Stock Analysis · Bulls vs Bears

Gorilla Technology Group GRRR

🐂 Bulls vs Bears 🐻Report #19 Oct 2026
★ Willem owns this stock · 1.9% of the portfolioOlder version — see the latest →
76
very high risk
🐂 6vs9 🐻
Bears lead
45%
▲ Chance of the biggest tailwind the forecasts miss
65%
▼ Chance of the biggest headwind the forecasts miss
The verdict

GRRR is a highly levered bet that a $100M-revenue integrator can turn about $5B of headline GPU contracts into rental income, financed mostly with other people's money and delivered through counterparties that are themselves stretched. The direction of travel is real (contracts signed, hardware ordered, Fitch rating), but every link (margin, financing, counterparty credit and delivery dates) is still unproven, so…

The Bulls vs Bears podcast is being made
Two AI hosts debate this report, NotebookLM-style. It appears here automatically within a few hours.

Written by Claude (Anthropic’s AI) from public sources · Willem owns GRRR · AI can make mistakes — numbers in brackets point to the sources · Not financial advice

🐂 Bull arguments 6

  1. 1A contracted step-change in revenue scaleMedium
  2. 2Scarce GPUs and power in India and Southeast Asia, with Nvidia in the chainMedium
  3. 3Proven access to capital, now with a credit ratingMedium
  4. 4The legacy Egypt contract funds the transition and is payingMedium
  5. 5Revenue guidance has a track record of being metWeak
  6. 6Optionality in Korat and the uncommitted parts of the dealsWeak

🐻 Bear arguments 9

  1. 1The high-margin rental business has not started, and current margins are near zeroHigh
  2. 2Counterparty and concentration riskHigh
  3. 3Dilution and financing overhangHigh
  4. 4Headline numbers don't reconcile with the accountsMedium
  5. 5Execution capacityMedium
  6. 6Governance and credibility flagsMedium
  7. 7GPU obsolescence and cost inflationMedium
  8. 8Egyptian pound and payment risk on the legacy contractLow
  9. 9Export controls and geopoliticsLow

Bull arguments show how well they survive the devil’s advocate (strong / medium / weak); bear arguments show the damage they could do after the counter-argument (high / medium / low).

Bull case — all positive points (most important first)

🐂 Bull #1A contracted step-change in revenue scale

  • The point: If the signed GPU contracts are delivered, revenue could grow about 4–5x in two years. The company is valued at less than one year of the 2027 guided revenue.
  • Evidence:
    • FY2027 revenue guidance is $450–500M, first given 2026-08-24 [3].
    • Management says it includes only contracted revenue with firm dates: Yotta 1 at about $100M a year, the first Yotta 2 batch at about $250M, and the first 300 Batam servers at about $75–80M a year (call summary) [4].
    • It excludes about 1,575 further Batam servers and part of Yotta 2 [4].
    • The Batam contract is described as "about $2.5bn" over 5 years [7].
    • Market value is about $345M (about $12 × about 28.8M shares) [22,18].
  • Why it matters: five-year fixed-fee rentals at high gross margins would turn a lumpy project business into a recurring one with visible cash flow.
  • 😈 Devil's advocate: the "contracted" figures are company-stated, and the counterparties are unnamed or unlisted.
    • The IFRS note on "unfulfilled long-term contracts" shows only $72.7M at 2026-06-30, down from $101.3M [3], against a claimed "signed backlog over $5 billion" (Q1 call summary) [4].
    • Headline values have shifted ($2.8bn to $2.5bn for Yotta 2) [6,21].
    • Freyr's $1.4bn has gone quiet [1,21].
    • Revenue timing has already slipped by about a quarter [18].
  • Verdict: Medium The contracts appear real, and hardware is being installed, but how much of the headline value becomes Gorilla revenue on time is unproven.

🐂 Bull #2Scarce GPUs and power in India and Southeast Asia, with Nvidia in the chain

  • The point: demand for AI compute in the region exceeds supply, so whoever holds installed GPUs with power can sign long, fixed-fee contracts.
  • Evidence:
    • Fitch (via the company's release) says power and GPU availability, not demand, limits growth in South and Southeast Asia [9].
    • Nvidia is expected to take about half the output of the Yotta 2 tranche under a four-year commitment tied to DGX Cloud (Nvidia's own rented-cloud service) [6].
    • Posted rental prices for the newest GPUs roughly doubled over the year (an aggregator index) [23].
    • Indonesia's Danantara (the sovereign wealth fund) and Telkom mapped up to 1 GW of state AI data-centre demand [24].
  • Why it matters: scarcity supports fixed-fee terms and reduces the risk of empty capacity.
  • 😈 Devil's advocate: scarcity attracts supply.
    • Firmus plans 170,000 GPUs on a 360 MW Batam campus from 2027–2028, and Singtel and others offer GPUaaS [19].
    • GPUs lose value fast as each Nvidia generation arrives. Gorilla is only protected if the customer is locked in and solvent for all five years.
    • Nvidia's role is with Yotta, not directly with Gorilla, and how Gorilla's revenue maps to Nvidia's offtake is undisclosed [6].
  • Verdict: Medium The demand backdrop is real today, but it protects Gorilla only through contract terms that are not public.

🐂 Bull #3Proven access to capital, now with a credit rating

  • The point: for a middleman, the moat is capital. Gorilla has raised it on better-than-feared terms.
  • Evidence:
    • It raised $107M of 7.5% convertibles in June 2026 and $125M in July 2026, with an initial conversion price of about $25.48, a 52% premium to the 2026-07-14 close of $16.77 [8,25].
    • Fitch assigned a first-time B- rating with a Positive outlook (2026-08-27). Fitch projects cash above $100M and leverage falling to 1.3x by 2030 [9].
    • Cash was $179.4M at 2026-06-30, before the July $125M [3].
    • Directors and officers accepted lock-ups [8].
  • Why it matters: each GPU tranche costs hundreds of millions of dollars. Without cheap capital the contracts cannot be delivered.
  • 😈 Devil's advocate:
    • B- is deep junk. Fitch calls the financial policy "aggressive", with no leverage target, and expects only $34M of parent-level EBITDA in 2027 [9].
    • The conversion price can reset down to $8.00 [25], so the "premium" is not locked in.
    • The project debt for Batam and Yotta 2 was still "offers" in August, and management said some initial offers were unfavourable (call summary) [4].
  • Verdict: Medium Access is proven at the equity-linked level but not yet at the project-debt level, where the large money is.

🐂 Bull #4The legacy Egypt contract funds the transition and is paying

  • The point: the Egypt build gives a cash base and proof that Gorilla can deliver a large government project.
  • Evidence:
    • Egypt revenue was $52.3M (2023), $59.1M (2024) and $77.5M (2025) [1].
    • More than $22M was collected from the largest customers in Jan–Feb 2026 [26].
    • Advance-payment guarantees were released in Q1 2026, and Q1 operating cash flow was +$6.6M [15].
    • Management expects about five years of recurring revenue after completion (call summary) [4].
    • The new Taiwan police contract (2026-09-22) shows the government franchise is alive [12].
  • Why it matters: cash from Egypt reduces how much new capital the GPU push needs.
  • 😈 Devil's advocate:
    • The contract is priced in a currency that lost about 11.5% on average in 2025. FX losses were $21.0M in 2025 and $18.9M in Q1 2026 [1,15].
    • Overdue receivables of 91–365 days were about $18.6M at June 2026 [3].
    • The contract ends around mid-2027 (four-year term, my inference) [1].
    • Egypt can terminate for delays of more than two months [1].
  • Verdict: Medium It is a real asset, but it is shrinking and FX-exposed, and it cannot fund GPU capex on its own.

🐂 Bull #5Revenue guidance has a track record of being met

  • The point: management has hit its revenue numbers for two years.
  • Evidence: FY2024 beat guidance [20], FY2025 landed in the raised range [1], Q2 2026 beat its pre-announcement [3], and FY2026 guidance was raised twice to at least $200M [3].
  • Why it matters: it gives some credibility to the FY2027 $450–500M guide.
  • 😈 Devil's advocate:
    • Revenue was met by adding low-margin hardware. Adjusted EBITDA swung from +$6.2M (H1 2025) to -$14.6M (H1 2026) [3].
    • The H1 2026 figures are "unaudited and unreviewed" [3].
    • Operational timelines (Yotta 1, Batam) have already slipped [18,21].
  • Verdict: Weak Hitting revenue numbers with pass-through hardware says little about delivering high-margin rental income.

🐂 Bull #6Optionality in Korat and the uncommitted parts of the deals

  • The point: several projects sit outside guidance and could add upside.
  • Evidence:
    • Korat: land bought for 200 MW. The company calls $16–18bn of illustrative five-year revenue "potential rather than contracted" [21,11].
    • Not in the 2027 guide: the rest of Batam phase 1, the 875-server Batam phase 2, and part of Yotta 2 [4].
  • Why it matters: if even part converts, the 2027–2028 numbers rise.
  • 😈 Devil's advocate:
    • Korat has no customer, no financing and no confirmed construction start [11].
    • Gorilla's Korat deck is labelled "v0.1" [11].
    • Each new tranche needs more capital, and therefore more dilution or debt.
  • Verdict: Weak It is real optionality but expensive to exercise, and it is speculative today.

Bear case — all negative points (most important first)

🐻 Bear #1The high-margin rental business has not started, and current margins are near zero

  • The point: the investment case rests on 75% gross margins that have never been reported, while today's business loses money at the gross level.
  • Evidence:
    • H1 2026 gross margin was 4.9%. Q2 had a gross loss of about $2.1M on $50.1M of revenue [3,15].
    • Hardware was $52.5M of H1 revenue [3].
    • The 40%+ 2027 gross margin is "not official guidance" (call summary) [4].
    • Memory costs are up 40–60%, and vendor prices are fixed only through December [4].
    • The IFRS operating loss was $47.2M in H1, including $25.4M of stock compensation [3].
  • Why it matters: if the GPU contracts are partly hardware resale (sold to Yotta, for example) rather than rentals, Gorilla stays a thin-margin reseller with big working-capital needs.
  • 🛡️ Counter-argument:
    • Management attributes the margin collapse to "mobilization" hardware ahead of rentals.
    • GPUaaS revenue only starts on customer acceptance, from Q4 2026 [4].
    • Fitch's analysis assumes take-or-pay SPV revenues [9].
  • Severity: High Until Q4 2026 or Q1 2027 shows rental revenue with high margins, the core thesis is unproven.

🐻 Bear #2Counterparty and concentration risk

  • The point: Gorilla depends on a handful of counterparties, several of them financially stretched or anonymous.
  • Evidence:
    • Egypt was 76.5% of FY2025 revenue [1].
    • Yotta: about $108M of FY25 revenue against about $1.1bn of net debt; it relies on an IPO in Jan–Mar 2027 to repay debt and buy GPUs; it has ordered tens of thousands more GPUs [14,13,16].
    • The Batam end-customer is unnamed [7].
    • Telkom is selling a majority of NeutraDC, the Batam landlord, with a target around end-2026 [24].
    • In H1 2026, two customers each held more than 10% of receivables plus contract assets [3].
  • Why it matters: a five-year fixed fee is worth only as much as the payer's ability to pay for five years.
  • 🛡️ Counter-argument:
    • Nvidia's expected offtake of about half the Yotta 2 output supports Yotta's revenue [6].
    • The Batam customer is described as "high investment-grade" [7].
    • Yotta raised $150M at about a $3.9bn valuation [13].
  • Severity: High Two of the three big GPU programmes run through one levered, unlisted Indian company.

🐻 Bear #3Dilution and financing overhang

  • The point: the plan needs far more capital than Gorilla has, and its financing tools dilute shareholders.
  • Evidence:
    • Share count: 18.06M (2025-01-01), 26.19M (2026-01-01) and about 28.8M (August 2026) [3,18].
    • $232M of convertibles: about 9.1M shares at the $25.48 conversion price, against a share price of about $12 [8,25,22].
    • Series B (the $125M July notes) can reset down to an $8.00 floor, which would mean about 15.6M shares for that series alone (my arithmetic) [25]. The June notes' floor is not verified; one note says $6 [17].
    • Interest can be paid in shares [25].
    • Committed equipment purchases were $228M [3].
    • Stock compensation was $25.4M in H1 2026, of which $22.0M went to key management [3].
    • As a Cayman foreign private issuer it can issue shares without a shareholder vote [27].
  • Why it matters: even if the business succeeds, value per share can be diluted away.
  • 🛡️ Counter-argument:
    • Management plans project-level debt for about 70% of costs, so most capex need not be equity [7].
    • The convertibles carry an upward reset too [25].
    • Buybacks of about $6.5M were made in 2025 and early 2026 [26].
  • Severity: High A downward reset at today's price would be the single largest per-share hit.

🐻 Bear #4Headline numbers don't reconcile with the accounts

  • The point: the "backlog" language far outruns what the audited accounts recognise as contracted.
  • Evidence:
    • "Signed backlog over $5 billion" (Q1 call summary) against IFRS unfulfilled long-term contracts of $72.7M at 2026-06-30 [4,3].
    • The pipeline has grown from $6bn (Mar 2025) to more than $7bn (Dec 2025) to about $10bn (2026, RedChip webinar title) [28].
    • The $2B Supermicro and $2.8bn Yotta figures are procurement or project values, not Gorilla revenue [5,6].
    • Freyr's $1.4bn produced no 2025 revenue, and no update was found in 2026 [1,21].
  • Why it matters: investors may be valuing headline totals that never become revenue.
  • 🛡️ Counter-argument: GPUaaS agreements may not yet qualify as IFRS contracts until delivery and acceptance, so the gap may be presentation, not substance. Fitch's committee reviewed the contracts [9].
  • Severity: Medium High. Unexplained until management reconciles it.

🐻 Bear #5Execution capacity

  • The point: a firm that delivered about $100M of revenue a year must install and run tens of thousands of top-end GPUs across three countries at once.
  • Evidence:
    • Yotta 1 went from "by end of August… ahead of internal schedules" on 2026-07-21 to about November [21,18].
    • Batam operations moved from Q3 to Q4 [18].
    • Headcount is to go from about 650 (about 300 full-time) to about 2,000 by mid-2027 (call summary) [4].
    • The CEO cites power ("electricity seems to be of a bit of a problem"), GPUs, memory and cabling shortages [4].
  • Why it matters: every month of delay is a month of depreciation and interest without rental income.
  • 🛡️ Counter-argument: the slips so far are about one quarter, and hardware is physically being installed (Yotta "physical execution", Sept 2026) [22].
  • Severity: Medium High.

🐻 Bear #6Governance and credibility flags

  • The point: several governance signals call for caution about the numbers and incentives.
  • Evidence:
    • Auditor Marcum Asia was dismissed on 2026-04-30, two weeks after signing the FY2025 20-F, and replaced by UHY. No reason is given; no disagreements are reported [29].
    • The interim accounts are "unaudited and unreviewed" [3].
    • A $20.9M Q1 2026 stock-compensation charge relates partly to services before 2025 (call summary) [4].
    • Key management pay was $5.9M in FY2025, up from $2.3M [1].
    • The CEO was formerly a co-founder of Shackleton Ventures, the parent of Shackleton Finance, which Gorilla acquired in 2026. The release does not disclose the link [1,30].
    • Short reports in 2025: The Bear Cave (2025-03-06) and Citron (2025-03-17). Law firms announced "investigations"; no filed class action was found [31,32].
    • Gorilla is suing its own PIPE investors (investors who bought shares privately at the SPAC listing) for alleged share manipulation, and is itself being sued over earn-out shares [1].
  • Why it matters: when the investment case rests on management's descriptions of unpublished contracts, trust in disclosure is the whole game.
  • 🛡️ Counter-argument:
    • Marcum's opinions were clean, and there were no reportable events [29].
    • Management reported effective internal controls [1].
    • No regulator has acted on the 2025 short claims.
  • Severity: Medium

🐻 Bear #7GPU obsolescence and cost inflation

  • The point: GPUs bought in 2026 must earn their return before newer chips (Nvidia Vera Rubin, from 2026–27) cut rental prices.
  • Evidence:
    • Management assumes 20–25% resale value after five years [4].
    • Hardware price locks run only through December, and memory costs are up 40–60% [4].
    • Yotta itself is ordering Vera Rubin GPUs [13].
  • Why it matters: if a customer renegotiates or walks away in year 2–3, the chips may be worth far less than the debt against them.
  • 🛡️ Counter-argument: take-or-pay, five-year flat fees shift price risk to the customer, if they are enforceable [4,9].
  • Severity: Medium

🐻 Bear #8Egyptian pound and payment risk on the legacy contract

  • The point: the fixed EGP price shrinks in dollars whenever the pound weakens.
  • Evidence: FX losses were $27.8M (2024), $21.0M (2025) and $18.9M (Q1 2026). Q2 saw a partial reversal as the pound rallied [1,15,18]. Overdue receivables are rising [3].
  • Why it matters: this is still most of the gross profit base until GPU rentals start.
  • 🛡️ Counter-argument: it is a declining share of the future business, and Q2 showed the swing can reverse.
  • Severity: Low Medium.

🐻 Bear #9Export controls and geopolitics

  • The point: US rules on AI chips, and Taiwan–China tension, sit over the whole model.
  • Evidence:
    • BIS (the US Commerce Department's export-control bureau) has not replaced the rescinded AI Diffusion Rule.
    • Licence requirements still apply to entities with a China-linked parent [33].
    • The Chip Security Act (location verification of AI chips) is pending in Congress [34].
    • Gorilla's management and staff are mostly in Taiwan [1].
  • Why it matters: a new rule on India or Southeast Asia could delay GPU deliveries.
  • 🛡️ Counter-argument: India and Indonesia face no country-specific licence requirement found today, and Gorilla says it does not sell to or host in China [1].
  • Severity: Low Medium.

Tailwinds NOT in the forecasts

  • Remaining Batam servers scheduled and financed — 45% chance within 36 months · Impact if it happens: High

    • What it is: about 1,575 servers (the rest of phase 1 plus the 875-server phase 2) get firm delivery dates and closed project debt.
    • Why it's not priced in: management explicitly excludes them from the $450–500M 2027 guide [4].
    • What would trigger it / the sign to watch: a 6-K announcing a closed Batam project-finance facility, and the first tranche accepted in Q4 2026.
    • Why this probability: the contract is signed, but financing, landlord ownership change and a new competing Firmus campus all add friction, so it is a bit worse than a coin-flip.

  • A named anchor customer for Korat — 22% chance · Impact: High

    • What it is: a signed long-term offtake for at least one 30 MW hall.
    • Why it's not priced in: the company itself calls Korat "potential rather than contracted" [21].
    • What would trigger it / the sign to watch: a customer announcement with financing, Thai BOI (Board of Investment) approval, and a power confirmation.
    • Why this probability: there are no customer, financing or construction proof points after five months, and Gorilla is already capital-constrained on its two signed programmes.

  • Yotta IPO succeeds and de-risks the counterparty — 55% chance · Impact: Medium

    • What it is: Yotta lists in Jan–Mar 2027 and raises up to $1.5bn, partly to repay debt [13].
    • Why it's not priced in: consensus treats Yotta revenue as given, but the market discounts GRRR heavily (about 0.7x 2027 guided revenue). A stronger counterparty would narrow that.
    • What would trigger it / the sign to watch: Yotta's draft IPO papers (DRHP) filing, expected October 2026, and their disclosure of the Gorilla contracts.
    • Why this probability: Indian IPO markets are open and Yotta has raised pre-IPO money, but valuations reported range from $3.9bn to $6bn and the timing has already slipped once.

  • Fitch upgrade to B / cheaper project debt — 35% chance · Impact: Medium

    • What it is: the Positive outlook converts into an upgrade, lowering the cost of the project debt.
    • Why it's not priced in: consensus models do not separate financing cost by rating.
    • What would trigger it / the sign to watch: first GPUaaS cash flows and SPV distributions.
    • Why this probability: Fitch flagged an upgrade path, but its projections already assume delivery that has slipped.

Headwinds NOT in the forecasts

  • Another dilutive raise, or a downward reset of the convertible conversion price — 65% chance within 12 months · Impact if it happens: High

    • What it is: new equity, new converts, or the reset clause moving toward the $8 floor (the share price is about $12 against $25.48).
    • Why it's not priced in: broker targets ($24–44) are set per share on the current count [22,18].
    • What would trigger it / the sign to watch: the reset-date mechanics in the indenture (not yet read), an F-3 or new 6-K financing, and Yotta 2 equity needs.
    • Why this probability: $228M of commitments, Korat ambitions, and a history of a raise roughly every 12 months make further dilution more likely than not.

  • 2027 revenue falls more than 20% short of the $450–500M guide — 45% chance · Impact: High

    • What it is: delivery, acceptance or financing delays push rental revenue into 2028.
    • Why it's not priced in: consensus (Zacks $436M, Compass Point $1.2bn) sits at or above guidance [18,35].
    • What would trigger it / the sign to watch: Yotta 1 billing not visible in Q4 2026, or Batam's first 300 servers not live by year-end.
    • Why this probability: two slips have already occurred, but the revenue items in the guide are the most advanced programmes.

  • Yotta renegotiates, delays or partly defaults — 20% chance · Impact: High

    • What it is: Yotta, levered and IPO-dependent, slows payments or reshapes its contracts.
    • Why it's not priced in: guidance treats Yotta as about $350M of 2027 revenue [4].
    • What would trigger it / the sign to watch: an IPO delay beyond mid-2027, Indian press on Yotta's funding, or Gorilla receivables from India rising.
    • Why this probability: Nvidia's offtake and the pre-IPO raise support Yotta, but its debt is about 10x its FY25 revenue.

  • Batam landlord change of control disrupts the project — 15% chance · Impact: Medium

    • What it is: a new NeutraDC owner reprices or reprioritises capacity.
    • Why it's not priced in: this is not discussed in company materials.
    • What would trigger it / the sign to watch: the announcement of Telkom's buyer [24].
    • Why this probability: capacity contracts usually survive a change of ownership, but power allocation at an 18 MW site is tight.

Investment Risk Score: 76/100

Risk area Weight Score (0–100) Why
Business & competition 25% 75 It is a middleman with no proven moat; its own software is $3.5M of revenue. It faces far larger, better-funded neoclouds and telcos, and 2026 margins are near zero.
Financial health & dilution 20% 80 It holds $232M of convertibles with an $8 reset floor, a B- rating, $228M of capex commitments and heavy stock compensation, and it raises equity roughly every year.
Execution (delivering the plan) 20% 80 It must grow revenue 4–5x while tripling headcount and running builds in three countries. Two timelines have already slipped.
Dependency & concentration 15% 85 Egypt was 76% of 2025 revenue; Yotta (levered, pre-IPO) is two of three GPU programmes; the Batam customer is unnamed. It also depends on Supermicro, Nvidia and landlords.
Regulation, policy & geopolitics 10% 60 US AI-chip export rules are in flux, it has Taiwan–China exposure, Egyptian pound risk, and Indonesian and Thai permits and power.
Sentiment & expectations 10% 65 The market is sceptical (about 0.7x 2027 guided revenue), but the broker targets of $24–44 and management's guidance embed full delivery.

Overall = the weighted average: 76/100 (very high). The score is driven by the stacking of risks: margin, financing, counterparty and execution all have to go right at the same time. Each is plausible, but none is proven yet. This is the first deep dive, so there is no prior score to compare.

Conclusion

  • What kind of company this is: a micro-cap, capital-markets-driven GPU leasing and installation venture, run by a financier CEO, attached to a shrinking one-customer integration business. Owning it is a bet that management can (a) finance, (b) install and (c) collect on about $3bn of GPU rentals through Yotta and an unnamed Batam customer before the chips age or the equity is diluted.
  • The 2–3 things that will decide the outcome:
    1. Whether Q4 2026 and Q1 2027 show real rental revenue at high gross margin.
    2. Whether project debt closes without new equity or a convertible reset.
    3. Whether Yotta stays solvent and lists.
  • Signposts:
    • Bull case confirmed by:
      • Yotta 1 accepted and billing by year-end 2026.
      • Q4 2026 or Q1 2027 gross margin above 25% with service revenue rising.
      • A closed Batam project-debt facility with no new equity.
      • Yotta's IPO completed with the Gorilla contracts disclosed.
    • Bull case broken by:
      • Another equity raise or a conversion-price reset toward $8.
      • Yotta 1 or Batam revenue slipping into mid-2027.
      • Egypt receivables over 180 days growing, or a new auditor issue.
      • The 2027 guide cut, or the Freyr deal formally dropped without replacement.
  • Balanced verdict:
    • The strongest bull version is that Gorilla got to the GPU scarcity window early. It has signed long fixed-fee contracts in two fast-growing sovereign-AI markets, ordered the hardware, and raised the capital at a premium. If the rentals switch on, 2027 revenue would be several times the current market value, at high margins.
    • The strongest bear version is that Gorilla is a thin-margin hardware pass-through whose headline contracts dwarf both its balance sheet and its audited backlog. It depends on a levered Indian partner, an anonymous Indonesian customer and debt it has not closed, with a convertible structure that punishes shareholders if the stock stays down.
    • Today the evidence leans to the bear side on risk, because no high-margin rental revenue, closed project debt or named Batam customer exists yet. It leans to neutral on direction, because installations are physically under way and the timeline slips are about a quarter, not years.

Open questions & data gaps

  • Earnings call transcripts: verbatim transcripts were not obtained, only summaries (Motley Fool, Webull/Benzinga). Quotes attributed to calls are summary-level. Medium importance.
  • Batam customer and terms: the customer's identity, and whether take-or-pay terms and minimums are in the contract text, were not found. High importance.
  • Hardware buyer and GPU ownership: who bought the $52.5M of H1 hardware, and whether Gorilla owns the Yotta GPUs or sells them to Yotta. Not disclosed. High importance, because it decides whether this is a rental or a resale business.
  • Convertible reset terms: reset dates and formula for both series, and the June notes' floor. The indenture was not read. High importance.
  • Freyr status: the $1.4bn deal's status in 2026. Not found. Medium importance.
  • Benchmark upgrade: a reported "upgrade to Strong-Buy" (2026-09-29) appears on one aggregator only and is unverified. Benchmark was the placement agent on both converts. Low importance.
  • Earlier auditor history: the reason for the April 2026 auditor change, and whether a prior auditor change preceded Marcum's January 2024 appointment. Medium importance.
  • Egypt contract status: completion percentage and the mid-2027 end date. Inferred, not confirmed. Medium importance.
  • Q3 2026 results date: about 2026-11-16 to 18, from data providers only. Low importance.
Sources
  1. Gorilla Technology Group Form 20-F FY2025 — SEC EDGAR — 2026-04-15 — https://www.sec.gov/Archives/edgar/data/1903145/000143774926012385/grrr20241231_20f.htm
  2. Gorilla Technology & Yotta Sign Landmark AI Infrastructure Deal — Gorilla IR / Newsfile — 2026-03-16 — https://www.newsfilecorp.com/release/288663
  3. H1 2026 results press release (Ex 99.1) and condensed interim financial statements (Ex 99.2), Form 6-K — SEC EDGAR — 2026-08-24 — https://www.sec.gov/Archives/edgar/data/1903145/000143774926028827/ex_974012.htm ; https://www.sec.gov/Archives/edgar/data/1903145/000143774926028827/ex_974013.htm
  4. Gorilla (GRRR) Q2 2026 Earnings Call Transcript (summaries only) — Motley Fool / Webull (Benzinga) — 2026-08-24 — https://www.fool.com/earnings/call-transcripts/2026/08/31/gorilla-grrr-q2-2026-earnings-call-transcript/ ; https://www.webull.com/news/15499649423107072 ; Q1 2026 call (summary) — Webull — 2026-05-27 — https://www.webull.hk/en/news/14957892135470080
  5. Gorilla Technology announces $2 billion AI infrastructure deal in India with Supermicro — Gorilla IR — 2026-06-02 — https://investors.gorilla-technology.com/gorilla-technology-announces-2-billion-ai-infrastructure-deal-in-india-with-supermicro-expanding-strategic-collaboration-across-asia-pacific/
  6. Gorilla and Yotta expand India AI infrastructure collaboration (~US$2.8B) — Newsfile — 2026-04-29 — https://www.newsfilecorp.com/release/294876
  7. Gorilla secures US$2.5B five-year AI GPUaaS compute contract (NeutraDC Batam) — Newsfile — 2026-06-23 — https://www.newsfilecorp.com/release/302522/
  8. Form 6-K: $107M convertible notes and lock-ups — SEC EDGAR — 2026-06-05 — https://www.sec.gov/Archives/edgar/data/1903145/000121390026065898/ea0293736-6k_gorilla.htm
  9. Fitch assigns Gorilla Technology Group first-time IDR of 'B-'; Outlook Positive — MarketScreener (Fitch text) / Newsfile — 2026-08-27 — https://www.marketscreener.com/news/fitch-assigns-gorilla-technology-group-inc-first-time-idr-of-b-outlook-positive-ce7858d8db8fff26 ; https://www.newsfilecorp.com/release/311788/
  10. Gorilla Technology secures 200MW Thailand AI campus — Newsfile — 2026-05-05 — https://www.newsfilecorp.com/release/295957/
  11. Korat AI Campus Strategic Deck v0.1 — Gorilla Technology — 2026-08 — https://www.gorilla-technology.com/wp-content/uploads/2026/08/Gorilla_Korat_AI_Campus_Strategic_Deck_v0.1-1.pdf
  12. Gorilla Technology wins major national AI intelligence infrastructure expansion in Taiwan — Media OutReach / Finviz — 2026-09-22 — https://finviz.com/news/394360/gorilla-technology-secures-multi-year-taiwan-ai-contract-worth-tens-of-millions-of-dollars
  13. Indian data centre operator Yotta targets 2027 IPO, seeks up to $1.5b (Reuters interview) — DealStreetAsia — 2026-09-02 — https://www.dealstreetasia.com/stories/data-centre-operator-yotta-india-ipo-493935
  14. Who's actually paying for Yotta rival's Nvidia orders (partly paywalled) — The Ken — 2026 — https://the-ken.com/story/whos-actually-paying-for-yotta-rivals-multi-billion-dollar-nvidia-orders/
  15. Q1 2026 results press release (Ex 99.1) — SEC EDGAR — 2026-05-27 — https://www.sec.gov/Archives/edgar/data/1903145/000143774926018444/ex_944447.htm
  16. Yotta Advances $1.5 Billion IPO Plan (GPU SPV financing structures) — CXO Digital Pulse — 2026 — https://www.cxodigitalpulse.com/?p=61508
  17. Gorilla Technology Group Inc (equity research note, opinion) — EAA Partners (Substack) — 2026-07 — https://eaapartners.substack.com/p/gorilla-technology-group-inc
  18. GRRR: The Good, The Bad and The Ugly (issuer-sponsored research) — Zacks SCR — 2026-08-25 — https://s27.q4cdn.com/906368049/files/News/2026/Zacks_SCR_Research_08252026_GRRR_Lantier.pdf
  19. Firmus to deploy 170,000 GPU cluster in Batam — DatacenterDynamics — 2026 — https://www.datacenterdynamics.com/en/news/firmus-to-deploy-170000-gpu-cluster-in-batam-indonesia/
  20. FY2024 results and 2025 guidance press release — SEC EDGAR — 2025-03-31 — https://www.sec.gov/Archives/edgar/data/1903145/000121390025026420/ea023624201ex99-1_gorilla.htm
  21. Gorilla Technology Accelerates Execution Across its Asian AI Infrastructure Platform — Newsfile — 2026-07-21 — https://www.newsfilecorp.com/release/305933/
  22. GRRR news feed and analyst table (price ~$11.94 on 2026-10-08; daily moves) — Stock Titan / MarketBeat — viewed 2026-10-09 — https://www.stocktitan.net/news/GRRR/ ; https://www.marketbeat.com/stocks/NASDAQ/GRRR/forecast/
  23. GPU price index — AIMultiple — viewed 2026-10 — https://aimultiple.com/gpu-index
  24. Telkom subsidiary NeutraDC launches AI data center in Batam (18MW) — Indonesia Business Post — 2025-10 — https://indonesiabusinesspost.com/5559/corporate-affairs/telkom-subsidiary-neutradc-launches-ai-data-center-in-batam-with-18mw-capacity ; Telkomsel seeks US$1–1.5bn to sell major stake in NeutraDC — w.media — 2026 — https://w.media/telkomsel-seeks-us1-us1-5-billion-to-sell-major-stake-in-neutradc/
  25. Series B convertible bonds pricing release (Ex 99.7, Form 6-K) — SEC EDGAR — 2026-07-15/17 — https://www.sec.gov/Archives/edgar/data/0001903145/000121390026079193/ea029837101ex99-7.htm
  26. Gorilla Technology Reports Full Year 2025 (Ex 99.1) — SEC EDGAR — 2026-03-02 — https://www.sec.gov/Archives/edgar/data/1903145/000143774926006382/ex_926172.htm
  27. 20-F FY2025, home-country practice and share authorisation — see [1]
  28. GRRR RedChip webinar: $10B pipeline — Fintool — 2026 — https://fintool.com/news/grrr-redchip-webinar-10b-pipeline
  29. Form 6-K: Change in registrant's certifying accountant — SEC EDGAR — 2026-05-06 — https://www.sec.gov/Archives/edgar/data/1903145/000143774926015208/grrr20260506_6k.htm
  30. Gorilla Technology signs MOU to acquire Shackleton Finance — Newsfile — 2026-02-18 — https://www.newsfilecorp.com/release/284311
  31. Problems at Gorilla Technology Group (free portion) — The Bear Cave (Substack) — 2025-03-06 — https://thebearcave.substack.com/p/problems-at-gorilla-technology-group
  32. Gorilla Technologies stock tumbles on Citron short-sell claims — Investing.com — 2025-03-17 — https://www.investing.com/news/stock-market-news/gorilla-technologies-stock-tumbles-on-citron-shortsell-claims-93CH-3932752
  33. BIS clarifies licence requirements for advanced computing items to D:5/Macau-headquartered entities — Baker McKenzie Sanctions Blog — 2026-06-08 — https://sanctionsnews.bakermckenzie.com/bis-clarifies-that-license-requirements-for-advanced-computing-items-to-country-group-d5-and-macau-headquartered-entities-remain-in-force-despite-ai-diffusion-rule-non-enforcement/
  34. Chip Security Act 2026 — Carra Globe — 2026-04-17 — https://carraglobe.com/chip-security-act-ai-hardware-importers/
  35. GRRR's $5.5B commercial book: Compass Point initiation — Stocktwits — 2026-08-18 — https://stocktwits.com/news-articles/markets/equity/grrr-5-5b-commercial-book-wall-street-attention-compass-point-193-upside/cZYcWj8RJj0
All GRRR Bulls vs Bears reports
Disclaimer. AI-generated research and podcast (Claude by Anthropic, voices by Google Gemini) for educational purposes — not financial advice and not a recommendation to buy or sell. Willem (MR.INVEST) owns this stock. Check the company’s own filings before making any decision.
← All Bulls vs Bears reports