🌾 Welcome to StableBread's Newsletter!
This post breaks down my quarterly analysis for six names I’m still following from the write-up archive, in the order they filed: QXO (QXO), Boost Run (BRUN), Reviva Pharmaceuticals (RVPH), DigiPowerX (DGXX), Agility Robotics (AGLT), which merged with the SPAC Churchill Capital Corp XI (CCXI) on November 4, and Newbury Street II (NTWO), the SPAC taking FORT Robotics public.
For paid subscribers, I’ll also share whether I'm buying, selling, or holding each name (not all are portfolio holdings), plus key catalysts to watch next.
QXO
QXO (NYSE: QXO), the building products distributor run by CEO Brad Jacobs, held its Q3 call on November 9 and filed its full Q3 results on November 12. Q3 was QXO’s first full quarter with TopBuild, the insulation installer and distributor QXO acquired on July 1.
QXO closed at $12.60 on November 13, below the $12.80-13.60 bear case from my deep dive, but up 7.9% since my November 3 macro update, where I said I’d need (1) a clean quarter of blended margins and (2) lower rates before adding.
Margin Test
QXO’s adjusted EBITDA margin rose to 11.4% from 11.1% a year ago, but TopBuild carried the increase.
If TopBuild earned the 17.8% margin it guided for 2026, the rest of QXO earned ~8.1%, below the 8.4% of Q2, which already included Kodiak (the lumber and building materials distributor QXO bought on April 1):

QXO: Adjusted EBITDA Without TopBuild (Form 10-Q; Forms 8-K; TopBuild Form 8-K)
In the legacy roofing business, roofing product sales fell 4.3% y/y to $2.0B, even as pricing in roofing and waterproofing went from -1.1% y/y in April to +4.1% in September.
Still, sales picked up after the quarter. Jacobs said on the Q3 call that October daily sales rose 5.4% y/y company-wide, and 9.1% in roofing and waterproofing, the business QXO has owned the longest.
Debt
Net debt of ~$8.45B as of September 30 landed within the $7-9B I expected in June, but it’s ~4.8x the ~$1.75B of consensus adjusted EBITDA Jacobs used as his starting point on the Q3 call, not the 3.3-4.3x I estimated on $2.1B.
Interest (including a one-time $15M fee on an unused bridge loan) and preferred dividends took $229M in Q3, 41.1% of the $557M of adjusted EBITDA, and operating cash flow was only ~$1M, down from $213M in Q3 2025.
The $4B of adjusted EBITDA QXO is targeting for 2030 assumes all free cash flow until then goes to paying down debt, Jacobs said, and spending the same cash on small acquisitions could lift 2030 adjusted EBITDA to $5.5B instead.
Jacobs also announced his ninth company on the Q3 call, leading a $1B investment in CKX Lands (NYSE American: CKX), a land management company, with $900M of his own money. He’ll become CKX’s chairman and CEO while remaining chairman and CEO of QXO, and said the industry for the new venture is “still to be determined.”
QXO also named Dev Ahuja, CFO of aluminum maker Novelis since 2016, to replace CFO Ihsan Essaid in January 2027, with Essaid moving to vice chairman to oversee M&A and capital raising.
Valuation
At $12.60/share, QXO has ~1.46B diluted shares, not 1.04B, counting the in-the-money convertible preferred and warrants from the $1B investment Jacobs’ firm led in June 2024, plus its mandatory convertible preferred, stock awards, and pre-funded warrants.
So QXO’s equity is worth ~$18.4B, and with the $3.0B of Series C preferred and ~$8.45B of net debt, QXO’s EV is ~$29.9B, ~17x the ~$1.75B of consensus adjusted EBITDA.
Boost Run (BRUN)
Boost Run (Nasdaq: BRUN), the AI data center company that rents out NVIDIA GPUs to AI startups, reported Q3 on November 12, and a data center delay in Georgia pushed $12.7M/month of revenue into Q1 2027.
What I Expected vs. What Happened
Here’s how Q3 compares with what I said I’d watch in October:

BRUN: Q3 2026 vs. My October 19 Watch Items (Form 10-Q; Form 8-K; Q3 2026 earnings call)
Q3 revenue of $39.9M came in 23.6% below the $52.2M implied by a straight line to $400M of ARR, and CFO Erik Guckel said year-end ARR would’ve been $435M without the Georgia delay.
CEO Andrew Karos said a city council request paused progress at the Georgia site after permitting, an issue he says is now resolved. But the project, worth $144-152M of ARR, won’t go live until Q1 2027.
The delay leaves year-end ARR at $283-291M, right at my $280M bear case. If the Georgia project bills by March 31, 2027, ARR clears my $400M base case a quarter late:

BRUN: ARR vs. My October Cases (Q2 and Q3 2026 earnings calls; Q3 2026 investor presentation)
Financing
CEO Karos said Boost Run has secured $600M of non-dilutive project financing in 2026 so far, and the 10-Q confirms $486.9M of new GPU leases through September 30, plus five more equipment leases signed in October that require $83.1M of payments.
So I’m dropping the $62.2M stock sale from my bear case. Even after counting the 6.23M restricted stock units (RSUs) Boost Run granted in Q3, mostly to its CFO, CTO, and COO, the bear case rises to $5.04/share.
Boost Run paid $96M of its own cash up front on two of its three Q3 GPU leases, and with the new leases, GPU lease payments due in 2027 rose to $186.2M (as of September 30) from $103.5M (as of June 30).

BRUN: Where Q3 2026 Cash Came From and Went (Q3 2026 earnings call; Form 10-Q)
Warrant exercises brought in $88.2M in Q3, and all 11.4M warrants are now exercised, so don’t expect any further cash or dilution from warrants.
Valuation
At $5.53/share, Boost Run trades 9.7% above my $5.04 bear case.
My base case, $400M of ARR at 4.2x (CoreWeave’s multiple of year-end ARR and the median of the three peers I used), is worth $15.88/share after counting the RSUs, a quarter later than I expected.
Reviva Pharmaceuticals (RVPH)
Reviva Pharmaceuticals (OTCQB: RVPH), the biotech developing brilaroxazine for schizophrenia, filed its Q3 10-Q on November 13 and plans to apply to relist on Nasdaq in H1 2027.
Its December 18 annual meeting asks shareholders to approve a reverse split of up to 1-for-100, nine months after its last reverse split.
What Changed
Reviva’s timelines moved in both directions between the Q2 and Q3 reports:

RVPH: What Changed Since the Q2 Report (Form 10-Q; Form 8-K; DEF 14A)
The USPTO granted Track One (its fast-track review) to Reviva’s August 11 patent application on a new formulation of brilaroxazine, which could extend patent protection to 2046.
That 2046 date matters for a partnership or sale, since Reviva’s original patents expire in 2030 (possibly 2035 with an extension after approval), and the short patent life was big pharma’s main objection to the drug, per CEO Bhat.
Track One aims for a final decision within ~12 months, although that decision can be a rejection (the evidence suggests that’s unlikely).
Relisting
The shelf registration behind Reviva’s ATM expires February 13, 2027, and with a public float under $75M, Reviva requires an exchange listing to file a new shelf.
CEO Bhat estimated in April that RECOVER-2 (Reviva’s second Phase 3 trial of brilaroxazine) and the FDA filing would cost $50-55M, ~3x Reviva’s $17.5M of cash as of September 30.
So back on Nasdaq, Reviva could sell stock to fund most of the $50-55M before RECOVER-2 starts, which would end the pause in dilution I noted in August.
The 10-Q says Reviva does “not currently have any committed external sources of capital,” and the Q3 release says Reviva is “pursuing partnership opportunities.”
Valuation
At $0.353/share, Reviva is valued at ~$4.6M, 26.5% of its $17.5M of cash (~$1.33/share).
But even the $32.5-37.5M of the $50-55M that cash doesn’t cover is ~7-8x Reviva’s market cap:

RVPH: Market Cap vs. Cash and Funding Need (Form 10-Q; April 2026 fireside chat)
DigiPowerX (DGXX)
DigiPowerX (Nasdaq: DGXX), which is building an AI data center campus in Columbiana, Alabama for AI chipmaker Cerebras (CBRS), reported Q3 on November 13 and still expects to close the $200M loan to finish the campus in Q4 2026.
I’ve covered DigiPowerX since April at $3.50/share, before the Cerebras deal, and added below $4.00 in August.
What I Expected vs. What Happened
Here’s where each item I said I’d watch in August stands after Q3:
Loan: Still not closed. CEO Michel Amar said on the Q3 call that DigiPowerX has the loan documents and expects to close “definitely in Q4 2026” at a single-digit interest rate.
ATM: The share count has held at 101.4M since August 14, but DigiPowerX has already sold $174.7M of the $175M its ATM prospectus covers, including 2.85M shares for $11.3M in Q3, even though an October 13 release said it “did not issue any shares” under the ATM that quarter.
Phase 1: Still on track, with Cerebras getting access on December 15 and billing from January 2027. Phase 2 access is targeted for March 2027, “subject to various conditions,” with billing from April.
Q3 revenue: Up 85% q/q to $12.2M, short of the more-than-100% increase guided on the Q2 call.
Adjusted EBITDA: $0.6M, now excluding currency moves (a $0.5M loss this quarter), after Q2’s figure included a $7.7M currency gain.
Funding
Amar said DigiPowerX generated $4.7M of positive cash flow in Q3 with the help of a $12M upfront payment from Cerebras, which he called “non-refundable.”
But the 10-Q says the payment is refundable to Cerebras in certain circumstances before Phase 1 starts, including if DigiPowerX misses the final delivery date, less what’s been spent on finished construction.
Even with its cash, crypto, and the loan, DigiPowerX is ~$30M short of its ~$350M 2027 plan, mostly for ~10 MW of AI inference capacity and Columbiana’s Phase 2, before any Q4 2026 spending:

DGXX: 2027 Capex Plan vs. Funding in Hand (Form 10-Q; Form 8-K)
Amar said DigiPowerX will fund the 2027 plan mainly with customer deposits and prepayments, along with debt and its cash, so deposits like the $12M Cerebras paid up front, or more debt, would have to cover the ~$30M gap.
He also said new customers now offer deposits of 30%, sometimes up to 50-70%, vs. 15% six months ago.
Valuation
At $2.28/share, DigiPowerX has a market cap of ~$231M and an EV of ~$112M after its $95.2M of cash and $24.5M of bitcoin and ether.
Amar targets $250-300M of annualized revenue by the end of Q3 2027, subject to financing, and said $143M of that run rate is supported by current operations and the Cerebras agreement, vs. ~$49M of Q3 revenue annualized:

DGXX: Annualized Revenue vs. Its Q3 2027 Target (Q3 2026 earnings call; Form 8-K)
Agility Robotics (AGLT)
Agility Robotics (Nasdaq: AGLT), the maker of the Digit humanoid robot, closed its merger with Churchill Capital Corp XI (CCXI) on November 4 and filed its Q3 financials on November 13.
Early access to Digit 5, the next version of Digit, is now expected to begin in H1 2027, after Chief Business Officer Daniel Diez told Forbes in August that Digit 5 would reach early customers in December 2026.
In my September 6 deep dive, at $13.40/share, I said it was too early to buy and that I’d consider Agility after the post-close washout, once lock-up supply clears and the Digit 5 ramp is clearer.
AGLT closed at $7.36 on November 13, down 36.1% from its $11.51 close on November 5, its first day trading as AGLT, and 28.5% below the ~$10.29 that redeeming holders received.
Cash Situation
With 61.5% of public shares redeemed, Agility netted ~$334.1M from the deal, vs. ~$576M with no redemptions, the full-trust case I said would fund three years without another raise:

AGLT: Merger Cash, No Redemptions vs. Actual (Form 8-K/A; Forms S-4 and 8-K)
Agility spent $131.6M of cash on operations in the first nine months of 2026 (up 87% y/y), plus $6.8M on capex, while revenue fell 79% y/y to $0.36M, since the 2025 period included one-time robot sales.
So with the ~$334.1M from the deal and $64.9M of cash as of September 30, Agility has ~$399M, a little over two years at its ~$184M/year cash burn (operations plus capex), while Digit 5 rolls out through 2027.
Valuation
At $7.36/share, Agility’s market cap is ~$2.26B on ~307M shares, counting ~61.0M options and RSUs and 4.2M warrants.
Insiders and the sponsor stay locked up until early May 2027, but the 20.1M shares from the PIPE (the private placement that funded at the merger close), more than the ~16M public shares left, are being registered for resale.
Newbury Street II (NTWO)
Newbury Street II (Nasdaq: NTWO), the SPAC taking robot-safety maker FORT Robotics public, filed its S-4 (the merger filing) on November 13, with FORT’s numbers through September 30.
In my November 8 update, I said I wasn’t buying NTWO at $11.00, and that FROB (the ticker after the merger) stays a pass until the S-4 shows what $41-59M of cash would fund. NTWO closed at $10.89 on November 13.
What the S-4 Shows
FORT’s revenue grew 53% y/y to $12.5M in the first nine months of 2026, but operating expenses grew 93%, and cash fell to $1.7M as of September 30:

NTWO: FORT Robotics, Nine Months Ended September 30 (Form S-4)
The S-4 says FORT’s losses and low cash “initially raised substantial doubt” about whether FORT could continue as a going concern, which $10.0M of new SAFEs from existing investors (October 21 to November 6) and $1.5M more in bank loan draws alleviated.
After 82% of public shares were redeemed at the October 28 extension vote, the S-4 estimates FORT will have $36.0-54.2M of cash at closing, including those SAFEs, after $17.8M of deal costs (vs. $19M in the August investor deck).
The $500M that FORT’s existing holders get in the merger (50.0M shares and options, valued at $10.00/share) is unchanged, there’s no fairness opinion, and the S-4 says the board “did not receive any financial projections of Fort Robotics.”
Valuation
At $10.89/share, NTWO’s price implies an EV of ~$582M for FORT at either end of the cash range, or 36.7x FORT’s $15.9M of TTM revenue.
At the 7.5-11.2x sales multiples from my September 16 deep dive, FORT would need $52-78M of revenue to justify ~$582M, 3.3-4.9x its TTM revenue:

NTWO: FORT Robotics Revenue vs. What the Price Requires (Form S-4)
Disclaimer: This write-up is for informational and educational purposes only and is not investment advice. I may hold positions in the securities discussed. Do your own research before buying or selling any security.
