Solar Microcap AI Power Pivot

🌾 Welcome to StableBread's Newsletter!

Note: PowerBank reports in Canadian dollars under IFRS, so the financial figures here are in Canadian dollars. Share price, market cap, and valuation are in U.S. dollars (US$), the currency the stock trades in.

Without AI data centers, processing the vast datasets behind today's real-time AI applications would be impossible. But AI data centers are very power-hungry.

The challenge isn't necessarily generating that power, but getting it to the data center, since transmission and grid interconnection can take 5+ years in the U.S.

One overlooked company addressing that bottleneck is PowerBank Corporation (Nasdaq: SUUN), a Toronto-based solar developer with a ~US$35M market cap.

Its operating and late-stage projects across Upstate New York and Ontario already hold site control, permits, and grid interconnection positions, which is exactly what a data center now waits years to secure.

But SUUN's operations are ugly. Revenue is lumpy ($19.2M in Q1 FY2026, then negative $34,000 two quarters later), EBITDA is negative, the company burns cash, and it faces a Nasdaq delisting deadline this September.

Moreover, PowerBank keeps announcing into hot themes. It unveiled a Bitcoin treasury strategy in June 2025, and in November 2025 a collaboration with Smartlink AI, "Orbit AI," to put solar-powered AI compute in low-Earth orbit.

PowerBank's money is barely in the satellite project. The November announcement set its commitment at a US$50,000 initial investment, with an option to invest more, up to a 20% stake in Orbit AI for US$10M.

By the time the first satellite, Genesis-1, launched that December, PowerBank had elected not to make any additional investment, declining that option. Beyond that, PowerBank only says it "intends to contribute" solar and thermal components to future satellites.

The point is that these two initiatives have not produced any revenue.

The data center pivot is the newest of these moves. The fact sheet now splits the company into a "today" business (developer fees and IPP) and a "tomorrow" business it calls "Speed to Power for the Digital Economy."

The skeptical read is serial narrative-chasing. The optimistic read is that data centers are the one theme where PowerBank actually owns the scarce input, which is powered, interconnected land.

Still, the caveat stays attached. PowerBank's own investor deck says the company does not currently have any data center projects or customers, and may never obtain them. Contracted data center revenue today is zero.

At ~US$0.72/share, SUUN is a wager that a data center deal gets done:

I read through SUUN's filings, the 2026 investor deck, and the May 19, 2026 earnings call to evaluate whether the bet is worth taking.

PowerBank builds solar projects. It was incorporated in Ontario in 2013, runs out of Toronto, and trades as SUUN on Nasdaq, SUNN on Cboe Canada, and FRA:103 on the Frankfurt Stock Exchange.

Operations sit mostly in Upstate New York and Ontario, with a smaller presence in Nova Scotia.

Those regions have what data centers want, including land to build on, a cooler climate, and existing grid interconnection.

Across the nine months ended March 31, 2026, PowerBank booked $22.2M of revenue:

Development fees are the lumpy line. IPP is the small recurring base. Neither is the reason to own the stock. The data center pivot is.

PowerBank runs lean, just 18 employees, with strong energy credentials:

These are people who have permitted, financed, and grid-connected solar before, which is the valuable skill here.

None have run a data center, but they wouldn't need to, since PowerBank intends to host and power the sites, not operate the compute.

PowerBank's closing bid stayed below the US$1.00 Nasdaq minimum for 30 straight trading days between February 19 and April 1, 2026, tripping the exchange's listing rule.

It has until September 29, 2026 to close above US$1.00 for 10 straight days. Miss that, and it may get a second 180-day period on the Nasdaq Capital Market, though that isn't guaranteed.

At about US$0.72 and burning cash, climbing back over US$1.00 within four months is possible, but still uncertain. A reverse split is the likelier fix. That solves the listing but does nothing for the business.

The reported numbers are poor, and revenue is too lumpy to underwrite, as the last three quarters show:

That negative quarter has a cause. Qcells, a buyer PowerBank features as a "customer success," exercised a conditional right to sell two projects back to PowerBank (Grantsville and Highway 28), reversing ~$2.5M of booked revenue.

In other words, a completed customer sale turned out not to be final. That's a real risk to watch on future deals.

Here's the nine months through March 31, 2026:

The margin expansion is worth crediting. But the one line you can actually underwrite is IPP revenue, since it’s recurring.

It hit $9.3M in FY2025, up from $0.6M a year earlier, and ran $6.0M over the first nine months of FY2026.

That $9.3M wasn't built, it was bought. PowerBank's recurring revenue is essentially Solar Flow-Through Funds, a portfolio of more than 70 operating Ontario solar sites it acquired in July 2024 for $45M, paid entirely in stock, including 2.3M contingent value rights still to convert into shares.

The recurring line is real, but it was purchased, not developed. Everything else is project-based revenue that can swing from $19.2M in a quarter to nothing.

The balance sheet is no healthier than the income statement—heavily levered, low on cash, and steadily diluting.

Working capital turned positive, to +$10.7M at March 31, 2026 from a $1.8M deficit at the June 30, 2025 FY-end, but only because debt and stock sales funded it, not the business.

As of March 31, 2026, PowerBank had $135M of total assets against $106M of liabilities, leaving $29M of equity. Net debt of ~$67M runs well above the company's entire market value, with a debt-to-equity ratio of about 2.5x.

Management says the debt is mostly project-level, non-recourse financing tied to specific operating projects. That limits the claim on the parent if a project fails, but the company is still heavily levered.

Cash is the bigger worry. PowerBank ended the quarter with $11.3M of cash, and operations burned $11.4M over the nine months of FY2026.

That balance held up only because of outside money. Financing brought in $21.1M over the same nine months, part of it an ATM equity program.

Common shares have climbed to ~47M from ~40M at the end of 2025, with roughly 15M more in warrants, options, and convertible rights waiting to dilute further.

More raises are close to certain ($11.3M of cash roughly matches the $11.4M nine-month operating burn, leaving well under a year of runway), and each one dilutes holders.

The weak numbers don't make this a dead company.

PowerBank's "today" business, developing and building solar projects, selling most to third-party owners, and keeping a few as IPP assets, keeps running. The data center pivot is layered on top of it, not a replacement.

PowerBank has sold community solar portfolios to credible buyers, including a US$53M sale to an affiliate of Charley's Philly Steaks, more than US$41M to Honeywell, and US$23M to Hanwha Qcells.

But the Qcells reversal noted earlier is a reminder these sales aren't always final.

It's mobilizing nine New York projects totaling 42 MW of solar and 22 MWh of storage, and has 15 projects safe harbored to lock in ~US$65M of federal Investment Tax Credits, against an estimated construction value near US$168M.

Its SFF-06 battery project in Ontario reached commercial operation on April 20, 2026, adding a recurring IPP stream under a long-term contract with the Ontario system operator.

There's also a mandate letter with CIM Group for up to US$100M of project financing for a 97 MW New York portfolio, structured so no PowerBank shares are issued, though that deal isn't final yet.

PowerBank's 2026 deck lays out the demand case. It projects U.S. data center power demand growing from 80 GW in 2025 to 338 GW by 2030, lifting data centers from 6.5% to 11.7% of U.S. electricity sales.

The deck puts interconnection and substation upgrades at 5-10 years, the slowest part of data center development.

Lawrence Berkeley National Laboratory data shows the typical U.S. power project now waits almost five years from interconnection request to commercial operation, up from under two years in 2008.

So the scarce asset isn't the building or the GPUs. It's a site that can actually be powered on a useful timeline.

CEO Richard Lu on the May 19 call said the new wave of electricity demand is the data center market, and that PowerBank is positioned for it through its sites, its property-owner relationships, its work with 50+ municipal permitting authorities, and its utility interconnection experience.

PowerBank's fact sheet lists a pipeline of 1 GW+ of solar and ~1 GWh of battery storage.

The headline is real, but it reads like power available now, and it isn't. It bundles projects at very different stages, and only a small slice produces power today.

Here's the pipeline as of December 31, 2025, with the figures management refreshed on the May 2026 Q3 call shown in parentheses:

PowerBank has developed 100+ MW over its history but sold most of it to fund the business, which is why it owns and operates only 35 MWp today.

As of the May 2026 Q3 call, another ~220 MW of solar and battery is under construction or in advanced development, with interconnection and permits already behind it—the slow part of solar development.

The rest is early-stage, mostly land and queue positions that take years to build, and industry-wide most never reach construction. So the headline gigawatt overstates what PowerBank can realistically build.

First Real Step

On April 8, 2026, PowerBank signed a non-binding letter of intent with Nodiac Corp. to deploy distributed AI compute, described on the call as edge data centers, across its solar and BESS sites.

Nodiac builds modular data centers at sites that already have power and permits, putting compute online in months instead of years. It funds and operates the buildout while the site owner just contributes the site.

So PowerBank adds a revenue stream on assets it already owns, with no capital outlay.

It's the cleanest expression yet of the data center thesis, the first attempt to turn PowerBank's ~100 sites and permitting relationships into compute hosting.

But it's only an LOI. Nothing is built, no revenue is committed, the economics are unspecified, and Nodiac is itself a young company.

At US$0.72 and ~47M shares, PowerBank carries a US$35M market cap and a US$83M enterprise value, about 3.3x trailing sales.

That multiple doesn't mean much with a company this volatile. SUUN is better understood as a small bet on the data center pivot, with almost all the value riding on whether it works.

Valuing a company like PowerBank is difficult, so this is more of a back-of-the-envelope estimate. I'll walk through my assumptions below so you can understand them and adjust as needed.