DOCN · DigitalOcean Holdings, Inc. · Stock research

Current Thesis

The fundamental leg delivered and the price structure broke anyway. Q2 2026 printed 2026-08-04: revenue $281.184M (+29% YoY) against $279.190M consensus, adjusted EPS $0.45 against $0.26, AI customer ARR $234M (+212% YoY), a record $93M of incremental ARR, and RPO of $894M — roughly 12x year-ago and above the ">$800M" preannounced on 2026-07-07. FY2026 revenue guidance moved to $1.170B–$1.180B from $1.130B–$1.145B; FY2026 adjusted EPS guidance to $1.35–$1.40 from $1.10–$1.20 against a $1.25 consensus. Shares fell 10.9% that session (StockStory, 2026-08-04).

The $132 weekly shelf named as the thesis-break level in the 2026-07-12 coverage did not hold. The last completed close is $124.15 (2026-08-07), 31.5% below the 52-week high of $181.29 and -24.3% over three months. The narrative on offer is second-order AI compute for the developer and SMB long tail — inference revenue +800% YoY and now over 70% of AI customer ARR, management guiding ~30% growth for 2026, ~35% by Q4 2026, and stating confidence in >50% for 2027. Between 2026-07-15 and 2026-07-24 the company placed ~12.5 million shares at $117.54, roughly $1.47B gross, and that supply is what the story is now being asked to absorb.

Bullish and bearish views on DigitalOcean Holdings, Inc.

The model's bull view on DigitalOcean Holdings, Inc. (DOCN), in brief: Acceleration is printed, not projected (2026-08-04): revenue $281.184M, +29% YoY, versus 14% in the year-ago quarter; the beat on the top line was small ($279.190M consensus) but the growth rate itself doubled. The bear view: A beat-and-raise was sold: -10.9% on 2026-08-04, and the 2026-07-07 RPO preannouncement spike also faded. Both cases follow in full.

Bull Case

• Acceleration is printed, not projected (2026-08-04): revenue $281.184M, +29% YoY, versus 14% in the year-ago quarter; the beat on the top line was small ($279.190M consensus) but the growth rate itself doubled.
• Margin leverage on the beat (2026-08-04): adjusted EPS $0.45 versus $0.26 consensus, and FY2026 adjusted EPS guidance raised to $1.35–$1.40 from $1.10–$1.20 — a larger raise proportionally than the revenue raise.
• Backlog is real and contracted (2026-08-04): RPO $894M, ~12x year-ago, following the 2026-07-07 preannouncement of ">$800M" plus 20 MW of capacity contracted for 2027/2028.
• Mix shift toward inference (Q2 call, 2026-08-04): AI customer ARR $234M (+212% YoY), inference +800% YoY and >70% of AI customer ARR, $1M+ customer ARR $259M (+214% YoY) — the AI revenue is recurring workload, not one-off training rentals.
• Balance sheet simplified (closed 2026-07-24): ~$472M principal of 2030 convertible notes retired, ~$153M outstanding, funded by the equity placement; the CFO framed the note retirement as covering 96% of the shares issued.
• Sell-side has not capitulated (2026-08-05): Citi raised to $190 (Buy) and Barclays to $161 (Overweight) on the print; KeyBanc initiated at $200 on 2026-06-03.

Bear Case

• A beat-and-raise was sold: -10.9% on 2026-08-04, and the 2026-07-07 RPO preannouncement spike also faded. Two consecutive positive fundamental events failed to produce a durable bid.
• Fresh supply at a discount to where the story was being marked: ~12.5M shares at $117.54 (announced 2026-07-15, closed 2026-07-24). The clearing price for $1.47B of stock sat well below the 52-week high of $181.29.
• Concentrated-holder distribution: an Access Industries affiliate block-sold ~3.3M shares for roughly $496M in May 2026 and is reported at 21.66M shares thereafter (IBTimes, 2026-08-07).
• The bear got more bearish (2026-08-05): UBS lowered its Neutral target to $140, its second cut of the year after $175 on 2026-05-12.
• 2027 is a confidence statement: management's ">50% growth in 2027" and the ~35% Q4 2026 exit rate are guidance, not contracted revenue; the RPO covers a multi-year build. Any Q3 guide that does not step revenue toward that exit rate re-prices the multiple directly.
• Q3 guidance was a modest beat, not a blowout: $304M–$307M versus $299.925M consensus, adjusted EPS $0.28–$0.30 versus $0.28 — in line on earnings.

Setup & Price Structure

• Life-cycle: SATURATED. Dated by 2026-08-04, when a beat with raised full-year revenue and EPS guidance produced a 10.9% single-session decline; by 2026-07-15/2026-07-24, when ~12.5M shares cleared at $117.54; and by the 2026-08-07 close of $124.15, still 31.5% under the 52-week high of $181.29 after three months of -24.3%. Coverage is broad, targets are widely published, and incremental good news is not buying price.
• The prior structure is gone. The post-preannouncement gap base that supported a $132 weekly level in July was lost; the reference now is the July placement price of $117.54, the level at which the last large block of stock actually changed hands.
• Momentum is neutral, not washed out. RSI(14) 52.6 on 2026-08-07, three sessions after a double-digit down day — the drawdown has produced no oversold reading to mean-revert from.
• Positioning observables: every published target sits above spot — UBS $140 (2026-08-05, lowered), Barclays $161 (2026-08-05, raised), Citi $190 (2026-08-05, raised), KeyBanc $200 (2026-06-03), BofA and Canaccord $200 (2026-05-06) — against a $124.15 close. The dispersion between the low and high target is now $50 wide. Russell 1000 membership since 2026-06-30 changed the passive cohort. Retail-facing coverage clustered around the print (Benzinga pre-market movers 2026-08-04, "how much you would have made" retrospective 2026-08-06), which is late-cycle attention rather than new attention.
• No imminent earnings risk: the next print is roughly three months out, so the near-term path is supply digestion and sector beta rather than a company-specific binary.

Catalyst Calendar (next 30 days)

• 2026-08-14 — 13F deadline for holdings as of 2026-06-30. Shows institutional positioning into the Russell 1000 add, though it pre-dates the July placement and so will not reveal who took the ~12.5M shares.
• No confirmed company-dated event inside the window. No guidance update, product event or regulatory date is scheduled between 2026-08-08 and 2026-09-07 on the public calendar as of 2026-08-07.
• ~2026-11-04 (est.) — Q3 FY2026 print. Tests the $304M–$307M guide and whether revenue steps toward the ~35% Q4 exit growth rate management described on 2026-08-04.

What Would Change Our Mind

The structure that mattered has already failed once: the $132 weekly level published on 2026-07-12 broke, which is why the reference has moved down to where the July equity placement cleared. The next break is a weekly close below $117 — beneath the $117.54 placement price, the last observable level at which roughly $1.47B of stock found buyers. That would say the marginal holder created in July is underwater and the base has no floor beneath it until the pre-2026 range. The upside case reverses on evidence, not on price alone: RPO extending past $894M at the Q3 print, a Q4 revenue implication consistent with ~35% growth, and UBS moving off $140 would each argue the de-rating was supply-driven and temporary. The opposite — a Q3 guide implying sub-30% Q4 growth, or another equity or convertible issuance before that print — would confirm that the capacity build is being financed out of the share count.

Correlation Notes

• Co-moves with the neocloud/GPU-capacity complex (CRWV, NBIS) and with memory and AI-infrastructure names such as MU; the May 4–8 large-cap gainers list grouped DOCN with MU and RKLB.
• Index cohort changed on 2026-06-30 with the move from the Russell 2000 to the Russell 1000; small-cap flow sensitivity persists but the passive bid now tracks a different rebalance.
• Customer base skews to startups and SMBs rather than hyperscaler offtake, so venture funding conditions and small-business IT spend matter more here than a single hyperscaler capex line.
• The July placement ties the equity to AI-infrastructure financing conditions generally: a widening in convertible/high-yield spreads for compute buildouts raises the cost of the 2027/2028 capacity plan.