$DUOT Duos Technologies: The Story Is No Longer Rail, But The Proof Still Has To Show Up

I have followed Duos Technologies for a long time.

This was originally a rail technology story. The company had what looked like an impressive railcar inspection platform, and I believed the technology had real value. The problem was never that the product did not make sense. The problem was that rail is slow. Very slow. Adoption cycles were long, revenue was lumpy, and the business never scaled the way many long term shareholders hoped it would.

That history matters.

Investors should not pretend the company is starting with a clean slate. Duos already had one big vision that did not fully turn into the business shareholders expected. So the burden of proof is higher this time.

But I also think this pivot is very different.

Duos is no longer trying to sell rail inspection technology into a slow moving industry. The company is now repositioning itself around AI infrastructure, modular Edge Data Centers, GPU as a Service, high power colocation, and Technology Solutions.

That is a very different market.

Rail was slow because the customer base was slow.

AI infrastructure is dealing with the opposite problem.

Demand is moving faster than supply.

# Why Doug Recker Matters

One of the reasons I take this pivot seriously is Doug Recker.

This is not a random rail company suddenly chasing an AI buzzword. Recker has real experience in data centers, edge infrastructure, and colocation. He founded Colo5 Data Centers, which was later acquired by Cologix, and he founded EdgePresence, which was acquired by Ubiquity.

That matters because the new Duos strategy is directly in his lane.

It also matters that Recker was already familiar with Duos before this pivot. Duos had edge compute needs inside its rail inspection business. The rail portals produced massive amounts of visual data, and that data had to be processed close to the source. That was already an edge compute problem.

So the pivot is not as random as it may look from the outside.

The old rail business did not scale, but it did give Duos experience operating physical technology infrastructure in the field. Recker brought the missing piece, which is how to turn that edge infrastructure concept into a broader data center business.

That does not guarantee success.

But it makes the pivot much more credible.

This is not a rail guy pretending to understand data centers. This is now a data center operator trying to use Duos as the platform to scale.

# Q1 Was Not The Thesis

The first quarter numbers looked weak on the surface. Revenue fell sharply, mostly because the Asset Management Agreement with New APR Energy ramped down faster than expected.

That bothered me at first.

But the AMA decline makes more sense now given the potential APR transaction backdrop. If APR is being acquired or repositioned, then it makes sense that the service agreement with Duos would wind down quickly.

Management was clear that the company is moving away from the old structure. Doug said Q1 reflected the “continued execution” of the company’s transformation toward a data center focused platform, while also confirming that the AMA will conclude later this year and that Duos will retain its 5 percent equity stake in APR’s parent.

I am not modeling value from that APR stake yet because management made clear there is a waterfall structure. But it is potential optionality.

More importantly, Q1 revenue was never the reason to own the stock.

The real question is whether the new businesses can replace the old bridge and scale fast enough.

That means Hydra, Tech Solutions, high power colocation, and future funding.

# Hydra Is The First Major Proof Point

Hydra is the biggest near term test.

The Hydra agreement involves 2,304 NVIDIA GPUs and is expected to generate roughly $176 million over 36 months. Management now expects about $26 million of revenue from this project in 2026, mostly in the second half.

That is important.

But what matters even more is execution.

On the latest call, Doug said Super Micro and NVIDIA have what they need, rack and stack is happening, and the timeline may move from August to July. His exact words were, “Instead of August, July 1. So everything is pointing in that direction.”

That changes the tone.

Before the call, the question was whether August was still real.

Now it sounds like they may be a full month ahead of schedule.

Hydra is not just one deal. It is the first real proof point that Duos can deploy high density AI infrastructure on schedule.

If Hydra works, the entire story becomes more credible.

If it slips, the market will question the model.

# High Power Colocation May Be The Cleaner Long Term Model

The first Hydra deal has Duos buying the GPUs. That creates big revenue potential, but it is also capital intensive.

The cleaner long term model is high power colocation.

In that model, the customer brings the GPUs and Duos provides the infrastructure.

Power.

Cooling.

Connectivity.

Backup power.

Reliability.

Uptime.

That is a much cleaner model because it reduces GPU ownership risk and makes the business easier to finance.

Doug made this point directly on the call. He said the first Hydra deployment involved Duos buying the GPUs, but added, “That’s not our model going forward.” He then explained that Hydra has more than 12 customers interested in 5MW or 10MW deployments where the customers already have GPUs and need a place to deploy them.

That matters.

The first Hydra deal proves Duos can participate in GPU infrastructure. The next version of the model may be less capital intensive and more repeatable.

# Tech Solutions Is The Bridge

The second major piece is Technology Solutions.

This business helps Duos source equipment more efficiently for its own builds, but it also creates revenue by serving other data center operators and infrastructure customers.

In Q1, Tech Solutions added 8 large data center customers and increased backlog to about $14 million, all expected to ship and invoice in 2026. Management also said the pipeline is much larger than current backlog.

That is important because AMA revenue is winding down.

Something has to bridge the gap.

Tech Solutions may be that bridge.

Doug said the company expects Technology Solutions revenue to not only replace the New APR AMA revenue, but also provide better margins.

It is not as high quality as long term colocation revenue, but it can scale faster, requires less capital, and gives Duos a way to participate in the data center supply chain while Hydra and the EDC model ramp.

# The 5MW To 10MW Opportunity Is The Interesting Part

The market is obsessed with giant AI campuses.

100MW.

500MW.

Gigawatt scale.

Duos is focused on a different part of the market.

The company is targeting smaller high power deployments in the 5MW to 10MW range. These sites can be deployed faster and can support inference workloads closer to where data is created and consumed.

That is the key idea.

Duos is not trying to beat Google, Microsoft, Amazon, or the other hyperscalers.

It is trying to help serve demand that cannot wait years for traditional data center capacity.

Doug’s commentary on the call was very direct. He said the 5MW to 10MW sector is becoming attractive because companies need to deploy GPUs quickly and Duos can bring those sites online much faster than traditional builds. He also said, “I have 21 neoclouds. If I had 5 or 10 meg, they would take it.”

That is promotional, but it also fits what we are seeing across the broader AI infrastructure market.

The opportunity is speed.

The opportunity is stranded power.

The opportunity is getting capacity online quickly.

That is why this is interesting.

# The Economics Are Attractive If They Hold

Management has laid out very attractive unit economics.

Doug said high power capacity costs about $6.5 million per MW to build, with revenue a little under $2 million per MW per year. He also said colocation contracts are typically 5 to 10 years, while management said high power colocation and GPU as a Service margins are around 80 percent.

If those numbers hold, this becomes an infrastructure asset story with recurring revenue.

But those numbers still need to be proven.

That is why the next few quarters matter so much.

# Future Funding Could Become Much Cleaner

This business is capital intensive, so funding matters.

The good news is that management does not appear to want to keep using equity as the primary growth engine.

Doug was very clear on the call: “I don’t want to go out for any more equity.” He wants to get in front of strategic partners like NVIDIA, Super Micro, Dell, and others that may be willing to help backstop future deployments.

That is the right playbook.

Use equity to prove the model.

Get Hydra live.

Get another high power site deployed.

Show contracted recurring revenue.

Then fund future growth through debt or strategic capital.

If Duos can prove that its high power EDCs produce long term contracted cash flow, then the business should become financeable with debt. These are physical infrastructure assets with customers, power access, equipment, and recurring revenue.

That is very different from funding a speculative software idea.

This is why execution over the next few quarters matters so much.

If Hydra goes live on time and the next sites follow, dilution risk should come down.

If they cannot prove the model quickly enough, then equity risk remains.

So future funding is not solved yet.

But the path is clear.

Prove the assets, then finance growth with debt or strategic partners instead of more dilution.

# Google Is The Market Signal

One of the reasons I like this setup is that Google recently confirmed the demand side of the thesis.

Google said it is compute constrained. It also talked about demand for AI infrastructure and even discussed putting TPUs into customer data centers.

That matters.

It does not mean Google is a Duos customer.

But it validates the market problem Duos is trying to solve.

Doug even referenced Google on the call, saying he is watching Google closely because he would like to tell them, “we can do this for you.” His point was not that Google is a customer today. His point was that large players need distributed inference infrastructure, and Duos wants to be positioned as a faster deployment partner.

That is the opportunity.

Duos does not need to create the market.

It needs to capture a small piece of an existing supply constrained market.

# What Still Bothers Me

I am interested in the setup, but I do not think everything is proven.

Hosting revenue was only $30,000 in Q1. That is very light given the number of deployed EDCs and the recent open houses. I understand that deployment may be ahead of monetization, but revenue eventually needs to show up.

The company also needs to make the revenue bridge cleaner.

There are too many moving pieces right now.

Hydra.

Tech Solutions.

Bookings.

Backlog.

Deposits.

Deferred revenue.

Colocation.

Hosting.

AMA.

APR optionality.

Investors need a clearer segment table so they can understand what is booked, what is expected to convert this year, what is recurring, and what margins look like by business line.

Funding is another major watch item.

Doug said he does not want to keep funding growth with equity. That is exactly what shareholders want to hear. The next step is proving the model enough to access debt or strategic capital.

If Duos can move from equity funded proof of concept to debt or strategic funded growth, the model becomes much more attractive.

If not, dilution remains a risk.

# What I Am Watching Next

For me, the checklist is simple.

Hydra needs to go live by July or August.

Tech Solutions backlog needs to convert into revenue.

High power colocation needs to keep building.

Hosting revenue needs to move higher.

Rail needs to be divested.

The company needs to secure cleaner funding for future MW growth.

APR optionality needs to be clarified.

And management needs to improve revenue disclosure.

# Bottom Line

Duos Technologies is not a clean story.

There is baggage from the rail business. The current revenue base is messy. AMA is winding down. Hosting revenue is still tiny. The model is capital intensive. The revenue bridge needs to be clearer.

But the opportunity is real.

Duos is no longer just talking about a pivot. They have deployed EDCs, a major Hydra contract, customer cash, growing Tech Solutions backlog, 10MW contracted, and a target of 25MW this year.

The market may still be trying to figure out what Duos is.

That is where the opportunity may be.

This is not a bet on Q1 revenue.

It is not a bet on the old rail business.

It is a bet on whether Duos can execute in one of the most supply constrained areas of the market.

AI infrastructure.

If Hydra goes live on time, Tech Solutions converts backlog, and future growth gets funded without heavy dilution, the market may start valuing Duos very differently.

That is the thesis.

Execution.