Ubiquiti UI

Reference table

So why not buy it? Three reasons, all specific. First, the March quarter broke a four-print beat streak — EPS $3.88 against $4.29 expected (−9.6%), revenue $788.2M against $814.3M (−3.2%), and a sequential decline from December's $814.9M. That is the first sequential fall in the file and it is unexplained by anything in the data. Second, after a 49% drawdown the stock still costs ~31.8x FY27 earnings — the de-rating has not made it cheap, only less expensive. Third, and most important: there is almost no information here. The estimate block carries one analyst in every year; the price-target block reports one number ($672) as high, low, median and consensus; there are no forecasts beyond FY27; and a search of 51,928 knowledge-base entries returns zero claims about the company. Excellent business, unresolved downtrend, and nothing to triangulate against. Hold, and add materially lower.

1. Data integrity — the thinnest information set in the batch

There is no corruption of the kind found elsewhere in this batch. There is something arguably more consequential: near-total absence of forward information. Stated plainly:

SEVERE GAP — the entire estimate block rests on one analyst. Every year in est — FY2020 through FY2027 — reports numAnalystsRevenue: 1 and numAnalystsEps: 1. There is no consensus for this company in this feed; there is a single contributor's model. Compare the same field for Twilio in this batch (24 revenue contributors, 18 on EPS) or Illumina (13 and 12). Every forward figure in this dive derived from est carries that caveat, and the base case is deliberately built on a haircut to it rather than on the number itself.

SEVERE GAP — no forecasts beyond FY2027. The est block terminates at 2027-06-30. There is no FY28, FY29 or FY30 line at all. For a company whose entire investment case is multi-year compounding, the absence of any forecast beyond nineteen months is a material limitation, and it means the bull case in §6 is constructed from our own extrapolation, explicitly labelled as such.

SEVERE GAP — the price target is a single number. pt reports targetHigh: 672, targetLow: 672, targetMedian: 672, targetConsensus: 672. One analyst, one target, presented four ways. Note the internal inconsistency: grades records 21 raters (6 Buy, 10 Hold, 5 Sell). So twenty-one firms rate the stock and one publishes a target into this feed. We quote $672 as "street consensus" in the front matter because the field requires it, but it should be read as a single data point, not a consensus.

FLAGGED — duplicated estimate values across unrelated years. est reports epsAvg: 9.65 for FY2021 (2021-06-29) and epsAvg: 9.65 for FY2025 (2025-06-30). Two unrelated fiscal years carrying an identical estimate to the cent is a vendor artifact, not a coincidence. Relatedly, the historical rows carry implausibly wide EBITDA ranges — FY2023 low $545.6M against high $818.3M, a 50% spread — that look like placeholder scaling rather than analyst dispersion. All historical est rows are treated as unreliable; only FY2026 and FY2027 are used, and only for revenue and EPS.

CLEAN — the reported financials. The actuals are internally consistent and cross-check well:

• Trailing revenue from inc_q (Q4 FY25 $759.152M + Q1 FY26 $733.773M + Q2 $814.867M + Q3 $788.199M) = $3,095.99M. ratios_ttm revenuePerShare of $51.156 × 60.52M shares = $3,095.9M. Confirmed.
• Trailing net income ($266.705M + $207.876M + $233.610M + $233.914M) = $942.11M. ratios_ttm netIncomePerShare of $15.567 × 60.52M = $942.1M. Confirmed.
• FY25 annual revenue $2,573.545M with EPS $11.77 on 60.534M diluted shares: $711.922M ÷ 60.534M = $11.76. Confirmed.
• inc_q[0] (2026-03-31) — the usual offender — is clean: revenue $788.199M, gross profit $370.711M (47.0%), operating income $290.821M, net income $233.914M, EPS $3.87 on 60.572M diluted shares ($233.914M ÷ 60.572M = $3.862). Confirmed. No sign flips, no zeroed share counts, no duplicated rows.

NOTE — fiscal year ends 30 June. inc_q labels the March 2026 quarter as "Q3" and the December 2025 quarter as "Q2". All references in this dive use fiscal labelling with calendar dates attached to avoid confusion.

NOTE — net cash is implied, not stated. bal_a FY25 (30 June 2025) shows total debt $297.4M against cash of $149.7M, i.e. net debt of $147.7M at the fiscal year end. But km_ttm reports an enterprise value of $32,904.3M against a market capitalisation of $33,206.0M — enterprise value below market capitalisation, implying roughly $302M of net cash — and netDebtToEBITDATTM of −0.27x confirms the sign. The company moved from net debt to net cash during fiscal 2026. We use ~$302M of net cash and label it as implied.

2. The economics — why this business is worth studying regardless of the price

Three of these deserve to be pulled out and stated on their own.

Total operating expense of 10.3% of revenue is the whole company. A conventional enterprise networking vendor spends heavily on direct sales, channel margin, marketing and professional services. Ubiquiti does not: it sells through an installer community and online distribution, and the saving lands almost entirely in operating margin. That is the mechanism, and it is why a 46% gross margin produces a 36% operating margin — a conversion rate that most software companies cannot match.

Stock compensation of 0.23% of revenue means reported earnings are real earnings. This is the sharpest contrast available inside this batch. Twilio's FY27 consensus EPS is roughly half added-back stock compensation; Ubiquiti's is essentially none. When this dive says 31.8x FY27 earnings, that figure needs no burdening adjustment — which makes it a genuinely different 32x from the one quoted elsewhere.

The deleveraging is nearly complete and largely unremarked. Total debt: $1,138.3M (FY23) → $757.3M (FY24) → $297.4M (FY25) → net cash today. Equity went from negative $115.7M (FY23) to $95.1M (FY24) to $668.3M (FY25). The company had negative book value three years ago because of aggressive historical buybacks ($618.1M in FY22 alone); it has rebuilt the balance sheet entirely from operating cash flow while growing revenue 33.5%. Interest expense fell from $16.8M in the June 2024 quarter to $0.7M in the March 2026 quarter — a 96% reduction, which is roughly $0.90 per share of annualised earnings recovered from the balance sheet alone.

3. Segments — one engine, one flat tail

Enterprise Technology is the company. It compounded from $409.7M (FY17) to $2,254.3M (FY25) — 23.7% annually over eight years — and grew 39.4% in FY25 alone. This is the UniFi family named in the profile: cloud gateways, wireless access points, switches, video surveillance (UniFi Protect), door access (UniFi Access), telephony (UniFi Talk).

Service Provider Technology is a flat tail. It peaked at $623.2M in FY21 and has sat between $310M and $375M for four consecutive years. It contributes nothing to growth and is now 12.4% of revenue. Any forward model should treat it as a constant.

Geography (FY25): North America $1,295.5M (50.3%, of which the United States is $1,193.3M), EMEA $999.4M (38.8%), Asia-Pacific $168.8M (6.6%), South America $109.8M (4.3%). EMEA grew 35.0% in FY25 ($740.1M → $999.4M) and North America 36.9% — genuinely broad-based, not a single-region story. Note that FY25 is the first year the file breaks out the United States separately from North America, so year-over-year US comparison is unavailable.

4. The March-quarter stumble — the thing this dive is actually about

Five consecutive prints, actual versus estimate:

Four enormous beats, then a double miss. And the revenue figure declined sequentially — $814.9M in December to $788.2M in March, the first sequential fall anywhere in the eight quarters the file records.

The market's reaction tells you how much of the price was the beat streak: the three-month return is −46.2% against +5.1% for the S&P. Roughly half the market capitalisation was removed on one quarter.

What the file can and cannot tell us about why. It can tell us the miss was not a margin event: gross margin in the March quarter was 47.0% ($370.7M on $788.2M), the highest in the file, up from 45.9% in December and 45.7% in September. Operating margin was 36.9%, also the highest. Pricing and mix held; volume did not. It can also tell us that the estimate the company missed had risen very fast — from $2.93 for the September quarter to $4.29 for March, a 46% increase in the bar in two quarters, driven by the preceding beats. A meaningful part of this "miss" is an estimate that chased four consecutive surprises.

It cannot tell us the cause. There is no segment-level quarterly disclosure, no backlog figure, no channel-inventory metric, and — critically — no second analyst to compare interpretations with. This is the single clearest instance in this batch where the absence of coverage is itself the risk.

One relevant data point on the demand question: inventory stood at $675.1M at the June 2025 fiscal year end, with days-inventory-outstanding of 142.8 — roughly five months of stock. That is inherent to a build-ahead, low-channel-cost model, but it is also the line item that reveals a demand misjudgement first. The file does not carry a quarterly balance sheet, so we cannot see whether inventory built through the March quarter. We flag that as a specific unknown and a specific thing to check on 2026-08-21.

5. Financials — the receipts

• Revenue: FY25 $2,573.5M (+33.5%); trailing $3,096.0M. Nine months of FY26: $2,336.8M against $1,814.4M — +28.8%.
• Profitability: FY25 operating income $836.3M (32.5% margin), net income $711.9M (27.7%), EPS $11.77. Trailing operating margin 35.8%, net margin 30.4%, trailing EPS $15.57.
• Gross margin: 38.4% (FY24) → 43.4% (FY25) → 46.0% trailing → 47.0% (March quarter). Expanding through the revenue stumble, which is the most reassuring single fact in the file.
• Operating expense: R&D $169.7M (6.6% of FY25 revenue; 6.4% trailing), SG&A $111.5M (4.3%; 3.9% trailing). Total 10.3% of revenue.
• Cash flow: FY25 operating cash flow $640.0M, capex $12.6M, free cash flow $627.4M. FY24 $541.5M / $12.0M / $529.5M. FY23 was negative (−$145.4M operating) — the working-capital cost of the inventory build. Trailing free cash flow per share $12.39.
• Balance sheet (30 June 2025): cash $149.7M, total debt $297.4M (all short-term), net debt $147.7M at that date — but the trailing enterprise value of $32,904.3M sits below the $33,206.0M market capitalisation, implying ~$302M of NET CASH today, with netDebtToEBITDA of −0.27x. Current ratio 3.56x, quick ratio 1.90x. Interest coverage 128.4x.
• Inventory: $675.1M (FY25), up from $462.0M (FY24). Days inventory 142.8, cash conversion cycle 136.3 days. The working-capital intensity is the cost of the distribution model.
• Returns: ROE 101.5%, ROIC 71.3%, ROA 54.9%, return on capital employed 83.8%. These are inflated by the small equity base ($668.3M against $942.1M of trailing net income) — a legacy of historical buybacks that drove book value negative in FY23 — but even normalised they are exceptional.
• Dividend: $3.20 per share annually, a 0.58% yield at $548.66, payout ratio 19.3%. Modest, growing, fully covered.
• Buyback: essentially none currently ($0.8M in FY25) after $618.1M in FY22. The company chose deleveraging over repurchase, which was correct at $1.1B of debt and is worth watching now that it is in net cash — a resumed buyback at these prices would be a meaningful signal.

6. Valuation — what is priced at $548.66

At $548.66 (market cap $33.21B, net cash ~$0.30B, EV ~$32.90B):

The FY26 figure is essentially already known. Nine months of actual EPS is $11.17 ($3.44 + $3.86 + $3.87); adding the $4.03 Q4 consensus gives $15.20, within a cent of the single analyst's $15.252. So the real forward question is FY27, and it rests on one model.

The base case therefore uses a haircut rather than the estimate. The single analyst models FY27 revenue growth of +15.7% and EPS growth of +13.0%. Given that the March quarter declined sequentially and that we have no visibility into why, we apply a haircut to $16.50 of FY27 EPS (a ~4% reduction from $17.234, implying FY27 EPS growth of ~8.6% rather than 13.0%) and build the scenarios from that. This is our number, not the vendor's, and it is labelled as such throughout.

6a. What today's price assumes (the inversion)

Reverse-engineering $548.66 into falsifiable claims.

• Revenue reaches ~$3,688M in FY27 — i.e. growth of 15.7% resumes after a sequential decline. This is the single analyst's estimate. At 8.9x FY27E EV/revenue the price is broadly consistent with it. Falsifiable at the 2026-08-21 print, where consensus revenue of $850.5M would represent a 7.9% sequential advance from the March quarter's $788.2M — the first test of whether the stumble reverses.
• Gross margin holds at or above 46%. It has expanded every year (38.4% → 43.4% → 46.0%) and reached 47.0% in the March quarter. This is the least fragile assumption in the price and the one that is currently being met most convincingly.
• Total operating expense stays near 10% of revenue. The 35.8% operating margin depends entirely on it. R&D at 6.4% of revenue is genuinely low for a company competing on product; the assumption is that the community-distribution model continues to substitute for a sales force. Falsifiable quarterly and, so far, holding.
• The market keeps paying ~32x forward earnings for a mid-teens grower. After a 49.4% drawdown the multiple is still 31.8x FY27E. This is the most fragile assumption in the price. A 15% grower with no dilution and net cash is a defensible 25-30x; 32x requires the growth rate to re-accelerate toward the 24-34% of FY25 and the first nine months of FY26. The March quarter is direct evidence against it.
• Nothing structural broke in the March quarter. The price does not embed a demand turn. Since the file offers no explanation and no second opinion, this is an assumption held on faith rather than on evidence — which is precisely why the verdict is Hold rather than Buy at a 49% discount.

6b. The return bridge (why the multiple moves)

Expected return over the eighteen-to-twenty-four-month horizon: EPS growth (~9-13% on our haircut versus the single estimate) + multiple drift + shareholder yield (0.58% dividend, no buyback).

Our base case assumes modest multiple COMPRESSION. We apply 31.5x to our haircut FY27 EPS of $16.50 against the 31.8x the market currently pays for the single analyst's $17.234 — but the EPS number itself is 4% lower, so the effective assumption is that the market pays roughly the same multiple on a slightly lower earnings base. The reason we do not assume expansion is that the growth rate is decelerating on every available measure: FY25 +33.5%, nine months of FY26 +28.8%, FY27 estimated +15.7%, and the most recent quarter negative sequentially. Multiples follow growth rates down.

So the base-case return of roughly −5% is earnings growth of ~9% offset by an estimate haircut and a starting price that has not yet de-rated to match the new growth rate. Note the important structural point: because stock compensation is 0.23% of revenue and capital expenditure is 0.35%, essentially all of this EPS growth is distributable — there is no add-back adjustment eroding it, which is not true of most companies trading at this multiple. That quality argument is real and is why the bear case is only −34% rather than worse.

The bull case explicitly requires multiple expansion or a growth re-acceleration: it assumes FY27 growth lands near 15% and FY28 sustains it, with the market paying 38x. That is the fragile leg, and it should only be underwritten after two consecutive quarters of restored sequential growth.

6c. Variant perception (where we differ, what would surprise)

• We are more cautious than the single available estimate on FY27. It models +15.7% revenue growth into a quarter that just declined sequentially, and the model was demonstrably chasing the prior beats (the bar rose from $2.93 to $4.29 in two quarters). Our haircut to $16.50 of FY27 EPS is the concrete expression of that caution. Negative surprise, watchable 2026-08-21: revenue below ~$800M, a second consecutive sequential decline.
• We think the gross-margin expansion through the revenue stumble is being underweighted. Margin reached its highest level in the file (47.0%) in the quarter the stock fell 46% over three months. Volume disappointments accompanied by margin expansion are demand-timing events far more often than they are competitive ones. If the August print restores sequential revenue growth at 46%+ margin, this drawdown will read as an overreaction. Positive surprise, watchable 2026-08-21: revenue above $850M with gross margin ≥46%.
• We think a resumed buyback is a live and underpriced possibility. The company repurchased $618.1M in FY22 when it had debt, then stopped entirely to deleverage. It is now in net cash with $627M of annual free cash flow, and the stock is 49% off its high. A material repurchase authorisation would be a strong signal of management's own view of value and is entirely within their control. Positive surprise, watchable in cash-flow statements.
• Where we have NO edge, and say so plainly: on the cause of the March miss. The file contains no segment-level quarterly data, no backlog, no channel-inventory figure, one analyst and zero knowledge-base claims. We cannot distinguish a demand air-pocket from a demand turn, and neither can anyone using this data set. Per house rule, a no-edge situation defaults toward Hold — which is exactly what has happened, and it is the honest reason this is not a conviction buy at a 49% discount.

Synthos fair values

• Bear ~$360 — 22x our haircut FY27 EPS of $16.50 = $363. Assumes the March quarter marked a demand turn, FY27 growth lands near zero to mid-single-digits, and the market re-rates toward a mature-hardware multiple. Approaches the fifty-two-week low of $388.83 and breaks it. −34%.
• Base ~$520 — 31.5x our haircut FY27 EPS of $16.50 = $519.8; cross-checked at 33.4x trailing EPS of $15.57 = $520. Assumes the stumble is an air-pocket, growth resumes at high single to low double digits, gross margin holds at 46%+, and the multiple holds roughly where it is. −5%.
• Bull ~$755 — 38x an FY28 EPS of ~$19.84 (our extrapolation: FY27 revenue of $3,687.7M growing 15% to ~$4,241M at the FY27 net margin of 28.3%, giving ~$1,200M of net income on 60.5M shares). This figure is our own construction — the file contains no FY28 estimate of any kind. Assumes growth re-accelerates and the market pays a compounder multiple again. +38%.

Base sits ~5% below spot, with a −34% bear and a +38% bull. Roughly symmetric, centred slightly below the current price — the signature of a Hold, and specifically of a Hold whose asymmetry improves sharply on any decline.

7. Technicals — the most clearly broken chart in the batch

• Price $548.66, −49.4% from the $1,084.50 fifty-two-week high. Fifty-two-week low $388.83, so the stock sits 41.1% above its low and roughly midway through the range.
• Below a falling 50-day average ($556.79) by 1.5% and 18.2% below a falling 200-day average ($671.09). Both averages are above the price and both are declining — the textbook definition of a broken trend, and the only such structure in this batch.
• RSI 49.6 — dead neutral. There is no oversold signal, which means the decline has stabilised without producing an exhaustion low. MACD is negative at −8.14.
• Returns: −46.2% over three months (S&P +5.1%), −0.5% over six months, +26.0% over twelve months. The damage is entirely concentrated in the period since the 2026-05-08 miss — over twelve months the stock is still ahead of the index.
• Read: the 50-day at $556.79 is 1.5% above spot and is the first level to reclaim; the 200-day at $671.09 is 22% above and is the level that would confirm the drawdown is over. There is no completed base. The practical implication: a buyer here is early by the chart, and the fundamentals do not yet compensate — which is the arithmetic case for waiting for either the print or a lower price.

8. Insider activity and ownership

The insider record in this file is sparse and unremarkable:

• Kevin Radigan (Chief Accounting Officer) sold 500 shares at $680.53 on 2026-05-13 (leaving 2,268), and 350 shares at $576.58 on 2025-11-12 (leaving 2,768). He received a 489-unit restricted-stock award on 2025-08-05.
• No other officer or director transactions appear in the file.

Two things worth stating. First, there is no evidence of concentrated insider selling into the highs, which given a 49% drawdown from $1,084 is a meaningful absence — the total recorded disposals are 850 shares. Second, there is also no insider buying into the drawdown, which would have been the strongest possible signal and is not present.

Governance note: CEO Robert J. Pera is the founder and has run the company since before its 2011 listing. The file does not disclose his ownership percentage, and we make no claim about it. What the file does show is a company run with unconventional discipline — 10.3% operating expense, 0.23% stock compensation, no analyst-relations effort evidenced by the single price target, minimal public disclosure. That is a coherent founder-operator posture and it is part of why the economics are what they are. It is also a genuine key-person consideration, and it is the reason the coverage vacuum described in §1 exists at all.

9. Knowledge base — zero claims, and a textbook name collision

A search of 51,928 knowledge-base entries for "UI" and "Ubiquiti" returned 26 string matches. Every single one refers to user interface design. A representative sample, verbatim:

• No Priors (2025-10-02, conviction 75): "The durable moat is the proprietary system-of-record data, not the app UI; models won't become the system of record — locked databases will."
• a16z (2026-04-14, conviction 85): "AI dissolves classic software moats — you can now throw money at problems, and lock-in from migration/data/UI is gone as AIs flexibly move data and drive UIs."
• Cognitive Revolution (2025-11-23, conviction 70): "AI will dynamically generate interfaces, but humans want predictability; constantly-changing UIs and chat-based form input fail — so dynamic generation won't dominate."
• Arthur Hayes (2022-12-13, conviction 75): "UX/UI is the biggest barrier to mainstream crypto adoption..."

Not one of the twenty-six concerns Ubiquiti Inc. The search term "Ubiquiti" returns zero matches independently.

kb_claim_count: 0. kb_breadth: 0. This is a pure name collision of exactly the type flagged in round one ("Ferguson", "MPC"), and the correct count is zero. No tracked voice has ever expressed a view on this company.

Consequence for the verdict, stated plainly: this dive is fundamentals- and quant-driven only — an accepted and precedented outcome, never a reason to manufacture conviction. Combined with the single-analyst estimate block and the single price target documented in §1, it means Ubiquiti is the least-triangulated name in this batch: there is no panel view, no consensus, and no second opinion of any kind. That is not an argument against the company. It is an argument for a smaller position and a lower entry price, and it is a substantial part of why the verdict is Hold rather than Buy at a 49% discount.

10. Moat and competitive position

Ubiquiti's moat is a cost structure competitors cannot replicate without abandoning their own distribution. An incumbent enterprise networking vendor carries a direct sales force, a channel partner programme with margin attached, professional services, and marketing — perhaps 30-40% of revenue. Ubiquiti carries 10.3%. To match Ubiquiti's pricing, an incumbent would have to dismantle the distribution machine that constitutes its own relationship with its customers. That is a classic disruptor's asymmetry and it has produced 23.7% annual Enterprise Technology compounding for eight years.

The reinforcing mechanism is the installer community: network professionals who learn the product, standardise on it, and recommend it, substituting for both the sales force and the marketing budget. This is a real network effect on the supply side of installation labour.

The bounds of the moat: it is a price-and-distribution moat, not a technology moat. R&D at 6.4% of revenue is low in absolute terms, and the products compete in categories (wireless, switching, surveillance, access control) where feature parity is achievable. The moat protects margin, not category leadership. And the model's working-capital intensity — 142.8 days of inventory — is the price paid for skipping the channel.

Moat: real, unusual and durable in mechanism; bounded by an existing addressable market. It justifies a premium multiple over commodity hardware. Whether it justifies 32x is the valuation question, and the March quarter is the first evidence that it might not.

11. Management and capital allocation

Founder-CEO Robert J. Pera. 1,667 employees.

The allocation record over the file's span is deliberate and, in hindsight, correct in sequence:

• FY22: $618.1M of buybacks — aggressive repurchase that drove book value to negative $115.7M by FY23 and left $1,138.3M of debt.
• FY23-FY26: deleveraging. Debt from $1,138.3M → $757.3M → $297.4M → net cash. Buybacks essentially suspended ($0.4M in FY24, $0.8M in FY25). Interest expense fell from $16.8M per quarter to $0.7M, recovering roughly $0.90 per share of annualised earnings.
• Throughout: a growing dividend, now $3.20 annually at a 19.3% payout ratio.
• Throughout: negligible capital expenditure ($12.6M in FY25) and negligible stock compensation ($7.2M).
• No acquisitions in any year of the file. Growth is entirely organic.

This is disciplined, unfashionable, shareholder-aligned capital allocation — buy back when leverage permits, repair the balance sheet when it does not, pay a modest dividend, dilute nobody. The forward question is whether the buyback resumes now that the company is in net cash with $627M of annual free cash flow and the stock 49% off its high. That decision would be the single most informative signal management could send, and it is worth more attention than any product announcement.

12. Verdict, kill-criteria and flip conditions

Hold. Ubiquiti is, on the operating numbers, the best business in this batch: a 30.4% net margin and 71.3% return on invested capital produced by 1,667 people spending 10.3% of revenue on all operating costs combined, with capital expenditure at 0.35% of sales, stock compensation at 0.23%, a balance sheet repaired from $1.1B of debt to net cash, and a gross margin that has expanded every year and reached 47.0% in the most recent quarter. Reported earnings here are very nearly owner's earnings, which is not true of most companies at this multiple.

It is also, on the setup, a name to be patient with. The March quarter broke a four-print beat streak with a 3.2% revenue miss, a 9.6% earnings miss and the first sequential revenue decline in the file — and nothing in the data explains it. The stock is below both a falling 50-day and a falling 200-day average with no completed base. After a 49.4% drawdown it still costs ~31.8x FY27 earnings. And the information environment is the thinnest in the batch: one analyst producing the entire forward estimate, one price target reported four ways, no forecasts beyond FY27, and zero knowledge-base claims.

Base fair value $520, five percent below spot. We do not add here.

Staged entry:

1. No addition at spot ($549).

2. Tranche 1 (1-2%) below ~$450 — roughly 27x our haircut FY27 EPS, approaching the fifty-two-week low ($388.83) zone, at a price where the base case pays ~16% instead of −5% and the bear case is largely in the price.

3. Tranche 2 (to ~3%) below ~$400 — near the fifty-two-week low, at ~24x, which for a net-cash business with a 46% gross margin and no dilution is a genuinely attractive entry irrespective of whether FY27 growth is 8% or 15%.

4. Alternative path: a first tranche on a 2026-08-21 print above ~$850M of revenue with gross margin ≥46% and a reclaim of the 50-day average ($557) — buying confirmed strength rather than unconfirmed weakness.

Pre-registered KILL / do-not-add criteria:

• A second consecutive sequential revenue decline on 2026-08-21 (revenue below ~$788M) — this would convert the March quarter from an air-pocket into a turn and would invalidate the entire FY27 estimate the price rests on.
• Gross margin below ~43% — the 38.4% → 47.0% expansion is the strongest evidence that the pricing position is intact; reversal would recast the revenue stumble as competitive rather than cyclical.
• Total operating expense rising above ~14% of revenue — the 10.3% figure is the entire mechanism of the 35.8% operating margin. Any structural increase (a build-out of direct sales, for example) would compress the model toward peer economics.
• Inventory rising materially above the $675.1M FY25 level while revenue is flat or falling — the clearest available signal of a demand misjudgement, and the specific line to check on 2026-08-21.
• Departure of the founder-CEO — a genuine key-person risk for a company run this unconventionally, with no succession information in the file.

Pre-registered FLIP TO BUY:

• A price below ~$450 with gross margin held at 46%+ — the single most likely path. Nothing about the company needs to change.
• Restored sequential revenue growth for two consecutive quarters (Q4 FY26 and Q1 FY27) with margin intact — this would prove the March quarter was timing, and would justify paying today's multiple.
• A material buyback authorisation — management, now in net cash with $627M of annual free cash flow, telling us what they think the stock is worth. This would be the strongest available signal and it is entirely within their control.
• Coverage deepening beyond the single analyst — not a fundamental event, but it would materially reduce the information risk that is currently a real component of our required discount.

Where UI fits in the Synthos Framework Portfolio. The networking / hardware sleeve, as a high-quality-compounder holding — target 2-3% at the right price, currently 0%. It is genuinely complementary rather than redundant to optical-infrastructure exposure: Ubiquiti sells enterprise-premises networking to businesses and installers, not datacentre interconnect to hyperscalers, so the demand driver is commercial-property and enterprise IT refresh rather than compute capital expenditure. Its most useful portfolio property is the earnings quality — with 0.23% stock compensation and 0.35% capital intensity, it is one of the few names where a quoted P/E requires no adjustment. Logged as a tracked Synthos call (Hold, do not add; Buy trigger at $450) as of 2026-08-04 at $548.66.

Single biggest risk: an unexplained demand stumble underwritten by one analyst. The March quarter missed both lines and declined sequentially; the file offers no cause, no segment detail, no backlog, no second forecast and no expert view. If the August print confirms a demand turn, there is no coverage depth to cushion the estimate revisions, and the bear case ($360, −34%) becomes the working case very quickly.

Most fragile assumption in the price: that the market keeps paying ~32x forward earnings for a business whose growth rate has gone from +33.5% (FY25) to +28.8% (nine months FY26) to a modelled +15.7% (FY27) to negative sequentially in the most recent quarter. Multiples follow growth rates, and this one has not yet finished following.

Provenance and disclosures

• Traceability: ZERO tagged knowledge-base claims name Ubiquiti. A search across 51,928 entries returned 26 string matches on "UI", every one of which refers to user interface design — a pure name collision of the type flagged in round one. The term "Ubiquiti" returns no matches independently. kb_claim_count: 0, kb_breadth: 0. No tracked voice has expressed a view on this company. This dive is fundamentals- and quant-driven only — an accepted and precedented outcome, and never a reason to manufacture conviction.
• Data as-of: fundamentals 2026-03-31 (Q3 FY26; fiscal year ends 30 June) · estimates, prices, technicals and knowledge-base claims 2026-08-04. All market data drawn exclusively from scripts/deepdive/vti_data/UI_data.json.
• Severe information gaps stated, not papered over: (a) the entire est block rests on ONE analyst — numAnalystsRevenue: 1 and numAnalystsEps: 1 in every year from FY2020 to FY2027; (b) there are no forecasts of any kind beyond FY2027, so the bull case in §6 is built from our own extrapolation and is explicitly labelled as such; (c) pt reports a single value of $672 as targetHigh, targetLow, targetMedian and targetConsensus simultaneously, while grades records 21 raters — one firm publishes a target into this feed. We report $672 as "street consensus" because the field requires a value, and it should be read as one data point.
• Vendor data flagged as unreliable: historical est rows carry a duplicated epsAvg of exactly 9.65 for both FY2021 and FY2025, and implausibly wide historical EBITDA ranges (FY2023: $545.6M low against $818.3M high). All historical estimate rows were discarded; only FY2026 and FY2027 revenue and EPS were used, and the FY2027 EPS was then haircut from $17.234 to $16.50 (our number, labelled throughout) to reflect the March-quarter sequential decline.
• Vendor data checked and found sound: inc_q[0] (2026-03-31) reconciles exactly — net income $233.914M ÷ 60.572M diluted shares = $3.862 against the stated $3.87 EPS. Trailing revenue from the four quarters ($3,095.99M) matches ratios_ttm revenuePerShare × shares ($3,095.9M). Trailing net income ($942.11M) matches netIncomePerShare × shares ($942.1M). No corruption found in the reported financials.
• Net cash is implied, not stated: bal_a at the 30 June 2025 fiscal year end shows net debt of $147.7M, but km_ttm reports enterprise value ($32,904.3M) below market capitalisation ($33,206.0M) and netDebtToEBITDATTM of −0.27x, implying roughly $302M of net cash today. We use the implied figure and label it as such.
• Fair-value method: scenario multiples applied to our haircut FY27E EPS of $16.50 (from the single analyst's $17.234). Bear 22x, base 31.5x (cross-checked at 33.4x trailing EPS of $15.57), bull 38x an FY28 EPS of ~$19.84 which is our own extrapolation — the file contains no FY28 estimate. Stated arithmetic only; no discounted-cash-flow model was built.
• Earnings-quality note: unlike most names at this multiple, no burdening adjustment is required. Stock compensation is 0.23% of revenue ($7.2M) and capital expenditure 0.35% ($12.6M), so reported EPS is very nearly owner's earnings. The 31.8x quoted here is a materially different 32x from the one quoted for Twilio in this same batch.
• Timing caveat: written seventeen days before the Q4 FY26 print (2026-08-21). The recommendation is explicitly to wait for it or for a materially lower price.
• Not investment advice. Independent research, educational and informational only, never personalised.
• Version: 2026-08-04-full.