The comfortable reading is that BigBear.ai simply failed to hold a contract rally. That reading is wrong. The shares closed at $2.16 on Friday 16 October 2026, down 7.30 percent from Thursday’s $2.33, and the year-end path from the last 90 sessions is a bull of $2.22, a base of $1.52 and a bear of $1.18. That bull is not the $4.00 one-year target Nasdaq still displays. The gap the coverage skips is the company’s own footnote. On 13 October BigBear.ai said it had secured more than $225 million in contract awards since July. The footnote calls that figure total potential value, including unexercised options and ceilings, and says it does not necessarily reflect funded amounts. October left the $135 million to $165 million revenue range where it was.
The closes used here are Friday’s official prints from the Nasdaq historical table saved for this piece. The bull, base and bear in the title are the distribution maths below, not a street target. This is not financial advice.
Key facts
- Friday 16 October 2026 official close was $2.16, down 7.30 percent from Thursday’s $2.33, on volume of 15,522,020 shares. Session range $2.14 to $2.32. Nasdaq historical table, saved 17 October 2026.
- Nasdaq’s quote summary put the 52-week range at $2.12 to $8.68 and a one-year target at $4.00. Thursday’s low was $2.12. Market value was $1,035,708,105, equal to $2.16 times the 479,494,493 shares outstanding at 30 June. Nasdaq quote summary, saved 17 October 2026.
- On 13 October 2026 the company said it had secured more than $225 million in total potential contract value since July, including more than 25 new generative AI awards. The footnote says the figure includes unexercised options and ceilings and is not necessarily funded. BigBear.ai release, 13 October 2026.
- Second-quarter revenue was $36.7 million, up 13 percent. Gross margin was 32.8 percent. Adjusted EBITDA was a loss of $11.6 million. Cash and investments were $409.8 million. Full-year guidance of $135 million to $165 million was affirmed and was not updated in October. BigBear.ai results, 30 July 2026.
- Backlog at 30 June was $269.597 million, of which funded backlog was $79.415 million and priced unexercised options were $148.996 million. Form 10-Q, filed 30 July 2026.
- Ask Sage contributed $6.098 million of second-quarter revenue. Excluding it, the quarter was $30.651 million, 5.6 percent below the year-earlier $32.472 million. Same 10-Q, arithmetic on the company’s figures.
- On 31 July 2026 the company agreed with Jefferies to offer up to 100,000,000 shares at the market. The agent’s compensation is up to 3.0 percent of gross proceeds. As of 30 June, the 10-Q said no capacity remained under the older programmes. Form 8-K, 31 July 2026.
- Third-quarter results are due Tuesday 27 October 2026, release at about 4:15 pm Eastern, call at 4:30 pm. BigBear.ai, 9 October 2026.
What just happened, and why the obvious reading is wrong
The week is not a two-day story. On Tuesday 13 October the company announced the awards at 9:15 am Eastern. The stock opened at $2.31, traded as high as $2.57, and closed at $2.39, up from Monday’s $2.30, on 36,645,240 shares. Wednesday already closed at $2.18. Thursday bounced to $2.33, and Friday gave that bounce back. Friday’s close was $2.16. From the Tuesday high to the Friday close the give-back is 16.0 percent. From the Tuesday close it is 9.6 percent.
The customer names are real, and they are not a revenue print. The release lists more than 25 generative AI awards at the FBI, NASA, the Coast Guard, the Department of Energy and Veterans Affairs, plus CBP recompetes with SAIC and Unissant, work at Central Command and Southern Command, and classified intelligence awards. The footnote says total contract value “represents the total potential value of the awards, including unexercised options and ceiling values, and does not necessarily reflect funded amounts.” It also says the revenue ultimately realized “may differ materially from the TCV presented.”
The same page cites a 30 June backlog of about $270 million, about 29 percent funded. The 10-Q splits that book into funded $79.415 million, unfunded $31.709 million, priced unexercised options $148.996 million and unpriced options $9.477 million, total $269.597 million. Priced options not yet exercised are 55 percent of it. Neither October release revises $135 million to $165 million, and the award note does not call the $225 million additive to that backlog.
Who is exposed
The company is BigBear.ai Holdings, Inc., NYSE: BBAI, headquartered in McLean, Virginia. It sells mission-ready AI to national security and to trade and travel customers. Customs and Border Protection work continues through SAIC on one recompete and Unissant on the other. The FBI and the Coast Guard are described as new. Central Command, Southern Command and unnamed intelligence customers sit where a funded dollar cannot be inspected from the outside.
The income statement is the other exposure. Second-quarter operating loss was $28.558 million on $36.749 million of revenue, and adjusted EBITDA was a loss of $11.6 million, worse than $8.5 million a year earlier, after selling, general and administrative expense rose $10.4 million. Operating cash use was $22.207 million in the quarter and $40.208 million in the half. Cash and investments of $409.8 million, against current debt of $16.643 million and equity of $770.324 million, are why a $2 stock is not a call on a thin balance sheet. Weighted-average shares were 479,119,921 against 320,591,204 a year earlier. Revenue rose 13 percent. The share count rose 49 percent.
As of 30 June the 10-Q said no at-the-market capacity remained, and 2026 cash flow shows zero proceeds from that channel. The 2025 programmes raised $293.431 million in the quarter ended 30 June 2025. On 31 July the company opened a new Jefferies programme for up to 100,000,000 shares, compensation up to 3.0 percent. No filing fetched here shows a sale under that agreement. Until the third-quarter report, the exposure is the option.
A full take-down at $2.16 is $216 million gross, about $209.5 million after a 3 percent commission, and lifts the count 20.9 percent to 579,494,493. Holding Friday’s equity value fixed and ignoring cash raised, that is $1.79 a share. That $1.79 is share-count arithmetic, not the bear. The bear is lower because the last 90 sessions have a negative drift. Using both at once would double-count.
What the tape and the filings actually show
The week splits into a headline, a backlog definition, a revenue range that did not move, and a distribution that does not reach the displayed target.
| Item | Figure | Date | What it is not |
|---|---|---|---|
| Friday close | $2.16, down 7.30 percent | 16 October 2026 | Not Tuesday’s $2.57 high |
| Award headline | More than $225 million total potential value | 13 October 2026 | Not funded revenue |
| Funded backlog | $79.415 million of $269.597 million | 30 June 2026 | Not the October ceiling |
| Priced unexercised options | $148.996 million | 30 June 2026 | Not revenue, and not exercised |
| Revenue guide | $135 million to $165 million | Affirmed 30 July, unchanged in October | Not a raise on the awards |
| Second-quarter revenue | $36.749 million, up 13 percent | Quarter ended 30 June 2026 | Not organic: Ask Sage was $6.098 million |
| Shares and new ATM capacity | 479,494,493 out, up to 100 million more | 30 June and 31 July 2026 | Not a counted sale since July |
| Bull, base, bear in this piece | $2.22, $1.52, $1.18 | 52 sessions to 31 December 2026 | Not Nasdaq’s displayed $4.00 target |
First-half revenue was $71.184 million. The $135 million floor leaves about $31.9 million a quarter still to book, below the second quarter’s $36.749 million, so the floor needs Ask Sage to stay, not a surge. The $165 million top needs about $46.9 million a quarter. McAleenan said on 30 July the firm was “on track for our target of 17 percent revenue growth.” The midpoint is $150 million, so that target sits in the middle of the range.
The 10-Q says the quarter’s $4.3 million revenue increase was “primarily due to the inclusion of Ask Sage,” which contributed $6.098 million in the quarter and $12.242 million in the half. Excluding it, the quarter was $30.651 million against $32.472 million, down 5.6 percent, and the half was $58.942 million against $67.229 million, down 12.3 percent, which the filing ties to lower Army volume and contracts that did not recur. Equity value of $1.036 billion, less $409.8 million of cash and investments, plus $16.643 million of debt, is an enterprise value near $643 million, about 4.3 times the $150 million midpoint. That multiple does not assume the $225 million ceiling becomes 2026 revenue. Palantir’s quarter, in FinanceFeeds’ report on the 93 percent print, is a different scale, and the FAA award Palantir did not win is the nearer lesson: a government AI headline and the contract record are different documents.
What the company has said, and what it has not
Kevin McAleenan, chief executive of BigBear.ai, said in the 13 October award release: “Retaining critical work with CBP while adding the FBI, the Coast Guard and new defense and intelligence customers is proof of a strong foundation and an expanding customer base.” The company account on X posted the same morning that the awards were “both successful recompetes and new wins, with more than 25 new generative AI awards across federal agencies, including the FBI, NASA and the U.S. Coast Guard.” The status adds no funded figure.
Sean Ricker, chief financial officer of BigBear.ai, said in the 30 July results: “BigBear.ai is in an excellent position to take advantage of the rapid expansion of investment and innovation in defense technology, which shows no signs of slowing down, given the macro environment. We have steadily been maturing the underlying financial discipline of the company and have significant cash in reserve so that when the right opportunity presents itself, we can move fast.” In that same release McAleenan tied the quarter to $410 million of cash and investments, a 17 percent growth target, and accretive M&A. July is still the last time either officer spoke to the numbers.
No fetched release says the $225 million is funded or that the guide is rising. Nasdaq’s saved summary still shows a one-year target of $4.00, separate from the $2.22 bull. A named broker note was not re-fetched. Ricker’s line about moving fast is the risk in the cash: a stock deal, or a draw on the Jefferies programme, changes the count under the bear.
The funding rule that sits under the headline
There is no new statute here. The 10-Q says most federal contracts are not fully funded at award, the company generally does not deliver beyond appropriated funding, and termination for convenience can leave incurred costs rather than the full value. The October release lists shutdowns and lapses in appropriations among the risks. A ceiling is a poor stand-in for funded backlog. The $148.996 million of priced options on the June book is where a slip shows first.
FinanceFeeds has already reported the Pentagon putting commercial models onto a secure network for defense users. The demand is real. The conversion problem is this issuer’s: a backlog about 29 percent funded at the last balance-sheet date, a revenue line that fell once Ask Sage is removed, and an award total the company will not call funded.
Bull, base and bear
The bull, base and bear in the title are the distribution maths on the saved Nasdaq closes, not the $4.00 figure Nasdaq displays as a one-year target. Take the last 90 daily log-returns through Friday 16 October 2026. The mean, mu, is -0.00681870. The sample standard deviation, sigma, dividing by 89, is 0.03515996. N, the New York sessions after Friday through 31 December 2026, is 52. The count runs from Monday 19 October through Thursday 31 December, skips weekends, and skips only the full closures on Thursday 26 November and Friday 25 December. Friday 27 November and Thursday 24 December are early closes and still count. The square root of 52 is 7.21110.
- Base = $2.16 x exp(52 x -0.00681870) = $2.16 x exp(-0.35457) = $1.52.
- Bull = $2.16 x exp(52 x mu + 1.5 x sigma x sqrt(52)) = $2.16 x exp(0.02574) = $2.22.
- Bear = $2.16 x exp(52 x mu – 1.0 x sigma x sqrt(52)) = $2.16 x exp(-0.60811) = $1.18.
The drift is steeply negative after the fall from an $8.68 high. The bull still prints above the spot, so the formula keeps the drift. A no-drift 1.5-sigma move is $3.16, a sensitivity, not the bull, and even that does not reach $4 by 31 December.
| Case | Price | Versus the $2.16 spot | What has to be true |
|---|---|---|---|
| Bull | $2.22 | +2.8 percent | The drift breaks, 27 October still affirms $135 million to $165 million, some of the October total is called funded, and Jefferies is not drawn in size. A return to $4 is outside this maths. |
| Base | $1.52 | -29.6 percent | The recent mean continues. Revenue can stay inside the guide, because the floor is undemanding, without the share count exploding. |
| Bear | $1.18 | -45.4 percent | The guide is cut, the ex-Ask Sage line keeps falling, or a block of the 100 million share capacity is sold while the drift is still negative. This is one sigma on the current count, not that stacked case. |
Versus Friday, $2.22 is 6 cents. The base at $1.52 is about $729 million of equity on the June count. The bear at $1.18 is about $566 million, near the enterprise value the cash already implies.
What happens next
On Tuesday 27 October 2026 the release is due at about 4:15 pm Eastern and the call at 4:30 pm. A third-quarter print near $36.749 million, on a first half of $71.184 million, keeps the $135 million floor easy and leaves the fight at the $150 million midpoint. A print back toward the ex-Ask Sage run-rate makes the floor a second-half rescue. Affirming $135 million to $165 million, or staying silent after July’s affirmation and October’s skip, is the tell.
That filing is also the first regular report that can show shares sold under the 31 July Jefferies programme. A count still near 479.5 million leaves dilution as capacity. A higher count shrinks the claim on $409.8 million of cash before any view of the bear. Each 10 million shares, holding Friday’s equity value fixed, is about 4 cents. A named deal on the call changes the maths. A repeated intention to “move fast” does not.
Through Thursday 31 December, 52 sessions on, the base is $1.52 and the bull is $2.22. A close above $2.22 needs 27 October to supply a funded number the 13 October release did not. Another ceiling leaves the base in charge. The bear at $1.18 is a guide cut, or a large sale of new shares, with recent volatility running to the downside into year-end.
Frequently asked questions
Is the year-end bull only what the stock’s own volatility allows?
Yes. It is only what the stock’s own volatility allows. From Friday’s $2.16 close, the last 90 log-returns and 52 sessions through 31 December produce a bull of $2.22, a base of $1.52 and a bear of $1.18. Nasdaq’s displayed one-year target of $4.00 is a separate number. Even the no-drift version of the same upside, which is not the bull used here, is $3.16. The awards headline does not pull the distribution up to $4.
Did BigBear.ai announce $225 million of new revenue on 13 October?
No. The company announced more than $225 million of total potential contract value since July, across new wins and recompetes, including more than 25 generative AI awards. Its footnote says that figure includes unexercised options and ceiling values and does not necessarily reflect funded amounts. The release then cited a 30 June backlog of about $270 million, about 29 percent funded. The $135 million to $165 million revenue range was not changed.
Why did the stock fall after the contract news?
The shares rallied on 13 October, from a Monday close of $2.30 to a Tuesday high of $2.57 and a close of $2.39. Wednesday closed at $2.18, Thursday bounced to $2.33, and Friday closed at $2.16, down 7.30 percent on the day and 16 percent below Tuesday’s high. The release did not add a funded figure, did not raise guidance, and reminded readers that government contracts are funded incrementally and can be cut.
Where does the bear sit if BigBear.ai keeps issuing shares?
The distribution bear is $1.18 on the 479,494,493 shares outstanding at 30 June. The 31 July programme allows up to 100 million more shares through Jefferies, and the first-half cash flow shows that programme had not sold stock by 30 June. A full issue at Friday’s price would lift the count by 20.9 percent. Holding today’s equity value fixed, that count is about $1.79, which is share-count arithmetic, not a second bear.
When is the next hard date, and what would actually change the view?
Tuesday 27 October 2026, with the release at about 4:15 pm Eastern. The result that would challenge the base is an affirmed $135 million to $165 million range plus a funded figure for the October awards, and a share count that has not jumped. The result that confirms the base is another ceiling with the guide untouched. A cut to the range, or a large disclosed draw on the at-the-market programme, is the bear’s evidence.
How were the bull, base and bear calculated?
From the last 90 daily log-returns in the saved Nasdaq series through 16 October 2026. Mu is -0.00681870 and sigma is 0.03515996, using the sample standard deviation. N is 52 sessions. Base is the spot times exp of N times mu. The bull adds 1.5 sigma times the square root of N. The bear subtracts 1.0 sigma times the square root of N. Rounded to the cent, that is $2.22, $1.52 and $1.18.
