BTBT · Bit Digital, Inc. · Stock research

Current Thesis

The two legs of the wrapper have separated. WhiteFiber (WYFI) — closed $29.52 on 2026-08-14, up 4.18% on the day and back above both the $27.57 print of 7/21 and the five-analyst $27 median target, after its own Q2 report in the week of 2026-08-10 (revenue $28.84M vs $19.72M consensus, EPS -$0.39 vs -$0.45). Bit Digital itself printed Q2 on 2026-08-13: revenue $32.113M against a $23.709M consensus, GAAP EPS $(0.31) versus $(0.05) expected, adjusted EPS $(0.06) versus $(0.04). The parent closed $1.57 on 2026-08-14, 62.8% below the $4.22 52-week high, with RSI(14) at 61.3 — the first firm momentum reading inside the $1.40–$1.70 band that replaced June's ~$2.01 shelf. What an investor buys here is the holdco discount closing, and management restated in the Q2 release that it does not intend to sell WhiteFiber shares in 2026. The asset side improved; the dated lever to monetize it did not appear.

Bullish and bearish views on Bit Digital, Inc.

The model's bull view on Bit Digital, Inc. (BTBT), in brief: The stake alone exceeds the parent's equity value. The bear view: Nothing scheduled closes the discount. The Q2 release repeats that the WhiteFiber position is "a long-term strategic asset" and that no shares will be sold in 2026 — the same language as the July disclosure, pushing any monetization to 2027. GAAP earnings missed by a wide… Both cases follow in full.

Bull Case

• The stake alone exceeds the parent's equity value. 27,043,750 WYFI shares × $29.52 (2026-08-14 close) ≈ $798M, against roughly $555M of market value at $1.57 on the ~353.4M share count used in the July calculation (the current 10-Q cover figure is the check on that denominator).
• The operating asset is compounding. Cloud services revenue was $23.8M in Q2, +42% sequentially and +29% year-over-year for H1 (Q2 2026 release, 2026-08-13); consolidated revenue of $32.113M beat the $23.709M consensus by ~35%.
• ETH position grew through the drawdown. 164,310.5 ETH at 2026-06-30 (marked at a ~$1,569 ETH close in the release), up from ~158,462 in the July disclosure; 75,757.5 ETH held directly at $118.9M fair value, with 73,235 ETH liquid-staked for 66,192 LsETH during the quarter.
• The loss is narrowing on the reported line. Net loss $(107.2)M in Q2 versus $(146.7)M in Q1, including a $46M non-cash impairment on liquid-staked ETH used in the WhiteFiber financing.
• WhiteFiber added dated capacity. The Krambu partnership announced 2026-08-10 targets 100MW of high-density GPU infrastructure for 2027; NC-1's 40 MW contracted IT load is guided to full run-rate billing by 2026-08-31.

Bear Case

• Nothing scheduled closes the discount. The Q2 release repeats that the WhiteFiber position is "a long-term strategic asset" and that no shares will be sold in 2026 — the same language as the July disclosure, pushing any monetization to 2027.
• GAAP earnings missed by a wide multiple. $(0.31) against $(0.05) consensus on 2026-08-13; the adjusted line also missed at $(0.06) versus $(0.04). Two quarters of nine-figure losses ($107.2M, $146.7M) sit against $83.6M of consolidated cash at 2026-06-30 — of which only $27.5M is at the parent, $56.1M at WhiteFiber.
• The ETH treasury remains offside. Spot opened $1,884.42 and traded near $1,872.97 on 2026-08-14, against a treasury average cost around $2,334 disclosed in prior quarters. Quarter-end marks, not operations, drive the headline EPS.
• The treasury is partly encumbered. 49,000 LsETH ($105.6M) is pledged as collateral and 17,192 LsETH ($27.6M) held as margin buffer against the up-to-$150M delayed-draw facility for WhiteFiber; that portion is not free-and-clear value.
• The subsidiary is levering. Coverage of WhiteFiber's Q2 filing (2026-08-12) describes widening losses and ~$230M of added debt alongside the HPC revenue growth.
• Sell-side marks are stale. The BTBT average 12-month target near $4.13 from mid-July is ~2.6x the 2026-08-14 close and has not been reset through two prints; WYFI's own five-analyst median of $27.0 sits below its $29.52 spot.

Setup & Price Structure

$1.57 on 2026-08-14, inside the $1.40–$1.70 range that has contained the tape since mid-July, roughly 26% above the $1.25 52-week low and 62.8% below the $4.22 high; three-month return -13.0%. RSI(14) at 61.3 is the strongest momentum reading of this range but has produced no close above $1.70, so no higher-high structure exists yet. The July divergence has inverted: the underlying now leads. WYFI ran from $27.57 (7/21) to $29.52 (8/14) on its Q2 beat, while the wrapper's move has been smaller — the discount widened on the asset side even as the parent ticked up.

The narrative is saturated. The 2026-08-13 print beat revenue by ~35% and the equity remains nearly two-thirds below its 52-week high with holdings marked around double the equity value; a beat that does not compress the discount is what a thin marginal bid looks like. The ETH-treasury-vehicle cohort has traded below stated asset value across both the 7/23 and 8/14 observations. The operating asset inside the wrapper is the part still accelerating — 42% sequential cloud growth, a 100MW partnership dated 2026-08-10 — and it is listed separately, which is where that flow is going.

Positioning observables, stated as observables: the quarterly print is behind (2026-08-13), so no binary sits inside the next 30 days; sell-side targets on the parent have not been marked down since mid-July while price fell; WYFI trades above its own median target; and the ETH line is partly pledged. No insider Form 4 activity is in hand for this window, and the share-count trajectory versus ~353.4M is unverified until the 10-Q cover is read — issuance into any strength is the specific thing to check there.

Catalyst Calendar (next 30 days)

• 2026-08-31 — WhiteFiber targets full run-rate billing across the 40 MW contracted IT load at NC-1 (North Carolina). Converts contracted capacity into recognized revenue and sets the Q3 revenue slope for the stake.
• ~2026-09-30 (est.) — WhiteFiber's stated Q3 2026 commercial-launch target for the cross-datacenter fabric tested 2026-07-09 at 111.2 Tbps over 83km of dark fiber.
• ~2026-09-30 (est.) — Quarter-end ETH mark. ETH near $1,873 on 2026-08-14 versus ~$2,334 average treasury cost determines whether another non-cash charge lands in the Q3 GAAP line.

Elapsed catalysts

• No company print inside the window. Next quarterly report ~mid-November 2026 (est.); the 2026-08-13 date has passed. (passed 13d ago)

What Would Change Our Mind

The structural break is the range floor going: the $1.40–$1.70 band held since mid-July fails and the $1.25 52-week low is back in play — specifically, a weekly close below $1.40 would confirm the discount is being validated rather than closed. Secondary conditions that would do the same work: a WYFI weekly close below $25, which cuts the stake mark below the level that made the sum-of-parts argument in the first place; ETH under $1,700, which pushes the treasury further from its ~$2,334 average cost and guarantees another mark against Q3 GAAP; or 2026-08-31 passing with no confirmation that NC-1's 40 MW is billing at run rate, which would be the first slipped operational date in the WhiteFiber build.

The other direction is equally specific: a weekly close above $1.70 while WYFI holds above $29 would turn the discount into a trending structure instead of a static one, and a 10-Q share count materially above ~353.4M would say the parent funded itself by issuing into the bounce.

Correlation Notes

BTBT is not an independent instrument — roughly $798M of its asset side is a single listed security (WYFI at $29.52, 2026-08-14) and the rest is an ETH book marked to spot. Directionally it tracks WYFI first and ETH second, and the two inputs are currently pulling opposite ways: WYFI +4.18% on 2026-08-14 with ETH drifting off $1,884 to ~$1,873 the same morning. Read alongside the neocloud/GPU-rental complex for the stake mark and the ETH-treasury vehicles for the discount regime; the parent will not re-rate on its own operating numbers while consolidated results are dominated by non-cash digital-asset marks. Bitcoin mining is winding down, so the legacy hashprice correlation that defined the name before 2025 no longer applies.