are designed around the first sale.

The screen must look good.

The camera must sell.

The battery must last long enough.

The body must feel premium.

The operating system must support the current ecosystem.

The device must compete in the consumer market.

All of that matters.

But it is no longer enough.

A device with high embodied carbon and significant compute capability should also be judged by what happens after the first consumer role ends.

The manufacturer should have to answer:

Can the compute board be recovered?

Can it be securely wiped?

Can the battery be safely removed?

Can the board be powered in a standard rack?

Can the board be cooled?

Can Linux be installed?

Can the device identity be attested?

Can it join a cluster?

Can it be maintained?

Can it be documented?

Can it work with boards from other manufacturers?

Can it serve a second life before it becomes material recycling?

If the answer is no, the design is incomplete.

A phone that cannot be reasonably transitioned into second-life use is a less mature industrial object than a phone that can.

This is a design ethics issue.

It is also an infrastructure issue.

11 The Difference Between Refurbishment, Recycling, And Factored-Network Reuse

There are three major post-consumer pathways.

Refurbishment preserves the device as a device.

Recycling recovers material after device usefulness has ended.

Factored-network reuse recovers function by converting the device into a component of another system.

Each pathway matters.

If a phone can still serve another person as a phone, refurbishment may be best.

If the phone is physically destroyed or its board is no longer usable, recycling may be best.

But if the phone is no longer useful as a consumer phone while its board remains functional, factored-network reuse may be the higher-value pathway.

This third category is underdeveloped.

It deserves its own policy language.

A device should not be forced directly from “consumer phone” to “recycling stream” when a second functional role remains available.

That is like demolishing a structurally sound building because its first tenant moved out.

The right question is not merely whether the original use continues.

The right question is whether the structure can be rerouted.

12 Workloads That Make Sense

Second-life phone clusters should not be oversold.

They are not the answer to every computing need.

They are not replacements for large GPU training clusters.

They are not automatically appropriate for high-memory workloads, high-reliability enterprise services, or mission-critical systems without extensive validation.

Their value appears in the correct workload band.

Educational computing.

Introductory programming environments.

Grading systems.

Jupyter notebooks.

Small cloud services.

Lightweight web applications.

Parallel classroom tasks.

Research sandboxes.

Edge services.

Local civic dashboards.

Noncritical municipal tools.

Batch jobs that can tolerate node failure.

Disaster-resilient local compute.

Public-interest compute pools.

This is why factored-network thinking is important. The goal is not to pretend every component can do every job. The goal is to route suitable jobs to suitable components.

A strong second-life compute policy must therefore include workload classification.

Not all compute belongs on recycled phones.

But some compute absolutely can.

13 Security And Chain Of Custody

No second-life compute system can be credible without security.

Phones contain personal data.

A serious program must include secure wipe procedures, device-owner consent, chain-of-custody documentation, firmware verification, hardware testing, and board attestation.

The second-life cluster must be treated as infrastructure, not hobby equipment.

That means identity and trust matter.

A board entering the system should be wiped, tested, logged, certified, and assigned a trusted role.

A board leaving the system should be decommissioned, wiped again, and recycled through documented channels.

The same standard that makes hardware reusable should make misuse harder.

Security must be designed into the second-life pathway from the beginning.

14 Labor, Repair, And Workforce Development

The article-level discussion often mentions teardown labor as a cost problem.

That is true, but it is also an opportunity.

A standardized second-life phone design could create a new technical workforce pathway.

Students could learn electronics disassembly.

Technicians could learn secure data destruction.

Repair workers could become compute-recovery specialists.

Community colleges could teach second-life infrastructure operations.

Local governments could create circular-computing jobs.

Universities could build training labs around retired hardware.

Right-to-repair policy could connect directly to public compute policy.

This is important because the labor should not be invisible.

A phone does not magically become a server.

Someone removes the battery.

Someone tests the board.

Someone installs the operating system.

Someone mounts it.

Someone cools it.

Someone secures it.

Someone replaces failed nodes.

Someone manages the cluster.

If this is done badly, the system fails.

If this is done well, it becomes a new layer of technical employment and civic capacity.

15 Manufacturer Responsibility

Manufacturers should not be allowed to treat second-life recovery as someone else’s problem.

If a company profits from manufacturing millions of phones, it should help make those devices recoverable as functional components.

This does not mean every manufacturer must operate its own phone data centers.

It means manufacturers should design and document devices so that certified institutions can recover functional boards safely and efficiently.

A future environmental score for smartphones should include:

repairability,

battery replaceability,

software support duration,

material recyclability,

and second-life compute readiness.

That last category is new but necessary.

A phone that cannot be converted after consumer retirement has a design weakness.

A phone that can be converted has civic value built into it.

16 Public Procurement As Driver

Public procurement could accelerate this transition.

Schools, universities, governments, and agencies buy large numbers of phones and tablets. They could begin requiring second-life readiness in procurement contracts.

A state university system could say:

We prefer devices with certified second-life compute pathways.

A county government could say:

We require end-of-life board recovery documentation.

A school system could say:

Devices purchased with public funds must be eligible for conversion into educational compute clusters after retirement.

A federal program could support standards development.

This would shift the market.

Manufacturers respond to purchasing requirements.

If large buyers reward second-life readiness, second-life design will improve quickly.

17 The Civic Cloud Possibility

Second-life phone clusters should also be understood as part of a civic cloud strategy.

A civic cloud is not merely a government cloud vendor account.

It is local, regional, educational, emergency, and public-interest compute capacity designed to serve community needs.

Second-life phone clusters could support pieces of that system.

Local data projects.

Public dashboards.

School coding environments.

Small research tools.

Library services.

Local archival access.

Environmental monitoring.

Emergency communications support.

Municipal training environments.

Civic AI experiments too small to justify expensive cloud spending.

The goal is not to replace commercial cloud infrastructure.

The goal is to create a low-cost, low-carbon, locally understandable compute layer for appropriate tasks.

That matters because cloud dependence can become expensive and abstract.

A school that cannot afford endless cloud bills may still have access to retired hardware.

A local government that cannot build a full data center may still operate a small second-life cluster.

A rural community may benefit from edge compute that does not depend on every task traveling outward.

Second-life compute is not only a sustainability story.

It is also a decentralization story.

18 The Correct Lesson From Google And UC San Diego

The correct lesson is not:

Google found a clever use for old phones.

The correct lesson is:

The industry has been under-designing the end-of-life pathway for high-value compute devices.

Google and UC San Diego are showing what can be done despite that under-design.

The next generation should do it on purpose.

This is the Secretary Suite contribution to the discussion.

We do not need to claim that Google copied Secretary Suite. The stronger and more responsible claim is that Google’s project now publicly demonstrates a direction Secretary Suite has already argued for: obsolete hardware should be reclassified as functional infrastructure when its remaining capacity can be routed back into service.

The convergence matters.

It shows that the idea is not merely theoretical.

It is becoming visible.

19 The Core Principle

The core principle can be stated plainly:

A device should not be judged only by the role it is leaving. It should also be judged by the network it can still enter.

That principle applies beyond phones.

Old laptops.

Old desktops.

Old tablets.

Old routers.

Old batteries.

Old industrial sites.

Old buildings.

Old grid infrastructure.

Old vehicles.

Old public facilities.

Old machines.

The question is always:

What functional layer remains?

What network could receive it?

What standard would make transition easier?

What public value could be recovered before destruction?

Secretary Suite is fundamentally about this kind of reclassification.

It sees wasted capacity where the ordinary system sees expired objects.

20 Proposed Policy Statement

The following policy statement should guide future smartphone design:

Smartphones and comparable mobile devices should be designed as dual-life compute objects. Their first life is consumer use. Their second life should be standardized recovery of the compute board into a factored network for appropriate educational, civic, research, edge, or low-carbon cloud workloads. Manufacturers, regulators, public buyers, and standards bodies should collaborate to create second-life compute compatibility requirements that make this transition safe, secure, economical, and interoperable across brands and generations.

That is the entire paper in one paragraph.

21 Conclusion

The phone in the junk drawer is not merely a dead object.

It may be a misclassified compute component.

It may contain embodied carbon already paid for.

It may contain processing capacity that earlier generations would have considered extraordinary.

It may not be useful as a smartphone anymore, but it may still be useful as part of a factored network.

That is the design lesson.

Google and UC San Diego have shown that retired smartphones can become clustered computing infrastructure. But the larger lesson is that future smartphones should be designed for that destiny from the beginning.

The next generation of devices should not stumble into reuse.

They should be born with a second life.

A phone should not become architectural waste when its consumer shell expires.

It should become a standardized compute component in a factored network.

That is how society moves from recycling matter to recovering function.

That is how the junk drawer becomes infrastructure.

That is how waste becomes civic capacity.

That is the Secretary Suite idea.

References

Averill, Charles. “A Brief Analysis of the Apollo Guidance Computer.” arXiv, 2022.

Google Research. “A Low-Carbon Computing Platform From Your Retired Phones.” June 12, 2026.

Switzer, Jennifer; Marcano, Gabriel; Kastner, Ryan; Pannuto, Pat. “Junkyard Computing: Repurposing Discarded Smartphones to Minimize Carbon.” arXiv, 2021; revised 2022.

UNITAR / ITU. “The Global E-waste Monitor 2024.”