Gorilla Technology Grew Revenue 138%. The Stock Fell 13%. Here's the Disconnect.

Gorilla Technology Group just delivered 138% year-over-year revenue growth. The stock dropped 13%.

Wall Street plays a specific game with companies like GRRR. Revenue doubles, margins are still in the red, and the reflex is to sell — the assumption being the business model itself can't scale to profitability. That reaction makes sense if you're looking at the income statement headline. It doesn't hold up if you look at the cash flow, the contracted backlog, or the margin structure that's coming online.

The story here is simpler than the panic suggests: Gorilla is burning through a capital-intensive investment phase, and the math shows it has the cash, the contracts, and the margin trajectory to get to the other side. The question for investors isn't whether the revenue growth is real. It's whether they can stomach a company that's spending heavily now for gross margins management says will be 40% or higher by 2027.

Let's work through the numbers from the Q2 2026 results, released August 24.

Revenue didn't just grow — it accelerated. Q2 brought in $50.1 million, up 78% from Q1. That topped the company's own upgraded guidance of $44 million by 14% and beat external estimates of about $33 million by a wider margin. First-half revenue reached $78.4 million, nearly double the same period in 2025.

That's the part Wall Street acknowledged before moving on to what scared it.

The headline loss for H1 was $47.2 million on an IFRS operating basis. Here's what that number hides. The Q1 operating loss was $41.1 million — most of it one-time share-based compensation charges, with over 80% of H1's SBC recognized in the first quarter. Q2's operating loss came in at $2.2 million, a 95% reduction. That's the normalized run rate when you strip out the accounting front-load.

The real tell isn't the loss, though. It's the cash consumption. Operating cash usage in H1 2026 was $4.3 million — down 65% from $12.5 million, even as revenue nearly doubled. Cash consumption as a percentage of revenue dropped from 31.8% to 5.5%. That's not a business that's hemorrhaging. That's a business where the operating model is working and the capex is the load.

And the capex is the point. Gorilla spent $14.1 million on property and equipment in H1. Management said it would increase "many multiples" in H2 as data center equipment deliveries and site setups accelerate. The company ended June with $179.4 million in cash and is moving toward project finance and debt for the heavier lifts — the Indonesia deal, for instance, uses 70% debt funding with the rest from Gorilla's balance sheet and customer prepayments. Management expects operating cash flow to reach breakeven for the full year.

So what is that cash building? The answer explains why the old loss numbers are being replaced by a different story.

Gorilla is transitioning from selling video intelligence and IoT security platforms to deploying and operating AI data center infrastructure — specifically GPU-as-a-Service (GPUaaS), where the company owns the compute and charges customers for it under long-term contracts. The economics of the two models are fundamentally different. Hardware sales and deployment have low margins. GPUaaS contracts, management says, carry gross margins of 75% or more. They're fixed-price, take-or-pay, with customers committing to 100% utilization over five years.

Full-year gross margins are projected in the 40% range by 2027, as the service revenue scales past the hardware-heavy deployment phase. That's not a projection based on hope. It's the margin math of a contract structure where power, connectivity, and compute are locked in at known costs and priced at fixed rates.

The contracted revenue backing this transition is what separates Gorilla from a story that lives on promises.

The $1.4 billion Southeast Asia data center deal with Freyr Singapore — announced September 2025, initial $300 million phase starting in Q4 2025 — targets Indonesia, Malaysia, and Thailand. The initial phase is expected to generate $100 million in annualized revenue from 2026 through 2028. Gorilla also closed a $2 billion AI infrastructure supply arrangement with Supermicro for India, delivering over 25,000 NVIDIA B300 and B200 GPUs, along with networking equipment and related infrastructure.

Management guided to at least $200 million in full-year 2026 revenue, up from a previous range of $137 million to $200 million. Q3 is planned at $48 million to $50 million, Q4 at $60 million to $70 million. That $200 million figure is based on contracted revenue with firm delivery timing — it excludes later phases of NeutraDC in Indonesia, Yotta Phase 2 Batch 2 in India, and additional projects still in the pipeline. The total contracted backlog exceeds $5 billion. The broader pipeline is reported at over $7 billion.

For 2027, management targeted $450 million to $500 million in revenue, also from contracted sources. Yotta Phase 2 Batch 1 alone would add roughly $250 million. The NeutraDC first 300 servers in Indonesia, annualized at $75 million to $80 million, are scheduled for deployment by end of October 2026.

That's the revenue ramp. Now, the valuation context.

At roughly $16 per share before the post-earnings drop, Gorilla's market cap sat around $290 million. Analyst estimates call for FY2026 EPS of approximately $1.24, putting the forward P/E at about 16x. That's for a company guided to at least $200 million in revenue, with management targeting $450 million to $500 million in 2027 and gross margins expanding from current depressed levels toward 40% or higher.

If even the low end of that trajectory holds — $200 million in 2026 revenue, 40% gross margins, operating cash flow hitting breakeven — the stock at 16x forward earnings is pricing in a company that's barely profitable, not one transitioning to a 40% margin business on a $200 million revenue base.

That doesn't mean the thesis is guaranteed. Execution is the variable here.

Five parallel deployment sites, GPU supply constraints that affect the entire industry, data center power readiness, customer workload migration, and the financing of later-phase equipment — these aren't small operational details. They're the difference between a 2027 target of $500 million and a delayed version of the same number. Management acknowledged the risks on the call: hardware delivery timing, site readiness, customer acceptance. The company is ordering hardware early and gating deployment on confirmed power, but those are mitigations, not guarantees.

There's also the convertible note overhang. The $107 million of 7.5% senior unsecured convertible notes, due June 2031, carry an initial conversion price of approximately $25.48 — well above the current $16 range — but include reset provisions with a floor at $6 and a cap at $31.85. That's dilution that hangs over the stock if it climbs, and debt service that adds to the cash flow picture as the notes age.

Here's where the case lands.

Gorilla is not a company with a structural profitability problem. It's a company in a heavy investment phase, converting $5 billion in contracted backlog into GPUaaS revenue with margins that management says will be 75% plus on individual projects. Revenue doubled while operating cash consumption fell 65%. The operating loss narrowed 95% sequentially. Cash on hand is $179.4 million. Operating cash flow is guided to breakeven for the year.

The market reacted to the adjusted loss headline and the capex ramp the same way it reacts to any company spending heavily before it turns — it sold. That reaction treats a temporary investment phase as a business-model problem. The math, the contracts, and the margin trajectory don't support that conclusion. But the execution risk is real, and a delay on any one of five sites changes the timing, if not the direction, of the case.

At 16x forward earnings for a company targeting $200 million in 2026 and $450 million to $500 million in 2027 with 40% plus gross margins, Gorilla doesn't price in the revenue growth already embedded in its contract backlog. Whether that disconnect closes depends on hardware arriving on schedule and GPUs going live — not on the market's mood.