Stock Analysis · Bulls vs Bears

Duos Technologies Group DUOT

🐂 Bulls vs Bears 🐻Report #311 Oct 2026
★ Willem owns this stock · 4.3% of the portfolio
66
high risk
🐂 9vs10 🐻
Bears lead
30%
▲ Chance of the biggest tailwind the forecasts miss
50%
▼ Chance of the biggest headwind the forecasts miss
The verdict

The evidence still leans bearish on risk: Axe Compute is the tenant on ~$611M of contracts and owes Duos $42.9M. Axe's own 10-Q warns its liquidity "may be insufficient to fund our capital requirements". Axe also says the 2026-09-30 GPU sale "significantly" improves its margin because Duos now pays the cluster's power and facility costs for five years. The bull case rests on real assets: a live Columbus campus, net…

🎙️ Bulls vs Bears podcast — Duos Technologies Group
Alex and Sam debate every bull and bear argument · 12:45
0:0012:45
⤓ Download

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

🐂 Bull arguments 9

  1. 1Contracted rents match big-campus leases on cheaper, faster buildsMedium
  2. 2Columbus proved a working AI site in under 60 daysMedium
  3. 3Net cash and no debt after shedding the GPU financingMedium
  4. 4Market value is ~0.37x contracted rent; targets average $20Medium
  5. 5The end user is a creditworthy Tier 1 hyperscaler on take-or-payWeak
  6. 6Small-site AI capacity is scarce and rents are still risingMedium
  7. 7Insiders bought in September; anchor holders backed every raiseWeak
  8. 8Landlord model sheds most GPU obsolescence riskWeak
  9. 9Technology Solutions backlog nearly doubled to $25MWeak

🐻 Bear arguments 10

  1. 1Nearly all contracted rent depends on Axe, a $22M-cash neocloudHigh
  2. 2The GPU sale moved value from Duos to AxeMedium
  3. 3Duos and Axe both claim the same Columbus customer contractHigh
  4. 4FY26 "$50M" guide reaffirmed after selling its biggest pieceHigh
  5. 5Shares up 54% this year, and short sellers keep addingMedium
  6. 6Financial reporting is loose: late 10-Q, internal inconsistencies, four CFOsMedium
  7. 7Rail unit sold to the interim CFO's company on seller financingMedium
  8. 8No moat, and five-year terms against 10–15 years at peersMedium
  9. 9The 55 MW sites are not bought, powered or financed yetMedium
  10. 10Air-side cooling may not suit Rubin-era liquid-cooled racksLow

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

🐂 Bull #1Contracted rents match big-campus leases on cheaper, faster builds

  • The point:
    • Duos earns rent per MW at or above what large AI-campus landlords sign for.
    • It builds for "under $6M a megawatt".
    • A site could therefore earn back its build cost within the five-year term.
  • Evidence:

    • The 10 MW Columbus order is worth >$111M over five years. The 55 MW Axe orders are worth >$500M of base payments, "excluding electricity" [1,10].

    • Interpretation (my arithmetic): that is ~$2.2M and ~$1.8M per MW-year, or ~$185 and ~$152 per kW-month.

    • Peers, per MW-year (my arithmetic on stated totals):

      Deal Per MW-year
      WhiteFiber–Nscale, 40 MW, 10 years [34] ~$2.16M
      Digi Power X–Cerebras, 40 MW, 10 years [35] ~$2.75M
      Applied Digital, 300 MW, 15 years [36] ~$1.67M
      TeraWulf–Fluidstack, 10 years [37] ~$1.85M

    • Build cost "under 6 million a megawatt" (CEO, Q2 call) [1]. Duos must "contribute basically 30 million per site" to the Axe JVs [1].

    • Cantor's Overweight thesis (as publicly reported) rests on the same point: Duos turns power into contracts quickly [28].

  • Why it matters: if 65 MW bills as contracted, Duos earns ~$120M a year of rent at claimed ~80% colocation EBITDA margins (unaudited) [1]. That is the core of the ≥$160M 2027 framework.
  • 😈 Devil's advocate:
    • The payer behind ~$611M of that rent is Axe (Bear #1).
    • The 55 MW sits in JVs where Duos owns 51%, so only about half of that profit belongs to Duos shareholders [1].
    • Five-year terms must recover ~$6M/MW plus land and power upgrades before 2031–32.
    • The <$6M/MW figure is an unaudited CEO remark. On the Q1 call it was "6 million to 6.5 million" [3].
  • Verdict: Medium The rent levels are in line with the market, but the cost figures are unaudited and the rent depends on one payer.

🐂 Bull #2Columbus proved a working AI site in under 60 days

  • The point: Duos turned an empty building into a live 2,304-GPU B300 site in weeks. Speed is scarce in AI infrastructure, so this is a skill Duos can sell.
  • Evidence:
    • ">7 MW of new infrastructure in under 60 days" [1].
    • The cluster is live and serving its end customer. The customer contract has been extended to 2031 [7].
    • Axe's CEO calls Duos "a world class data center builder" [7].
    • DDC Solutions (owned by Daikin) supplies the rack containment [41].
  • Why it matters: speed and a working reference site are what win the next tenant. Recker says more than 100 MW of interest is "in the funnel" [1].
  • 😈 Devil's advocate:
    • Columbus was an existing building with a grid feed.
    • Power above 5 MW is tied to a seller earn-out ($5M per extra 5 MW). No Georgia Power schedule has been published [12].
    • Not one of the 55 MW sites had been bought by the Q2 call [1].
    • One site is one data point.
  • Verdict: Medium The skill is real but has been shown only once.

🐂 Bull #3Net cash and no debt after shedding the GPU financing

  • The point: Duos has no bank debt, and it got the GPU loan off its books.
  • Evidence:
    • Cash was $112.3M at 30 Jun. Liabilities were mostly customer deposits; notes payable were only ~$0.33M [2].
    • The CFO, on the Q2 call: "effectively debt-free" [1].
    • At closing on 09-30, Axe repaid the SPV's ~$87.8M asset-backed loan [5,7].
    • Axe also owes Duos $42.9M. Any unpaid balance is due in full by the earlier of 12 months or an Axe financing secured by the cluster [5,7].
    • Axe's own release lists a "security interest held by Duos in the Georgia Cluster pending payment" [7].
    • About $10M of APR sale proceeds sit in escrow until ~May 2027 [2].
  • Why it matters: Duos can fund Columbus and its first JV equity without immediate borrowing. The $42.9M receivable is secured, short-dated cash.
  • 😈 Devil's advocate:
    • $68.8M of the 30 Jun cash had already gone out as GPU deposits [2]. What came back is the $42.9M receivable, not cash.
    • The Columbus purchase took $15M in cash in July [12].
    • Cash at 30 Sep is not reported. Estimate: ~$55–85M before receivables, depending on how the $18.8M customer deposit moved with the SPV.
    • Interpretation: the Columbus 10→20 MW expansion (~$60M at $6M/MW, estimate) plus ~$30M per Axe site exceeds that cash.
    • No debt facility has been signed. The GPU senior debt promised for H2 became the sale instead [1].
  • Verdict: Medium The balance sheet is clean today, but the cash on hand is smaller than the headline figure and the build plan needs more.

🐂 Bull #4Market value is ~0.37x contracted rent; targets average $20

  • The point: the market prices Duos as if most of its contracts will never bill.
  • Evidence:
    • Market cap ≈ $227M (31.47M shares × $7.20 on 10-09) [2,24]. On stale 30 Jun cash, EV is ≈ $115M [31].
    • Contracted value is ~$611M ($111M + $500M) [1,10]. Interpretation: market cap is ~0.37x contracted value and EV ~0.19x.
    • Interpretation: EV is ~$1.5M per contracted MW. Windsor Drake puts Applied Digital at ~$8.2M and TeraWulf at ~$13.2M per contracted MW [46].
    • Price targets are $9.50–27, average $20.17 [25]. Chardan's $9.50 is the only target near the price [27].
    • Seeking Alpha bull (Veurink, 2026-08-19, long) puts 2027 EV/EBITDA at ~2x [33].
  • Why it matters: if Axe pays and the sites get built, the gap to peers is very large.
  • 😈 Devil's advocate:
    • Peers trade on energised MW with investment-grade tenants. Duos has ~10 MW energised and a micro-cap tenant.
    • The 55 MW contract value is gross, and Duos owns 51%.
    • The EV uses 30 Jun cash. The true figure after the GPU deposits and capex is higher (estimate: EV ~$140–170M).
    • The discount mostly prices counterparty risk, not neglect.
  • Verdict: Medium It is cheap if the contracts hold, but the discount is explained by Bears #1–#4.

🐂 Bull #5The end user is a creditworthy Tier 1 hyperscaler on take-or-pay

  • The point: behind Axe sits a top-credit cloud company that prepays. That makes the payment chain safer than Axe's balance sheet suggests.
  • Evidence:
    • Recker, Q2 call: "they are a Tier 1 hyperscaler. They did visit the site in person last week… extremely credible and creditworthy… we funded our GPUs based on that customer" [1].
    • Axe's 10-Q says its largest contract "is supported by a customer deposit, prepayment, and monthly in-advance payments on a take-or-pay basis" [8].
    • Axe says the Columbus customer is "committed for five years", through 2031 [7].
  • Why it matters: if the end user prepays Axe monthly in advance, Axe has cash to pay Duos' rent before Duos delivers.
  • 😈 Devil's advocate:
    • No guarantee from the hyperscaler to Duos has been disclosed in any filing.
    • The same Axe risk factor goes on: these amounts "may be insufficient to fund our capital requirements" [8].
    • The hyperscaler is unnamed. No public source identifies it.
    • On the April release, the Columbus end-customer deal was a $260M, three-year "enterprise contract", not a hyperscaler [49].
  • Verdict: Weak to-Medium. It is a plausible cushion but cannot be verified.

🐂 Bull #6Small-site AI capacity is scarce and rents are still rising

  • The point: demand for powered AI capacity still exceeds supply, which supports Duos' pricing and its chance of finding new tenants.
  • Evidence:
    • CBRE: primary-market vacancy ~1.4%; asking rent for 250–500 kW requirements was a record $196.25/kW-month at YE25 (+6.6%) [38].
    • Hyperscaler 2026 capex guidance is ~$735–750B after Q2 [43].
    • No neocloud default was found in 2026. Nscale filed for a US IPO on 2026-09-18 with >$103B of contracted value [44].
  • Why it matters: scarce power raises the odds of re-letting, and of a second tenant that is not Axe.
  • 😈 Devil's advocate:
    • The mid-September AI selloff (chip index −5–6%) showed how fast sentiment can turn [29].
    • The Fed hiked on 2026-09-16, which raises funding costs for small builders [45].
    • CBRE warns landlords to be selective with young AI-cloud tenants.
  • Verdict: Medium

🐂 Bull #7Insiders bought in September; anchor holders backed every raise

  • The point: insiders bought on the open market, and large holders have supported the raises.
  • Evidence:
    • Open-market purchases (code P):
      • Recker: 9,250 shares at $9.14 (~$84.5K) on 09-22, in an IRA [21].
      • Director Nixon: 5,764 shares at $8.50–8.60 ($49K) on 09-18 [22].
    • No open-market insider sales in 2026 across 33 Form 4s [21,22].
    • Bleichroeder holds 6.75M shares (18.48%). Alyeska holds 3.52M (9.99%, up from 1.91M at 31 Mar) [23].
  • Why it matters: these holders have backed management's pivot with real money.
  • 😈 Devil's advocate:
    • The buys total ~$134K against a ~$227M market cap.
    • Recker holds 400K restricted shares, with up to 880K more tied to targets [15].
    • Large holders concentrated in a small float can also exit together.
  • Verdict: Weak

🐂 Bull #8Landlord model sheds most GPU obsolescence risk

  • The point: Duos no longer owns chips that lose value as Rubin arrives.
  • Evidence: "Compute equipment ownership, GPU utilization, and technology refresh sit with the counterparty" [6].
  • Why it matters: buildings outlive several chip generations.
  • 😈 Devil's advocate:
    • Duos now "will lease the underlying GPU compute capacity from Axe Compute under a capacity supply arrangement" to serve the Columbus customer [6].
    • Duos also fixes power and facility costs for five years [7].
    • Interpretation: Duos still has a GPU-shaped contract, now as a reseller rather than an owner, on terms nobody has seen.
  • Verdict: Weak to-Medium. Down from Medium in report #2, because the lease-back keeps Duos inside the GPU economics.

🐂 Bull #9Technology Solutions backlog nearly doubled to $25M

  • The point: equipment resale brings near-term revenue and better buying power.
  • Evidence: backlog was ~$14M at Q1 and ~$25M at Q2. Q2 revenue was $3.23M [1,3].
  • Why it matters: it covers about half of the FY26 bridge.
  • 😈 Devil's advocate: it is low-margin, lumpy resale with no segment margin disclosed.
  • Verdict: Weak

Bear case

🐻 Bear #1Nearly all contracted rent depends on Axe, a $22M-cash neocloud

  • The point: one thinly capitalised counterparty is Duos' tenant, its debtor and its proposed funding partner, all at once.
  • Evidence:

    • Prepared remarks, Q2 call: "a five-year colocation agreement with Axe Compute… 10 megawatts… at our Columbus, Georgia campus… more than $111 million" [1]. The 55 MW is also with Axe [10,11].

    • Axe at 30 Jun [8]:

      Item Figure
      Cash $21.9M
      Equity $34.0M
      Accumulated deficit $438.4M
      Q2 revenue $3.2M
      Q2 net loss $17.2M
      Take-or-pay GPU commitments of its own $262.6M ($107.7M due in 2027)

    • Its ATH token treasury cost $94.7M and was worth $11.1M [8].

    • Axe's own risk factor says its liquidity "may be insufficient to fund our capital requirements", citing reliance on "volatile sources of liquidity, including the price of ATH" [8].

    • Axe claims >$3B of 2026 contracts. Its expected $534M of prepayments "over the next 30 days" (from 07-27) is unconfirmed [9].

    • The source of the $87.8M Axe paid at closing: not found.

  • Why it matters:
    • If Axe stumbles, Duos loses most of its 2027 revenue, the $140M JV equity, and possibly the $42.9M receivable.
    • Duos would then hold half-built sites.
  • 🛡️ Counter-argument:
    • AGPU did not fall with DUOT in October: it closed $12.29 on 10-06 and $11.68 on 10-09 [24]. The market is not pricing Axe distress.
    • Axe did find $87.8M at closing without issuing equity [7].
    • Duos holds a security interest in the cluster until it is paid [7].
    • Rent is billed only after acceptance, so Duos also avoids building for nothing if Axe signals trouble early.
  • Severity: High This remains the single largest determinant of the outcome.

🐻 Bear #2The GPU sale moved value from Duos to Axe

  • The point: the sale reads as Duos giving Axe a better deal than Duos itself had.
  • Evidence:
    • Axe's headline: the deal is "Significantly Increasing the Contract's Expected Free Cash Flow and Gross Margin" [7].
    • The reason Axe gives: "Under a separate agreement, Duos covers the facility's colocation and energy costs for the full five-year term" [7].
    • The end-customer contract was extended from 3 to 5 years, with ~$364.6M of revenue going to Axe [7].
    • Duos had paid $68.8M of GPU deposits by 30 Jun and held an $18.8M customer deposit [2]. It received a $42.9M deferred price [5].
    • Interpretation: if the customer deposit moved with the SPV, Duos' net equity was ~$50M against $42.9M received, a loss of ~$7M. If it did not, the gap is ~$26M. Actual gain or loss: not found until the Q3 10-Q.
    • March 2026 thesis: the GPU contract would give ~$40M/yr of EBITDA at >80% gross margin [19]. That economics now sits with Axe.
  • Why it matters:
    • Duos swapped a high-margin GPU contract for a fixed-cost obligation (power and facility for five years) and a receivable.
    • Duos says the revised agreement is "expected to increase the revenue Duos recognizes" [6]. Revenue up while margin falls is a bad trade for shareholders, and Recker's bonus is 60% revenue-based [15].
  • 🛡️ Counter-argument:
    • Duos shed ~$87.8M of debt and the residual-value risk on 2,304 GPUs that will be a generation old in 2027.
    • It recycled cash into sites, which Duos says is "where it is deployed once into a long-lived asset" [6].
    • Power at Columbus is cheap (~$0.04–0.07/kWh per management [1]). Interpretation: on ~6 MW of total load, a 25% power price rise costs only ~$0.7M a year (estimate).
  • Severity: Medium High. The direction is clear, but the size is unknown until the capacity-supply and power agreements are disclosed.

🐻 Bear #3Duos and Axe both claim the same Columbus customer contract

  • The point: the two parties describe the same deal in incompatible ways.
  • Evidence:
    • Duos [6]: "Duos and the customer have also entered into a revised five-year agreement… Duos will lease the underlying GPU compute capacity from Axe." CFO: "We keep the customer and the economics."
    • Axe [7]: "Axe Compute's AI Cluster Customer Agreement Extended to Five Years… Axe Compute expects to generate approximately $364.6 million in revenue."
    • The 10 MW deal: the 10-05 Duos release still calls it "a 10 MW hyperscaler colocation agreement" [6], while Duos' own Q2 script says it is with Axe [1].
    • Debt retired: Duos says ~$98.1M [6]; both 8-Ks say $87.8M [5,7].
    • Payment timing: Duos says "over the next 60 months" [6]; the 8-K says due within 12 months at the latest [5].
    • Exhibits: none of the agreements has been filed as an exhibit. Axe says the equity purchase agreement comes with its Q3 10-Q [5].
  • Why it matters: investors cannot tell who books the end-customer revenue, who carries the credit risk, or what Duos pays Axe.
  • 🛡️ Counter-argument:
    • A three-way structure is possible: the customer pays Duos, and Duos pays Axe for capacity. In that case both companies "recognise" revenue, one gross and one as a supplier.
    • The 8-Ks themselves are consistent with each other; the conflicts are in the press releases.
  • Severity: High until the Q3 10-Q. It feeds Bears #2 and #4.

🐻 Bear #4FY26 ">$50M" guide reaffirmed after selling its biggest piece

  • The point: the FY26 guidance bridge depends on revenue from equipment Duos no longer owns.
  • Evidence:
    • Q2 bridge: ~$26M GPUaaS + ~$25M Technology Solutions [1].
    • H1 revenue was $8.32M, including a one-time $2.71M catch-up [2]. So H2 needs ~$42M.
    • GPU revenue of "$4.4 million… starting July 1" (Q1 call) [3] had not been recognised by the Q2 filing [2].
    • On 10-05 the CFO said: "Our outlook for 2026 is unchanged" [6]. No new bridge was given.
    • Q3 revenue consensus is $14.1M (range $12.3–18.8M) [26]. FY26 consensus is $51.5M [25].
    • The Q1 call's ~$9.5M revenue consensus could not be verified this run. Report #2's "Q1 miss" rests on a secondary source.
  • Why it matters: another miss would hurt management credibility a second time. Gross versus net presentation of the Axe capacity lease decides whether the number is met.
  • 🛡️ Counter-argument: if Duos books the end-customer bill gross, the old $4.4M/month line can survive. Duos' "increase the revenue Duos recognizes" wording points that way [6].
  • Severity: High for credibility, Medium for long-run value.

🐻 Bear #5Shares up 54% this year, and short sellers keep adding

  • The point: equity is the funding tool, and a growing short base is betting on more of it.
  • Evidence:
    • Shares: 20.45M (12/31/25) → 31.47M (08-18), plus 3.8M pre-funded warrants [2]. Raises were 8.67M shares at $7.50 in March [18] and $55M at $9.50 in June [17].
    • Overhang, on helper arithmetic from the 10-Q (≈44.5M fully diluted, +41%):
      • Series E convertible into ~4.79M shares at $2.61.
      • Series D ~0.33M.
      • 0.73M warrants.
      • 1.96M unvested restricted shares.
      • Columbus seller note convertible at $10.50 (~1.43M) [2,12].
    • Short interest (FINRA via Equibles [30]): 4.50M (07-15) → 6.10M (09-30). That is ~19.4% of shares, with days to cover rising from 5.7 to 11.5.
  • Why it matters: each raise below $9.50 cuts per-share value of the same megawatts.
  • 🛡️ Counter-argument:
    • Recker: "We're not in this game to dilute our investors… the debt financing will follow" [1].
    • The Axe JV equity of up to $140M is meant to fund most of the new-site capex [1].
    • At $7.20 the $10.50 seller-note conversion is out of the money.
  • Severity: Medium High.

🐻 Bear #6Financial reporting is loose: late 10-Q, internal inconsistencies, four CFOs

  • The point: the quality of the financial reporting is itself a risk.
  • Evidence:

    • Duos filed an NT 10-Q on 08-17: the auditor "requires additional time" [14].

    • The Q2 10-Q gives different figures for the same items in different sections [2]:

      Item Figures given
      Net income $44.35M and $45.35M
      Working capital $114.4M and $119.6M
      Columbus note a "two-year" payable note (10-Q) vs a "three-year" contingent earn-out (8-K [12])

    • CFOs: Brown → Goldfarb (interim) → DeAlmeida (08-24) [16].

    • Controls were concluded "effective". The auditor is Salberg & Company, auditor since 2013 [4,2].

  • Why it matters: investors are being asked to underwrite complex three-way contracts with a finance team that is new and stretched.
  • 🛡️ Counter-argument: there has been no restatement and no material weakness. A permanent CFO with public-company experience is now in place.
  • Severity: Medium

🐻 Bear #7Rail unit sold to the interim CFO's company on seller financing

  • The point: a related-party sale of the old core business.
  • Evidence:
    • The rail unit was sold on 08-05 to Sandbank Acosta LLC, 50% owned by interim CFO Goldfarb. Duos took a $5.4M, 5% note due 2031 [13,2].
    • Duos injected $3.5M of cash into the unit before the sale [13].
    • Board approval was stated; the Q1 call mentions a fairness-opinion process [3].
  • Why it matters: it signals how the board handles conflicts, just as the Axe agreements go undisclosed.
  • 🛡️ Counter-argument: the unit was loss-making and hard to sell, and Goldfarb is no longer CFO.
  • Severity: Medium Low.

🐻 Bear #8No moat, and five-year terms against 10–15 years at peers

  • The point: Duos' edge is easy to copy, and its contracts are short.
  • Evidence:
    • Recker: "we're not rocket science over here… I anticipate other people coming out" [1].
    • Peers sign 10–15-year leases [34,35,36,37]. Duos signs five-year terms [10].
  • Why it matters: re-letting risk falls in 2031–32, by when chips need 100 kW+ racks.
  • 🛡️ Counter-argument: a powered, live site is hard to replicate quickly, and the incumbent landlord usually re-signs.
  • Severity: Medium

🐻 Bear #9The 55 MW sites are not bought, powered or financed yet

  • The point: the bulk of the contracted revenue depends on physical work that has not started.
  • Evidence:
    • "Six to seven" candidate sites in TX, SC, IA and AL, "a few in LOI status" [1].
    • Billing starts only after ready-for-service testing and Axe's written acceptance [10].
    • Texas paused data-center grid approvals in the ERCOT queue "regardless of load size" on 2026-08-03 [39].
    • Transformer lead times are ~128 weeks [50].
  • Why it matters: every quarter of slip pushes ~$25M of gross rent out of 2027 (estimate: 55 MW × $1.82M ÷ 4).
  • 🛡️ Counter-argument:
    • Duos targets "stranded power" sites with existing grid lines [1].
    • Georgia's large-load rule bites only at 100 MW or more [40].
    • Panhandle sites served by SPS sit outside ERCOT (weak source).
  • Severity: Medium High.

🐻 Bear #10Air-side cooling may not suit Rubin-era liquid-cooled racks

  • The point: the cooling at Columbus may not fit the next chip generation.
  • Evidence:
    • DDC's Columbus units are airside containment on facility chilled water [41].
    • Nvidia's Vera Rubin racks are fully liquid-cooled [42].
  • Why it matters: a five-year tenant in 2031 may need liquid cooling, which would mean retrofit capex.
  • 🛡️ Counter-argument: chilled-water plants can feed liquid-cooling distribution units, and five years covers the B300 life.
  • Severity: Low Medium.

Tailwinds NOT in the forecasts

Consensus (S&P, 3 analysts) [25]:

  • FY26: revenue $51.5M, EPS $1.20. The EPS is inflated by the $53.2M APR gain.
  • FY27: revenue $162.1M, EPS $0.59.

The FY27 figures mirror the "≥$160M, only announced and contracted programs" framework [1]. So the contracted Axe megawatts are already "in". Sizing uses ~35.3M shares (31.47M plus 3.8M pre-funded warrants).

For lost or added colocation revenue I assume a 40% incremental after-tax margin (estimate): ~80% claimed EBITDA margin, less depreciation and tax, haircut for site opex. Duos' share of the 55 MW JVs is 51%.

  • Axe 49% equity becomes definitive and project debt follows — 30% chance within 12 months

    • Impact on revenue: none vs consensus FY27; it is a funding event.
    • Impact on EPS: avoids ~−$0.18 (−30%) of FY27 dilution vs a $100M equity-funded path.
    • How we got the number (estimate): $100M ÷ $6.50 ≈ 15.4M new shares; $0.59 × 35.3 ÷ 50.7 = $0.41 → +$0.18 saved.
    • Expected value: 0.30 × $0.18 ≈ +$0.05.
    • What it is: a signed equity purchase agreement for up to $140M of Axe cash, plus a construction or term loan against the contracted rent.
    • Why it's not priced in: the term sheet is non-binding [10]. The stock at $7.20 and the 19% short interest imply the market expects a raise.
    • Trigger to watch: an 8-K with the definitive agreement or loan; the Axe Q3 10-Q exhibits.
    • Why this probability: Axe has a $1B shelf and ~$83M of ATM room [8]. But nothing has been signed in eight weeks, and its own liquidity warning cuts the odds.

  • A non-Axe tenant of ≥10 MW billing from mid-2027 — 28% chance by 2027-06-30

    • Impact on revenue: +$10M (+6%) vs consensus FY27 $162.1M.
    • Impact on EPS: +$0.11 (+19%) vs FY27 $0.59.
    • How we got the number (estimate): 10 MW × $2.0M/MW-yr × 0.5 year = $10M × 40% ÷ 35.3M ≈ +$0.11.
    • Expected value: 0.28 × $0.11 ≈ +$0.03.
    • What it is: a direct lease at Columbus's second 10 MW or a new site, with a tenant other than Axe.
    • Why it's not priced in: guidance counts only contracted programs [1].
    • Trigger to watch: an 8-K naming a new tenant; the formal 2027 guide with Q3 results.
    • Why this probability: the market is tight [38], but in 15 months Duos has signed only Axe at scale.

  • Columbus GPU contract is booked gross at higher revenue — 40% chance (shown in Q3/Q4 2026 filings)

    • Impact on revenue: possibly +$20–50M a year. That is the old $4.4M/month line, if Duos is principal [3,6].
    • Impact on EPS: not quantifiable. The cost of the Axe capacity lease and of the power commitment is undisclosed, so the margin could be near zero.
    • How we got the number (estimate): $4.4M × 12 = $53M upper bound; no margin basis.
    • Expected value: not quantifiable.
    • What it is: Duos says the revised deal will "increase the revenue Duos recognizes" [6].
    • Why it's not priced in: consensus is unchanged since the sale, and it is unclear whether it includes the line.
    • Trigger to watch: the Q3 10-Q revenue note (principal versus agent).
    • Why this probability: the wording points to gross booking, but the same customer is claimed by Axe [7].

  • The hyperscaler signs or guarantees directly with Duos — 15% chance within 36 months

    • Impact on revenue/EPS: not quantifiable. It de-risks existing revenue rather than adding any.
    • What it is: the end user steps around Axe.
    • Why it's not priced in: nothing in the filings points to it.
    • Trigger to watch: a named-customer 8-K, or Axe exiting a site.
    • Why this probability: hyperscalers use intermediaries for small sites but go direct when an intermediary weakens.

  • Strategic sale of the company or its sites — 10% chance within 36 months

    • Impact on revenue/EPS: not quantifiable; no bid basis.
    • What it is / why not priced: Recker has sold two data-center firms before [15]. Change of control fully vests his equity [15].
    • Trigger to watch: a strategic review, or a 13D from Bleichroeder (18.48%).
    • Why this probability: a buyer wants energised MW with sound tenants, and Duos has ~10 MW.

  • 0Lat 10-year lease (15 sites, 225 cabinets) — 40% chance by year-end 2026

    • Impact: not quantifiable; the term-sheet value is undisclosed. Likely <1% of FY27 revenue.
    • Trigger to watch: an announcement before exclusivity ends (~mid-Nov).
    • Why this probability: small exclusive term sheets often close, but many don't.

Headwinds NOT in the forecasts

  • 55 MW readiness slips by six months or more — 50% chance (first billing after mid-2027)

    • Impact on revenue: −$50M (−31%) vs consensus FY27 $162.1M.
    • Impact on EPS: −$0.29 (−49%) vs FY27 $0.59.
    • How we got the number (estimate): 55 MW × $1.82M × 0.5 year ≈ $50M × 40% × 51% ÷ 35.3M ≈ −$0.29.
    • Expected value: 0.50 × −$0.29 ≈ −$0.14.
    • What it is: site purchases, power, equipment or financing delay acceptance.
    • Why it's not priced in: consensus follows the "late 2026 into early 2027" target [10].
    • Trigger to watch: no site purchase 8-K by year-end 2026; a revised timeline at Q3.
    • Why this probability: the sites were at LOI stage in August with no financing. Lead times and the Texas pause make the first-guided dates hard to hit.

  • Equity raise below $7.50 within 12 months — 60% chance

    • Impact on revenue: none.
    • Impact on EPS: −$0.18 (−30%) vs FY27 $0.59.
    • How we got the number (estimate): $100M ÷ $6.50 = 15.4M shares; $0.59 × 35.3 ÷ 50.7 = $0.41.
    • Expected value: 0.60 × −$0.18 ≈ −$0.11.
    • What it is: an offering to fund Columbus 10→20 MW and JV equity.
    • Why it's not priced in: the average target of $20 implies little extra dilution [25].
    • Trigger to watch: an S-3 or ATM filing, or Q3 cash below ~$60M.
    • Why this probability: capex needs exceed estimated cash. Duos raised equity twice in 2026, and the debt promised for H2 never came. Raised from 55% because the stock is lower and the short base larger.

  • Axe funding or payment failure, or renegotiation — 38% chance within 36 months

    • Impact on revenue: −$36M (−22%) vs FY27 consensus, assuming half of Axe's FY27 rent is lost or deferred.
    • Impact on EPS: −$0.27 (−46%) vs FY27 $0.59. Plus a possible one-off write-down of up to $42.9M (≈ −$1.22/share of book value).
    • How we got the number (estimate): Columbus $22M × 50% × 40% = $4.4M; 55 MW $50M × 50% × 40% × 51% = $5.1M; total $9.5M ÷ 35.3M ≈ −$0.27.
    • Expected value: 0.38 × −$0.27 ≈ −$0.10.
    • What it is: Axe can't fund its commitments, delays acceptance, misses deferred payments or renegotiates.
    • Why it's not priced in: consensus models the contracts at face value.
    • Trigger to watch: Axe Q3 10-Q liquidity and going-concern language; AGPU raising equity at falling prices; the deferred price not refinanced by ~Sep 2027.
    • Why this probability: Axe's commitments are more than 10x its cash, and it warns itself. Slightly below 40% because Axe found $87.8M at closing and its stock has held up.

  • FY26 revenue below $50M — 50% chance (reported ~Mar 2027)

    • Impact on revenue: −$15M (−29%) vs consensus FY26 $51.5M.
    • Impact on EPS: −$0.14 (−12%) vs FY26 $1.20, before any loss on the SPV sale (a possible extra one-off of $0 to −$0.83/share).
    • How we got the number (estimate): $15M × 30% blended resale/colocation margin ÷ 31.5M basic ≈ −$0.14. SPV loss range $0–26M ÷ 31.5M.
    • Expected value: 0.50 × −$0.14 ≈ −$0.07.
    • What it is: the GPU line is booked net, or not at all, and resale slips.
    • Why it's not priced in: consensus sits at guidance [25].
    • Trigger to watch: Q3 revenue below ~$12M, and the gain/loss line on the sale.
    • Why this probability: the GPU line slipped twice and is now sold, but the CFO reaffirmed the guide days after the sale.

  • Texas grid pause delays a Texas Axe site — 20% chance

    • Impact on revenue: −$18M (−11%) vs FY27.
    • Impact on EPS: −$0.11 (−18%).
    • How we got the number (estimate): 10 MW × $1.82M × 1 year × 40% × 51% ÷ 35.3M ≈ −$0.11.
    • Expected value: 0.20 × −$0.11 ≈ −$0.02.
    • What it is / why not priced: the ERCOT audit has no end date [39], and Duos has not disclosed which grid its sites use.
    • Trigger to watch: Duos choosing SC, IA or AL sites first.
    • Why this probability: Panhandle sites may be outside ERCOT, and Duos has non-Texas options.

  • Fixed power cost on the Axe cluster rises — 35% chance of a ≥25% power price rise over five years

    • Impact on revenue: none.
    • Impact on EPS: −$0.02 (−3%) a year.
    • How we got the number (estimate): ~6 MW total load × 8,760 h = 52.6 GWh × $0.055/kWh × 25% ≈ $0.72M × 79% ÷ 35.3M ≈ −$0.016.
    • Expected value: 0.35 × −$0.02 ≈ −$0.006.
    • What it is: Duos pays the cluster's energy for five years under a fixed arrangement [7].
    • Why it's not priced in: the agreement is undisclosed.
    • Trigger to watch: the terms in the Q3 10-Q.
    • Why this probability: utility rates are rising with data-center load, though Georgia Power has a base-rate freeze through 2028.

Investment Risk Score: 66/100 (high)

# Factor Weight Measured value Score Why
1 Liquidity & funding runway 15% Cash $112.3M at 30 Jun; at 30 Sep estimated ~$55–85M (not reported); H1 FCF ex-asset-sales ≈ −$65M; capex plan ~$60M (Columbus +10 MW) + ~$30M per Axe site; no credit line 68 Runway is 12–36 months only if capex is paced. The plan as stated needs new money within ~12 months.
2 Leverage 10% No bank debt; ~$0.33M notes; $15M contingent seller note; GPU loan gone with the SPV 18 Net cash. Lowered only by the contingent note and the undisclosed power commitment.
3 Dilution (12 months + overhang) 10% Shares 20.45M → 31.47M + 3.8M PFW (+72% incl. PFW in ~10 months); fully diluted ≈ 44.5M (+41% more) 92 Well above the +25% band, with a large convertible and RSU overhang.
4 Profitability 10% H1 operating margin −37.6%; Q2 ~0.8% only thanks to a $2.71M one-off; FY25 net loss −$9.8M on $27.0M; SBC ~21% of H1 revenue 70 Clearly loss-making on an underlying basis, improving if Columbus bills.
5 Revenue quality & concentration 10% ~100% of contracted future rent from Axe; H1 top customer 43%; current revenue mostly one-off resale 80 Top of the >35% band. Not 81+ because there is some revenue.
6 Execution & track record 10% GPU revenue start slipped twice; GPU thesis dropped in 6 months; NT 10-Q; inconsistent 10-Q figures; 4 CFOs; no restatement 76 Frequent slips and big delays, without a formal weakness.
7 Competitive position & disruption 10% CEO: "not rocket science"; 5-yr vs 10–15-yr peer terms; one live site 64 Easy to copy. Power-ready sites give only temporary advantage.
8 Cyclicality & macro sensitivity 10% Tenant is a loss-making neocloud tied to AI capex and GPU rental prices; Fed hiked 09-16 62 Rents are fixed, but the payer is highly cyclical.
9 Regulatory, legal & geopolitical 5% Texas ERCOT pause; Georgia rule only ≥100 MW; no litigation found 38 Material but contained.
10 Valuation & expectations 10% Mcap $227M ≈ 0.37x contracted value; targets avg $20.17; short interest 6.10M (~19.4%), 11.5 days to cover, rising 72 Cheap on paper, but heavy and rising short interest on a fragile story.

Overall: weighted average = 0.15×68 + 0.10×18 + 0.10×92 + 0.10×70 + 0.10×80 + 0.10×76 + 0.10×64 + 0.10×62 + 0.05×38 + 0.10×72 = 65.5 → 66.

Change vs last: 0 (66 → 66), but the components moved.

  • Method: the new method gives net cash full credit (leverage 18). The old method blended it into "financial health" at 62.
  • Pushing up:
    • Dilution and short interest are now scored separately (92 and 72).
    • Short interest rose from 4.5M to 6.1M shares.
    • Axe's own liquidity warning was found.
    • Axe's "significantly" better margin from Duos covering power was found.
  • Pushing down: nothing material.

Ranking check: DUOT (66) sits just below IPX (67) and well above REAX (50):

  • It is safer than IPX because it has net cash, a live, billing-ready site and contracted rents, where IPX is pre-revenue.
  • It is far riskier than REAX because nearly all of its future revenue rests on one micro-cap tenant, and it has diluted 54% this year.

Conclusion

  • What you are really betting on: that Axe Compute, and the unnamed hyperscaler behind it, funds and pays on schedule, and that Duos can build 55 MW of sites without selling a lot more stock below $7.50.

  • The 2–3 things that will decide the outcome:

    1. The Q3 10-Q (~mid-November, date unconfirmed). It should show the capacity-supply and power agreements, the gain or loss on the GPU sale, cash at 30 Sep, and gross or net revenue.
    2. Axe's Q3 10-Q: cash, any prepayments received, the definitive JV equity agreement, and going-concern wording.
    3. The first purchase and financing of a 55 MW site.

  • Signposts:

    • Confirming the bull case:
      1. A definitive Axe equity agreement plus a project loan.
      2. A site purchase 8-K by year-end with first readiness by Q2 2027.
      3. A named non-Axe tenant of ≥10 MW.
      4. Q3 revenue ≥$12M with the agreements filed and no large loss on the sale.
    • Breaking the bull case:
      1. An equity raise below $7.50.
      2. The Axe 10-Q showing cash falling against commitments, or going-concern language.
      3. An FY26 guide cut or a large SPV loss.
      4. Duos' power-and-facility commitment turning out to be uncapped, or larger than the colocation fee it receives.

  • New since the last report:

    • Axe's 10-Q warns its own liquidity "may be insufficient" [8]. Axe says the GPU sale "significantly" raises its margin because Duos pays the cluster's power and facility costs for five years [7]. Both point to value moving away from Duos.
    • Short interest rose from 4.5M to 6.1M shares (~19.4%, 11.5 days to cover) [30]. AGPU did not join DUOT's −22% fall from 10-07 to 10-09 [24], so the selloff was DUOT-specific. It coincided with quant downgrades (Weiss 10-08, Wall Street Zen) and Chardan's Neutral $9.50 [27,32].
    • The Q2 10-Q was filed late (NT 10-Q) and contains inconsistent figures [14,2]. No definitive Axe equity agreement, capacity-supply agreement or hosting-order exhibit has been filed [5].

Open questions & data gaps

These matter most:

  • Terms of the capacity-supply arrangement, the revised customer agreement and the separate power/facility agreement: not found (not filed). High importance: they decide revenue, margin and who owns the customer.
  • Cash at 30 Sep and gain or loss on the GPU sale: not found. High. Estimates above ($55–85M cash; $0–26M loss) are labelled as estimates.
  • Whether consensus ($162.1M FY27, 3 analysts) has been updated for the sale: unknown. Medium. Only Chardan's note post-dates the sale, and its text was not readable.
  • Source of Axe's $87.8M payoff and receipt of its $534M prepayments: not found. High for Bear #1.
  • Hyperscaler guarantee to Duos, or the end customer's identity: not found. High.
  • Columbus utility schedule above 5 MW, and the grid (ERCOT or SPP) for the Texas candidate sites: not found. Medium.
  • Bleichroeder share mismatch: 6.75M shares at 18.48% implies ~36.5M shares outstanding, against 31.47M reported. Probably counts the pre-funded warrants. Low.
  • Q1 2026 revenue consensus (~$9.5M) from report #2: not verified this run; only low-quality aggregators were found. Low-Medium: it affects the "missed" label in Bear #4.
Sources
  1. Q2 2026 earnings call transcript (8-K Ex. 99.2) — Duos Technologies via SEC EDGAR — 2026-08-17 — https://www.sec.gov/Archives/edgar/data/1396536/000107997326001150/ex99x2.htm
  2. Form 10-Q for quarter ended 2026-06-30 — Duos via SEC EDGAR — 2026-08-19 — https://www.sec.gov/Archives/edgar/data/1396536/000107997326001151/duos_10q-063026.htm
  3. Q1 2026 earnings call transcript (8-K Ex. 99.2) — Duos via SEC EDGAR — 2026-05-18 — https://www.sec.gov/Archives/edgar/data/1396536/000107997326000696/ex99x2.htm
  4. Form 10-K FY2025 — Duos via SEC EDGAR — 2026-03-31 — https://www.sec.gov/Archives/edgar/data/1396536/000107997326000405/duos_10k-123125.htm
  5. Form 8-K Items 1.01/2.01 (sale of GPUaaS SPV) — Duos via SEC EDGAR — 2026-10-05 — https://www.sec.gov/Archives/edgar/data/1396536/000107997326001318/duot_8k.htm
  6. Duos Technologies Sells Its GPU-as-a-Service Entity to Axe Compute (8-K Ex. 99.1) — Duos via SEC EDGAR — 2026-10-05 — https://www.sec.gov/Archives/edgar/data/1396536/000107997326001318/ex99x1.htm
  7. Axe Compute Takes Full Ownership of Georgia AI Cluster (8-K Ex. 99.1) — Axe Compute via SEC EDGAR — 2026-10-05 — https://www.sec.gov/Archives/edgar/data/1446159/000117184326006386/exh_991.htm
  8. Axe Compute Form 10-Q for quarter ended 2026-06-30 — Axe Compute via SEC EDGAR — 2026-08-14 — https://www.sec.gov/Archives/edgar/data/1446159/000117184326005556/agpu20260630_10q.htm
  9. Axe Compute 8-K and Ex. 99.1 ($1.5B five-year contract) — Axe Compute via SEC EDGAR — 2026-07-27 — https://www.sec.gov/Archives/edgar/data/1446159/000117184326004883/exh_991.htm
  10. Duos Signs Five-Year, 55 MW Hosting Agreements with Axe Compute (8-K Ex. 99.1) — Duos via SEC EDGAR — 2026-08-17 — https://www.sec.gov/Archives/edgar/data/1396536/000155335026000118/ex99x1.htm
  11. Axe Compute press release on 55 MW Duos agreements (8-K Ex. 99.1) — Axe Compute via SEC EDGAR — 2026-08-17 — https://www.sec.gov/Archives/edgar/data/1446159/000117184326005581/exh_991.htm
  12. Form 8-K (Columbus property purchase, Seller Contingent Earnout Note) — Duos via SEC EDGAR — 2026-07-20 — https://www.sec.gov/Archives/edgar/data/1396536/000107997326000961/duot_8k.htm
  13. Form 8-K (rail divestiture to Sandbank Acosta LLC) — Duos via SEC EDGAR — 2026-08-11 — https://www.sec.gov/Archives/edgar/data/1396536/000107997326001029/duot_8k.htm
  14. Form NT 10-Q — Duos via SEC EDGAR — 2026-08-17 — https://www.sec.gov/Archives/edgar/data/1396536/000107997326001131/duot_12b25.htm
  15. Form 8-K Item 5.02 (Recker and Patel employment agreements) — Duos via SEC EDGAR — 2026-09-17 — https://www.sec.gov/Archives/edgar/data/1396536/000155335026000123/duot_8k.htm
  16. Form 8-K Item 5.02 (CFO Christopher DeAlmeida) — Duos via SEC EDGAR — 2026-08-26 — https://www.sec.gov/Archives/edgar/data/1396536/000107997326001181/duot_8k.htm
  17. Form 8-K ($55M registered direct at $9.50) — Duos via SEC EDGAR — 2026-06-17 — https://www.sec.gov/Archives/edgar/data/1396536/000107997326000851/duot_8k.htm
  18. Form 8-K ($65M public offering at $7.50) — Duos via SEC EDGAR — 2026-03-02 — https://www.sec.gov/Archives/edgar/data/1396536/000107997326000270/0001079973-26-000270.txt
  19. Form 8-K (Hydra Host GPU-as-a-service contract) — Duos via SEC EDGAR — 2026-03-17 — https://www.sec.gov/Archives/edgar/data/1396536/000107997326000317/duot_8k.htm
  20. Form 8-K (USD.AI $98.1M GPU facility) — Duos via SEC EDGAR — 2026-06-11 — https://www.sec.gov/Archives/edgar/data/1396536/000107997326000826/duot_8k.htm
  21. Form 4, F. Douglas Recker (open-market purchase 2026-09-22) — SEC EDGAR — 2026-09-23 — https://www.sec.gov/Archives/edgar/data/1396536/000155335026000136/ownership.xml
  22. Form 4, James Nixon (open-market purchases 2026-09-18) — SEC EDGAR — 2026-09-22 — https://www.sec.gov/Archives/edgar/data/1396536/000107997326001266/0001079973-26-001266-index.htm
  23. Schedule 13G/A: Bleichroeder LP (2026-08-13), Alyeska (2026-08-14) — SEC EDGAR — https://www.sec.gov/Archives/edgar/data/1396536/000142050626001623/primary_doc.xml
  24. DUOT and AGPU daily price history — Yahoo Finance chart API — retrieved 2026-10-11 — https://query1.finance.yahoo.com/v8/finance/chart/DUOT?range=1y&interval=1d
  25. DUOT stock forecast (S&P Global consensus, price targets) — StockAnalysis — retrieved 2026-10-11 — https://stockanalysis.com/stocks/duot/forecast/
  26. DUOT forecast (Q3 revenue estimate) — TipRanks — retrieved 2026-10-11 — https://www.tipranks.com/stocks/duot/forecast
  27. Duos Technologies Group stock rated "Neutral" in new coverage at Chardan Capital — MarketBeat — 2026-10-09 — https://www.marketbeat.com/instant-alerts/analyst-duos-technologies-group-nasdaq-duot-stock-rated-neutral-in-new-coverage-at-chardan-capital-2026-10-09/
  28. Duos Technologies Group price target raised to $27.00 (Cantor) — Daily Political — 2026-08-20 — https://www.dailypolitical.com/2026/08/20/duos-technologies-group-nasdaqduot-price-target-raised-to-27-00.html
  29. Global AI stocks fall as industry chiefs call for slowing development — Khaleej Times (Reuters copy) — 2026-09 — https://www.khaleejtimes.com/business/global-ai-stocks-fall-industry-chiefs-call-for-slowing-development
  30. DUOT short interest (FINRA settlements 2026-07-15 to 2026-09-30) — Equibles — retrieved 2026-10-11 — https://equibles.com/stocks/duot/shortinterest
  31. DUOT statistics (market cap, EV, short interest) — StockAnalysis — retrieved 2026-10-11 — https://stockanalysis.com/stocks/duot/statistics/
  32. Duos Technologies Group cut to Sell at Wall Street Zen — Defense World — 2026-10-10 — https://www.defenseworld.net/2026/10/10/duos-technologies-group-nasdaqduot-cut-to-sell-at-wall-street-zen.html
  33. Duos Technologies: Not Buying Is Leaving A Lot Of Money On The Table — Seeking Alpha (A. Veurink) — 2026-08-19 — https://seekingalpha.com/article/4938141-duos-technologies-not-buying-is-leaving-a-lot-of-money-on-the-table
  34. WhiteFiber secures $865M 40 MW AI colocation deal with Nscale — Converge Digest — 2025-12-23 — https://convergedigest.com/whitefiber-secures-865m-40-mw-ai-colocation-deal-with-nscale/
  35. Digi Power X secures 40MW Cerebras colocation deal — Let's Data Science — 2026-05-18 — https://letsdatascience.com/news/digi-power-x-secures-40mw-cerebras-colocation-deal-86b4fca0
  36. Applied Digital secures $7.5B 15-year hyperscaler lease for 300 MW at Polaris Forge 3 — Whale Alert (secondary) — 2026 — https://whale-alert.io/stories/f1b701509a49a2/Applied-Digital-secures-75-billion-15-year-hyperscaler-lease-for-300-MW-at-Polaris-Forge-3-AI-campus
  37. TeraWulf secures 200MW lease with Fluidstack, Google to take stake — DatacenterDynamics — 2025-08 — https://www.datacenterdynamics.com/en/news/cryptominer-terawulf-secures-200mw-lease-with-fluidstack-google-to-take-stake-in-company/
  38. Global Data Center Trends 2026 — CBRE — 2026 (summary-level read) — https://indonesia.cbre.com/insights/reports/global-data-center-trends-2026
  39. Governor Abbott pauses Texas data center interconnections — Foley & Lardner — 2026-08-05 — https://www.foley.com/insights/publications/2026/08/governor-abbott-pauses-texas-data-center-interconnections-and-calls-for-verification-and-audit-what-data-center-developers-need-to-know-now/
  40. PSC approves rule to allow new power usage terms for data centers — Georgia Public Service Commission — 2025-01-23 — https://psc.ga.gov/site/assets/files/8617/media_advisory_data_centers_rule_1-23-2025.pdf
  41. DDC Solutions to cool high-density GPU infrastructure at Duos Edge AI's Columbus campus — PR Newswire via Finviz — 2026-09-01 — https://finviz.com/news/387502/ddc-solutions-to-cool-high-density-gpu-infrastructure-at-duos-edge-ais-columbus-georgia-campus
  42. No chillers needed for Vera Rubin server racks: Nvidia CEO — Facilities Dive — 2026 — https://www.facilitiesdive.com/news/no-chillers-needed-for-vera-rubin-server-racks-nvidia-ceo/809216/
  43. Follow the CAPEX: Q2 2026 Scoreboard — Platformonomics — 2026-07 — https://platformonomics.com/2026/07/follow-the-capex-q2-2026-scoreboard/
  44. AI cloud firm Nscale reveals revenue surge in US IPO filing — The Star (Reuters) — 2026-09-19 — https://www.thestar.com.my/tech/tech-news/2026/09/19/ai-cloud-firm-nscale-reveals-revenue-surge-in-us-ipo-filing
  45. FOMC rate decision: Fed hikes for the first time in three years — InvestingLive — 2026-09-16 — https://investinglive.com/central-banks/fomc-rate-decision-fed-hikes-for-the-first-time-in-three-years/
  46. AI Infrastructure Valuations Q3 2026 — Windsor Drake — 2026-09 — https://windsordrake.com/market-intelligence/reports/ai-infrastructure-valuations-q3-2026
  47. Duos Technologies Reports Second Quarter 2026 Results (8-K Ex. 99.1) — Duos via SEC EDGAR — 2026-08-17 — https://www.sec.gov/Archives/edgar/data/1396536/000107997326001150/ex99x1.htm
  48. Axe Compute 8-K (sale of Helomics to DataMEDS) — Axe Compute via SEC EDGAR — 2026-09-17 — https://www.sec.gov/Archives/edgar/data/1446159/000117184326006093/f8k_091626.htm
  49. Axe Compute secures $260 million three-year enterprise contract for 2,304-GPU NVIDIA B300 deployment — Barchart — 2026-04 — https://www.barchart.com/story/news/1446774/axe-compute-secures-260-million-three-year-enterprise-contract-for-2-304-gpu-nvidia-b300-deployment
  50. Transformers in 2026: shortage, scramble or self-inflicted crisis — POWER Magazine — 2026 — https://www.powermag.com/transformers-in-2026-shortage-scramble-or-self-inflicted-crisis
All DUOT 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